- Email: [email protected]
Interferences in Prompt γ Analysis of corrosive contaminants in concrete
ARTICLE IN PRESS
Nuclear Instruments and Methods in Physics Research A 569 (2006) 803–809 www.elsevier.com/locate/nima
Interferences in Prompt g Analysis of corrosive contaminants in concrete A.A. Naqvia,, M.M. Nagadia, O.S.B. Al-Amoudib a
Department of Physics, King Fahd University of Petroleum and Minerals, KFUPM Box 1815, Dhahran-31261, Saudi Arabia b Department of Civil Engineering, King Fahd University of Petroleum and Minerals, Dhahran, Saudi Arabia Received 12 May 2006; received in revised form 11 September 2006; accepted 13 September 2006 Available online 13 October 2006
Abstract An accelerator-based Prompt Gamma Neutron Activation Analysis (PGNAA) setup has been developed to measure the concentration of corrosive chloride and sulfate contaminants in concrete. The Minimum Detectable Concentration (MDC) limit of chlorine and sulfur in the concrete depends upon the g-ray used for elemental analysis. For more interfering g-rays, the MDC limit is higher than that for less interfering g-rays. The MDC limit of sulfur in concrete measured for the KFUPM PGNAA setup was calculated to be 0.6070.19 wt%. The MDC limit is equal to the upper limit of sulfur concentration in concrete set by the British Standards. The MDC limit of chlorine in concrete for the KFUPM PGNAA setup, which was calculated for less interfering 1.165 MeV g-rays, was found to be 0.07570.025 wt%. The lower limits of the MDC of chlorine in concrete was 73% higher than the limit set by American Concrete Institute. The limit of the MDC can be improved to the desired standard by increasing the intensity of neutron source. For moreinterfering 5.715 and 6.110 MeV chlorine g-rays the MDC limit was found to be 2–3 times larger than that of 1.165 MeV g-rays. When normalized to the same intensity of the neutron source, the MDC limits of chlorine and sulfur in concrete from the KFUPM PGNAA setup are better than MDC limits of chlorine in concrete obtained with the 241Am–Be source-based PGNAA setup. This study has shown that an accelerator-based PGNAA setup can be used in chlorine and sulfur analysis of concrete samples. r 2006 Elsevier B.V. All rights reserved. PACS: 82.80.Jp Keywords: PGNAA study; Concrete samples; Sulfur and chlorine measurement; 2.8 MeV neutrons; Monte Carlo simulations; Corrosion studies
1. Introduction The corrosion of reinforcing steel in concrete structures is a major problem faced by the building and construction industry [1]. The chloride and sulfate salt present in concrete initiate reinforcing steel corrosion [1–3]. Chloride-induced reinforcement corrosion is the principal cause of deterioration in concrete structures in the Arabian Gulf and the world over. In view of this, the American Concrete Standards Codes specify the allowable chloride concentrations in concrete as 0.03% by weight of concrete (or 0.15% by weight of cement). [2]. Similarly, the allowable limit of sulfate concentration in concrete is Corresponding author. Tel.: +966 3860 4196; fax: +966 3860 2293.
E-mail address: [email protected] (A.A. Naqvi). 0168-9002/$ - see front matter r 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.nima.2006.09.058
0.60% by weight of concrete (or 4.0% by weight of sulfur) [4]. These limits are being utilized by the construction industry as the corrosion–threshold concentrations in concrete for chloride and sulfate ions. Preventive measures against corrosion require maintaining the chloride and sulfur concentration in concrete below the threshold limits [2]. This requires monitoring the chloride concentration in concrete using a non-destructive technique. The Prompt Gamma Neutron Activation Analysis (PGNAA) technique can be used to monitor concentration of corrosive elements in concrete samples [5–12]. This technique has been successfully used to analyze the elemental composition of concrete samples using 241Am–Be radioisotope neutron sources [5–7]. The disadvantages of a radioisotope neutron source-based PGNAA setup are high g-ray dose, permanent
ARTICLE IN PRESS 804
A.A. Naqvi et al. / Nuclear Instruments and Methods in Physics Research A 569 (2006) 803–809
radiation hazards and limited life of the neutron source due to the short half-life of the radioisotopes source. It is desired to develop an accelerator-based PGNAA setup to monitor the chloride and sulfur concentration in concrete. The accelerator-based neutron source has several advantages such as controlled mechanism of neutron production and less radiation hazard [9–12]. The On–Off control of accelerator power ensures that the acceleratorbased PGNAA setup has radiation hazards only when in use. Furthermore, commercially manufactured compact accelerators have the physical dimensions almost comparable to with that of the radioisotope neutron source. Also the neutron flux available with a compact neutron generator is greater than that available with a commercial radioisotope source. A compact neutron generator can produce 3 108 neutrons/s or more [13] while a 5 Ci 241 Am–Be can produce 3 107 neutrons/s [14]. In short, an accelerator-based PGNAA setup is better, or at the least comparable to a radio-isotope neutron source-based PGNAA analyzer. In this regard, several studies have been undertaken at the King Fahd University of Petroleum and Minerals (KFUPM) 350 keV accelerator laboratory to develop an accelerator-based PGNAA setup. The setups were used to analyze concrete and cement samples [9–12]. An accelerator-based PGNAA facility was designed to analyze large samples of concrete [11–12]. The KFUPM accelerator, which is a 350 kV electrostatic accelerator, produces 2.8 MeV neutrons beam via D(d,n) reaction. Although the KFUPM accelerator has a limited source intensity of 3 105 n/s [11], the chlorine concentration in concrete was successfully measured using a large sample [12]. The MDC limit of a PGNAA setup depends upon several factors such as neutron source intensity, g-ray background under the peak etc. The g-ray background depends upon the g-ray used to characterize the element in
PGNAA analysis. For the interfering g-rays, the g-ray background is larger. In this study a prompt g-ray analysis has been carried out on corrosive chlorine and sulfur contaminants in concrete. Various g-ray lines from the corrosive contaminants have been studied to determine the chlorine and sulfur MDC limits in concrete. There are strong interferences between the g-rays from chlorine, sulfur and the concrete constituents. The prompt g-ray yield from sulfur has been measured using the KFUPM accelerator-based PGNAA setup, and sulfur concentration in concrete has been determined. Also, chlorine concentration in concrete has been analyzed for previously taken data [12] using strongly interfering 5.715 and 6.110 MeV chlorine g-rays. Earlier, the chlorine concentration in the concrete samples was analyzed using less interfering 1.165 and 7.431 MeV prompt g-rays [12]. The results of the recent g-ray analysis of sulfur and chlorine corrosive contaminants in concrete are described below.
2. Prompt c-ray yield measurements from sulfur The prompt g-ray yield from sulfur has been measured using the KFUPM accelerator-based PGNAA setup shown in Fig. 1. The setup has been reported earlier in detail elsewhere [11,12]; however, for sake of continuity it will be briefly described here. The setup mainly consists of a cylindrical sample enclosed in a rectangular paraffin moderator. The moderator is placed between a neutron target and a cylindrical NaI g-ray detector having a dimension of 25.5 25.5 cm (diameter height) [11]. The 2.8 MeV neutrons from the D(d, n) reaction are thermalized in the moderator and are captured in the concrete sample. The NaI g-ray detector measured the experimental yield of prompt g-rays from the capture of thermal neutrons in the concrete samples.
Fig. 1. Schematic representation of the PGNAA setup used to measure prompt g-ray yield [12].
ARTICLE IN PRESS A.A. Naqvi et al. / Nuclear Instruments and Methods in Physics Research A 569 (2006) 803–809
3. Results and discussion 3.1. Prompt g analysis of sulfur in concrete samples Fig. 2 shows the prompt g-ray pulse height spectrum of three concrete samples containing sulfur in different weight proportions. An important feature of Fig. 2 is the presence of a peak of 5.421 MeV prompt g-ray due to the capture of thermal neutrons in sulfur present in the concrete samples. The highest peak of sulfur in Fig. 2 corresponds to a 6 wt% concentration, while the smallest peak corresponds to 2 wt% and the middle one corresponds to 4 wt%. The 5.421 MeV g-ray peak is located between the channels 550 and 600, next to the single escape peak Ca(S) from calcium in concrete. The location of the sulfur 5.421 MeV peak coincides with the double escape peak of calcium but the intensity of the double escape peak from calcium is insignificant (typically less than 0.5% of the full energy peak). Therefore, the background under the sulfur peak, due to the double escape peak of calcium peak is insignificant. This is clearly shown in Fig. 2, where height
12000
Si(FE), 4.94
Ca(SE), 6.42
Ca(FE), 6.42
10000
COUNTS
8000
6000
S(FE), 5.42
4000
2000
4.75
5.75
6.75 Energy (MeV)
7.75
Fig. 2. Prompt g-rays spectra of the concrete samples above 3 MeV energy, containing 2.0, 4.0 and 6.0 wt% sulfur. The sulfur peak intensity increases with sulfur concentration, top most spectrum for 6 wt% sulfur concentration, middle one for 4 wt% and bottom most spectrum for 2.0 wt% sulfur concentration.
40000 5.42 MeV Gamma Ray Experimental Yield
The sulfur contaminated concrete samples used in the study were 14 cm long and 12.5 cm in radius. The procedure used to calculate their size, using the Monte Carlo simulations, is described elsewhere [11,12], The front moderator thickness used here was 5 cm. The sulfurcontaminated concrete samples were prepared by mixing the sulfur powder with the concrete mix in 2, 4, 6 and 8 wt% proportions. The concrete mix was prepared by mixing the concrete ingredients in the following ratio: Portland cement (370 kg), coarse aggregate (1110 kg); fine aggregate (680.4 kg) and water (198 kg) [11]. Then the concrete ingredients and the sulfur powder were mixed together and cast in specially designed 14 cm long plastic moulds with a radius of 11.5 cm. After demoulding, the concrete samples were cured in water for 7 days and then air dried in the laboratory environment (22721 C) for 7 more days. Thereafter, they were dried in an electric oven at 701 C temperatures for about 3 days until a constant weight was obtained [11]. The sulfur contaminated concrete samples were irradiated in the KFUPM PGNAA setup with the rectangular moderator [12]. The g-ray yield from the sulfur mixed concrete samples was measured using pulsed beam of 2.8 MeV neutrons produced via D(d, n) reaction. The deuteron pulse has a width of 5–6 ns and a repetition rate of 31.25 kHz [12]. The experiment was conducted with a typical beam current of 3.5–4.8 mA. The prompt g-rays produced in the concrete samples were detected by a 25 25 cm (diameter length) NaI detector. The prompt g-ray data were acquired for a preset number of charges measured at the electrically isolated neutron-producing target. The prompt g-ray detector signals were acquired in coincidence with a gate signal being derived from the beam pick-up signal.
805
Exp. Yield-5.42 MeV Monte Carlo fit-5.42 MeV
30000
20000
10000
0 0
2
4
6
8
10
Sulfur concentration (wt. %) Fig. 3. Experimental yield of 5.42 MeV prompt g-rays from sulfur in the four concrete samples plotted as a function of sulfur concentration.
of the sulfur peak increases with increasing sulfur concentration in concrete, over 2–6 wt% range. The net counts under each peak of 5.421 MeV g-ray were obtained by subtracting background from the sulfur gamma peak in the pulse height spectra of all four samples. Counts under each peak were integrated and normalized to the same amount of integrated beam charge measured at the target. The experimental yield of 5.421 MeV prompt g-ray from sulfur in the four concrete samples is plotted in Fig. 3 as a function of the sulfur concentration. Within the
ARTICLE IN PRESS A.A. Naqvi et al. / Nuclear Instruments and Methods in Physics Research A 569 (2006) 803–809
806
experimental uncertainties, there is a linear correlation between the 5.417 MeV prompt g-ray yield and the sulfur concentration. Solid lines in Fig. 3 represent the calculated yield of the 5.417 MeV prompt g-ray from the sulfur obtained through the Monte Carlo simulations. Consequently there exists an excellent agreement between the experimental and the calculated yield of 5.421 MeV sulfur g-rays from the four samples. 3.2. Prompt g-analysis of pure concrete samples For a better understanding of the g-ray spectrum of the pure concrete sample, it may be worthwhile to discuss the mutual interference amongst the g-rays emitted by the concrete constituents. The data is taken from ref. [12]. Fig. 4 shows an experimentally measured g-ray spectrum of a pure concrete sample above 3 MeV showing full energy and single escape peaks of calcium and silicon [12]. Also superimposed on the spectrum is the g-ray background spectrum (lower spectrum) taken without the sample. As shown in Table 1, there is interference between the g-rays from calcium and silicon in concrete. To be consistent with experimental data, where only g-rays with energies in excess of 3 MeV were analyzed, we will focus upon those prompt g-rays with energies above 3 MeV. The calcium has two g-rays with energies above 3.0 MeV, namely 4.42 MeV (14.9%) and 6.42 MeV (38.9%) energies and intensities respectively. Silicon has two prompt g-rays with energies 3.54 MeV (68%) and 4.94 MeV (62.7%). The full energy peak of 4.42 MeV (14.9%) from calcium interferes with the single escape peak of 4.94 MeV (62.7%) g-ray from silicon and its intensity is added under the peak resulting in the higher intensity of the single escape peak than the corresponding full energy peak of 4.94 MeV g-ray from silicon, as shown in Fig. 4.
14000 Si(FE), 3.54
Si(SE), 4.94 Si(FE), 4.94 Ca(SE), 6.42 Ca(FE), 6.42
12000
Electronics sum Peak
COUNTS
10000 8000 6000 4000 2000 0 3.09
4.83 6.56 Energy (MeV)
8.30
Fig. 4. Experimentally measured prompt g-ray spectrum of the concrete sample above 3 MeV showing full energy and single escape peaks of calcium and silicon. Also superimposed on the spectrum is the g-ray background spectrum (lower spectrum) taken without the sample [12].
Table 1 Energies and intensities of capture g-rays of concrete [15] Element
g-rays energy (MeV)
Intensity (%)
Calcium
1.942 4.418 6.420
72.5 14.9 38.9
Silicon
3.539 4.934
68 62.7
Aluminum
7.724
27.4
Iron
7.631 7.646
28.5 24.1
Hydrogen
2.223
Sulfur
5.421
100 59.1
Table 2 Energies and intensities of capture g-rays of chlorine [15] Eg (MeV) (keV)
Intensity (%)
788 1165 1950 2675 2863 3061 4979 5715 6110 6619 7413 7790
15.00 19.93 21.72 2.58 6.93 3.95 4.04 5.50 20.00 8.01 10.42 7.78
3.3. Prompt g-analysis of chlorine in concrete sample Chlorine has several capture g-rays. Their energies and intensities are listed in Table 2 [15]. Although chlorine g-ray energies below 3 MeV are listed in Table 2, but they are not used in the analysis of this study. They were used in a previous study to analyze 1.165 MeV g-ray from chlorine in the concrete [12]. Fig. 5 represents the prompt g-ray experimental pulse height spectra with energies above 3 MeV from four chloride-contaminated concrete samples, containing 0.5–3 wt% chlorine concentration and superimposed upon each other [12]. Full energy and escape peaks of chlorine are quiet prominent amongst the g-rays from the concrete constituents shown in Fig. 5. The highest chlorine peak spectrum corresponds to 3 wt% chlorine concentration, while the lowest peak spectrum corresponds to 0.5 wt%, next-to-the highest peak spectrum corresponds to 2 wt% chlorine concentration, while next-to-the lowest spectrum corresponds to 1.5 wt% chlorine concentration. The main feature of Fig. 5 is the increased intensities of some peaks due to constructive interference. The 6.110 and 6.619 MeV g-rays from chlorine have strong interference due to the full energy and single escape peaks of the 6.42 MeV g-ray from calcium in concrete. The single escape
ARTICLE IN PRESS A.A. Naqvi et al. / Nuclear Instruments and Methods in Physics Research A 569 (2006) 803–809
Si(FE), 3.54
Si(SE), 4.94 Si(FE), 4.94
Electronics Peak
COUNTS
20000
90000
Ca(FE), 6.42
15000 CI(FE),7.41
10000
Exp. Yield -6.110 MeV
Gamma Ray Experimental Yield
25000
807
5000
Monte Carlo fit-6.110
60000
30000
0 CI(FE), 5.72
CI(FE), 6.11
0
CI(FE), 6.62
0.5
1
1.5
2
2.5
3
3.5
Chlorine Concentration (wt. %) 0 4.55
5.35 6.15 Energy (MeV)
6.95
7.75
Fig. 5. Prompt g-rays spectra of the concrete samples above 3 MeV energy, containing 0.5, 1.5, 2.0 and 3 wt% chlorine. The chlorine peak intensity increases with chlorine concentration, top most spectra for 3 wt% chlorine concentration while bottom most spectrum for 0.5 wt% chlorine concentration [12].
peak of the 6.619 MeV chlorine peak interferes with full energy peak of the 6.110 MeV chlorine peak. This resulted in an increased intensity of the 6.11 MeV peak, as shown in Fig. 5. The 6.619 MeV g-ray peak from chlorine is located between the electronics sum peak and the full energy peak of the 6.42 MeV g-ray from calcium. The full energy peak of the 6.11 MeV g-ray interferes with the escape peak of the 6.42 MeV g-ray from calcium. However, the 7.413 MeV g-ray peak from chlorine is located in an isolated region and this peak was used to extract the chlorine concentration information, described earlier [12]. In order to obtain net counts under each peak, spectra were analyzed using the XSYS based data analysis system of the KFUPM 350 keV accelerator [9]. Background was defined around a peak by choosing two gates on either side of the peak of interest. The integral counts of the two gates were used to interpolate a pre-selected linear or polynomial background under the peak. Then net counts were determined from the difference between the integrated total counts under the peak counts and background counts under the peak determined by the pre-chosen background fit under the peak. Fig. 6 shows the experimental yield (net counts) of the 6.110 MeV prompt g- from chlorine in the five concrete samples as a function of the chlorine concentration. Within the experimental uncertainties, there was a linear correlation between the 6.110 MeV prompt g-ray yield and chlorine concentrations. The solid line in Fig. 6 shows the normalized-calculated yield of the 6.110 MeV prompt g-ray from chlorine obtained through the Monte Carlo simulations. The 5.715 MeV prompt g-ray from chlorine interferes with the double escape peak of 6.42 MeV g-ray from
Fig. 6. Experimental yield of 6.110 MeV prompt g-rays from chlorine in the five concrete samples plotted as a function of chlorine concentration.
60000 Gamma Ray Experimental Yield
3.75
Exp. Yield -5.715 MeV Monte Carlo fit-5.715 MeV
40000
20000
0 0
0.5
1 1.5 2 2.5 Chlorine Concentration (wt.%)
3
3.5
Fig. 7. Experimental yield of 5.715 MeV prompt g-rays from chlorine in the five concrete samples plotted as a function of chlorine concentration.
calcium, resulting in a broad peak. Fig. 7 shows the experimental yield of the 5.715 MeV prompt g-ray from chlorine in the five concrete samples as a function of the chlorine concentration. Within the experimental uncertainties, there is a linear correlation between the 5.715 MeV prompt g-ray yield and chlorine concentrations. The solid line in Fig. 7 shows the calculated yield of the 5.715 MeV prompt g-rays from chlorine obtained through the Monte Carlo simulations. 3.4. Minimum detection limit of chlorine and sulfur in concrete samples Finally, the minimum detection limit of chlorine and sulfur concentration in concrete MDC was calculated for the KFUPM PGNAA setup using the procedure described in Ref. [16]. The detection limit for an elemental concentration MDC measured under a peak with net
ARTICLE IN PRESS A.A. Naqvi et al. / Nuclear Instruments and Methods in Physics Research A 569 (2006) 803–809
808
counts P and associated background counts B (under the peak) is defined by [16] pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi MDC ¼ 3:29 Cf ½ð1 þ ZP =ZB Þ= t0 ðP=BÞðP=tÞg (1) where C is the element’s concentration in the peak, t0 is the counting time, P/t is net count rate, and ZP and ZB are the number of channels used to integrate the peak and background areas to calculate P and B counts. If ZP and ZB are equal and t0 and t are equal then the equation reduces to: pffiffiffiffi MDC ¼ 4:653 ðC=PÞ B (2) where C/P is concentration (wt%)/counts, i.e. the calibration constant of the setup for a specific g-ray peak. This is the Curie equation of Minimum Detection Limit (MDL) of counts given by Knoll [14], with counts converted into element concentration. The error in MDC i.e., pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi sMDC ¼ ðC=PÞ ½ ð2 BÞ. Table 3 shows the Minimum Detectable Concentration (MDC) limit of chlorine and sulfur in concrete samples calculated for the KFUPM PGNAA setup. For chlorine, MDC has been calculated for 1.165 and 7.413 MeV chlorine g-rays data published earlier [12]. The MDC of sulfur in concrete for the KFUPM setup using 5.421 MeV g-rays was calculated to be 0.6007 0.187 wt%. The minimum limit of sulfur in concrete set by British Standards is 0.60 wt% [4]. It shows that despite the small neutron intensity of 5 105 n/s of the KFUPM accelerator, the KFUPM PGNAA setup can detect sulfur in concrete within the limit set by the British Standards [4]. The MDC of chlorine in concrete for the KFUPM PGNAA setup was calculated using 1.165, 5.715, 6.111 and 7.413 MeV g-rays from chlorine in concrete. The MDC limit of chlorine in concrete obtained for less interfering 1.165 and 7.413 MeV g-rays is better than that obtained for strong interfering 5.715 and 6.111 MeV g-rays. For the 1.165 and 7.413 MeV g-rays from chlorine, the MDC limit was calculated to be 0.07570.025 and 0.07370.025 wt% respectively while for the 6.111 and 5.72 MeV g-rays, the MDC limit was higher and was found to be 0.1470.07 and 0.2670.05 wt%, respectively. The maximum permissible limit of chlorine concentration in concrete set by the American Concrete Institute is 0.03 wt%. The MDC limit of chlorine in the concrete for
the KFUPM neutron source strength of 3 105 n/s [11] was found to be 0.07370.025 wt%. Within a statistical uncertainty of 70.025 wt%, the MDC lower bound of 0.048 wt% is 60% higher than the maximum permissible limit of 0.03 wt% of chlorine set by ACI Standards [2]. If the KFUPM neutron source intensity is increased by a factor of 100, MDC limit of chlorine can be improved approximately by a factor of 10 [16]. It is desired to compare the MDC limits of chlorine and sulfur in concrete for radioisotope source-based PGNAA setups [5,6] and the KFUPM accelerator-based PGNAA setup. The MDC limit of chlorine in concrete of the KFUPM PGNAA setup is 30% lower than the value (0.1070.20 wt%) reported by Savio et al. [6] for an 241 Am–Be source with 3 Ci activity (3 107 n/s source strength). The MDC limit of chlorine for the KFUPM PGNAA setup is about two times larger than the value of 0.03 wt% reported by Saleh and Livingston [5] for a 5 Ci 241 Am–Be source (5 107 n/s source intensity). Although the KFUPM neutron source has 200 times smaller intensity than that used by Saleh and Livingston [5], the MDC limit of chlorine in concrete for the KFUPM setup is larger by a factor of two only. It can be concluded that after normalization of neutron source intensities amongst various radioisotope-based and accelerator-based PGNAA setups [5,6], KFUPM accelerator-based PGNAA setup has better performance, than the reported performance of the radioisotope source-based PGNAA setups [5,6]. 4. Conclusion Sulfur and chlorine concentrations in concrete samples were analyzed using the KFUPM accelerator-based PGNAA facility. The MDC limit of sulfur in concrete for the KFUPM accelerator-based PGNAA facility was found to be 0.6070.19 wt%. Similarly, the MDC limit of chlorine in concrete for the KFUPM PGNAA facility was found to be 0.07370.023 wt%. This study has demonstrated that chlorine analysis in concrete can be carried out using an accelerator-based PGNAA setup. With the accelerator-based PGNAA setups, better MDC limits of sulfur and chlorine in concrete can be achieved than the radioisotope-based PGNAA setup utilizing the neutron sources of the same strengths. Acknowledgement
Table 3 Minimum Detection Limit (MDL) of the KFUPM PGNAA setup for chlorine and sulfur in concrete samples Element
Eg (MeV)
MDL (wt%)
Limits in concrete (wt%)
Sulfur Cl Cl Cl Cl
5.421 1.165 7.413 6.111 5.72
0.60070.187 0.07770.025 0.07570.023 0.14070.068 0.25570.050
0.60 0.03 0.03 0.03 0.03
The authors acknowledge the support of King Fahd University of Petroleum and Minerals, Dhahran, Saudi Arabia, for this study through Project No. SABIC/2002-03. References [1] Portland Cement Association website: www.cement.org. [2] Protection of Metals in Concrete Against Corrosion, Report ACI222R-01, page 222R-11, American Concrete Institute Manual
ARTICLE IN PRESS A.A. Naqvi et al. / Nuclear Instruments and Methods in Physics Research A 569 (2006) 803–809
[3]
[4] [5]
[6]
[7]
[8]
of Concrete Practise 2004(MCP2004), CD-ROM edition (Farmington Hill, Michgan, USA: American Concrete Institute, 2004). O.S.B. Al-Amoudi, M. Maslehuddin, The effect of chloride and sulfate ions on reinforcement corrosion, Cement Concrete Res 23 (1993) 139. British Standards : BS 1881 Part 124. H.H. Saleh, R.A. Livingston, Experimental evaluation of a portable neutron-based gamma-spectroscopy system for chloride measurements in reinforced concretem, J. Radioanal. Nucl. Chem. 244 (2000) 367. D.L.C. Savio, M.A.J. Mariscotti, S.R. Guevara, Elemental analysis of a concrete sample by capture gamma rays with a radioisotope neutron source, Nucl. Instr. and Meth. 95 (1995) 379. R. Khelifi, Z. Idiri, L. Omari, M. Seghir, Prompt gamma neutron activation analysis of bulk concrete samples with an am-be neutron sources, Appl. Radiat. Isot. 51 (1999) 9. C.S. Lim, J.R. Tickner, B.D. Sowerby, D.A. Abernethy, A.J. McEwan, S. Rainey, R. Stevans, C. Manias, D. Retallack, An onbelt elemental analyzer for the cement industry, Appl. Radiat. Isot. 54 (2001) 11.
809
[9] A.A. Naqvi, M.M. Nagadi, Khateeb-ur-Rehman, M. Maslehuddin, S. Kidwai, Monte Carlo simulations for the design of the KFUPM PGNAA facility, Rad. Phys. Chem. 66 (2003) 89. [10] A.A. Naqvi, Fazul-ur-Rehman, M.I. Al-Jarallah, F. Abujarad, M. Maslehuddin, Performance tests of external moderators of a PGNAA setup, Appl. Radiat. Isot. 58 (2003) 27. [11] A.A. Naqvi, M.M. Nagadi, O.S. B Al-Amoudi, Elemental analysis of concrete samples using an accelerator-based PGNAA setup, Nucl. Instr. and Meth. 225 (2004) 331. [12] A.A. Naqvi, M.M. Nagadi, O.S. Baghaabra Al-Amoudi, Prompt gamma analysis of chlorine in concrete for corrosion study, Appl. Radiat. Isot. 64 (2005) 283. [13] Thermo Neutron Generators; Thermo Electron Corporation, 81 Wyman Street, Waltham, MA 02454; USA; website: www.thermo.com. [14] G.F. Knoll. Radiation Detection and Measurements, 2000. [15] M.A. Lone, R.A. Leavitt, D.A. Harrison, Prompt gamma rays from thermal neutron capture, Atomic Data Nucl. Data Tables 26 (1981) 511. [16] D.A. Gedcke, How counting statistics controls detection limits and peak precession. ORTEC Application Notes AN59.website: www. ortec-online.com.
Nuclear Instruments and Methods in Physics Research A 569 (2006) 803–809 www.elsevier.com/locate/nima
Interferences in Prompt g Analysis of corrosive contaminants in concrete A.A. Naqvia,, M.M. Nagadia, O.S.B. Al-Amoudib a
Department of Physics, King Fahd University of Petroleum and Minerals, KFUPM Box 1815, Dhahran-31261, Saudi Arabia b Department of Civil Engineering, King Fahd University of Petroleum and Minerals, Dhahran, Saudi Arabia Received 12 May 2006; received in revised form 11 September 2006; accepted 13 September 2006 Available online 13 October 2006
Abstract An accelerator-based Prompt Gamma Neutron Activation Analysis (PGNAA) setup has been developed to measure the concentration of corrosive chloride and sulfate contaminants in concrete. The Minimum Detectable Concentration (MDC) limit of chlorine and sulfur in the concrete depends upon the g-ray used for elemental analysis. For more interfering g-rays, the MDC limit is higher than that for less interfering g-rays. The MDC limit of sulfur in concrete measured for the KFUPM PGNAA setup was calculated to be 0.6070.19 wt%. The MDC limit is equal to the upper limit of sulfur concentration in concrete set by the British Standards. The MDC limit of chlorine in concrete for the KFUPM PGNAA setup, which was calculated for less interfering 1.165 MeV g-rays, was found to be 0.07570.025 wt%. The lower limits of the MDC of chlorine in concrete was 73% higher than the limit set by American Concrete Institute. The limit of the MDC can be improved to the desired standard by increasing the intensity of neutron source. For moreinterfering 5.715 and 6.110 MeV chlorine g-rays the MDC limit was found to be 2–3 times larger than that of 1.165 MeV g-rays. When normalized to the same intensity of the neutron source, the MDC limits of chlorine and sulfur in concrete from the KFUPM PGNAA setup are better than MDC limits of chlorine in concrete obtained with the 241Am–Be source-based PGNAA setup. This study has shown that an accelerator-based PGNAA setup can be used in chlorine and sulfur analysis of concrete samples. r 2006 Elsevier B.V. All rights reserved. PACS: 82.80.Jp Keywords: PGNAA study; Concrete samples; Sulfur and chlorine measurement; 2.8 MeV neutrons; Monte Carlo simulations; Corrosion studies
1. Introduction The corrosion of reinforcing steel in concrete structures is a major problem faced by the building and construction industry [1]. The chloride and sulfate salt present in concrete initiate reinforcing steel corrosion [1–3]. Chloride-induced reinforcement corrosion is the principal cause of deterioration in concrete structures in the Arabian Gulf and the world over. In view of this, the American Concrete Standards Codes specify the allowable chloride concentrations in concrete as 0.03% by weight of concrete (or 0.15% by weight of cement). [2]. Similarly, the allowable limit of sulfate concentration in concrete is Corresponding author. Tel.: +966 3860 4196; fax: +966 3860 2293.
E-mail address: [email protected] (A.A. Naqvi). 0168-9002/$ - see front matter r 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.nima.2006.09.058
0.60% by weight of concrete (or 4.0% by weight of sulfur) [4]. These limits are being utilized by the construction industry as the corrosion–threshold concentrations in concrete for chloride and sulfate ions. Preventive measures against corrosion require maintaining the chloride and sulfur concentration in concrete below the threshold limits [2]. This requires monitoring the chloride concentration in concrete using a non-destructive technique. The Prompt Gamma Neutron Activation Analysis (PGNAA) technique can be used to monitor concentration of corrosive elements in concrete samples [5–12]. This technique has been successfully used to analyze the elemental composition of concrete samples using 241Am–Be radioisotope neutron sources [5–7]. The disadvantages of a radioisotope neutron source-based PGNAA setup are high g-ray dose, permanent
ARTICLE IN PRESS 804
A.A. Naqvi et al. / Nuclear Instruments and Methods in Physics Research A 569 (2006) 803–809
radiation hazards and limited life of the neutron source due to the short half-life of the radioisotopes source. It is desired to develop an accelerator-based PGNAA setup to monitor the chloride and sulfur concentration in concrete. The accelerator-based neutron source has several advantages such as controlled mechanism of neutron production and less radiation hazard [9–12]. The On–Off control of accelerator power ensures that the acceleratorbased PGNAA setup has radiation hazards only when in use. Furthermore, commercially manufactured compact accelerators have the physical dimensions almost comparable to with that of the radioisotope neutron source. Also the neutron flux available with a compact neutron generator is greater than that available with a commercial radioisotope source. A compact neutron generator can produce 3 108 neutrons/s or more [13] while a 5 Ci 241 Am–Be can produce 3 107 neutrons/s [14]. In short, an accelerator-based PGNAA setup is better, or at the least comparable to a radio-isotope neutron source-based PGNAA analyzer. In this regard, several studies have been undertaken at the King Fahd University of Petroleum and Minerals (KFUPM) 350 keV accelerator laboratory to develop an accelerator-based PGNAA setup. The setups were used to analyze concrete and cement samples [9–12]. An accelerator-based PGNAA facility was designed to analyze large samples of concrete [11–12]. The KFUPM accelerator, which is a 350 kV electrostatic accelerator, produces 2.8 MeV neutrons beam via D(d,n) reaction. Although the KFUPM accelerator has a limited source intensity of 3 105 n/s [11], the chlorine concentration in concrete was successfully measured using a large sample [12]. The MDC limit of a PGNAA setup depends upon several factors such as neutron source intensity, g-ray background under the peak etc. The g-ray background depends upon the g-ray used to characterize the element in
PGNAA analysis. For the interfering g-rays, the g-ray background is larger. In this study a prompt g-ray analysis has been carried out on corrosive chlorine and sulfur contaminants in concrete. Various g-ray lines from the corrosive contaminants have been studied to determine the chlorine and sulfur MDC limits in concrete. There are strong interferences between the g-rays from chlorine, sulfur and the concrete constituents. The prompt g-ray yield from sulfur has been measured using the KFUPM accelerator-based PGNAA setup, and sulfur concentration in concrete has been determined. Also, chlorine concentration in concrete has been analyzed for previously taken data [12] using strongly interfering 5.715 and 6.110 MeV chlorine g-rays. Earlier, the chlorine concentration in the concrete samples was analyzed using less interfering 1.165 and 7.431 MeV prompt g-rays [12]. The results of the recent g-ray analysis of sulfur and chlorine corrosive contaminants in concrete are described below.
2. Prompt c-ray yield measurements from sulfur The prompt g-ray yield from sulfur has been measured using the KFUPM accelerator-based PGNAA setup shown in Fig. 1. The setup has been reported earlier in detail elsewhere [11,12]; however, for sake of continuity it will be briefly described here. The setup mainly consists of a cylindrical sample enclosed in a rectangular paraffin moderator. The moderator is placed between a neutron target and a cylindrical NaI g-ray detector having a dimension of 25.5 25.5 cm (diameter height) [11]. The 2.8 MeV neutrons from the D(d, n) reaction are thermalized in the moderator and are captured in the concrete sample. The NaI g-ray detector measured the experimental yield of prompt g-rays from the capture of thermal neutrons in the concrete samples.
Fig. 1. Schematic representation of the PGNAA setup used to measure prompt g-ray yield [12].
ARTICLE IN PRESS A.A. Naqvi et al. / Nuclear Instruments and Methods in Physics Research A 569 (2006) 803–809
3. Results and discussion 3.1. Prompt g analysis of sulfur in concrete samples Fig. 2 shows the prompt g-ray pulse height spectrum of three concrete samples containing sulfur in different weight proportions. An important feature of Fig. 2 is the presence of a peak of 5.421 MeV prompt g-ray due to the capture of thermal neutrons in sulfur present in the concrete samples. The highest peak of sulfur in Fig. 2 corresponds to a 6 wt% concentration, while the smallest peak corresponds to 2 wt% and the middle one corresponds to 4 wt%. The 5.421 MeV g-ray peak is located between the channels 550 and 600, next to the single escape peak Ca(S) from calcium in concrete. The location of the sulfur 5.421 MeV peak coincides with the double escape peak of calcium but the intensity of the double escape peak from calcium is insignificant (typically less than 0.5% of the full energy peak). Therefore, the background under the sulfur peak, due to the double escape peak of calcium peak is insignificant. This is clearly shown in Fig. 2, where height
12000
Si(FE), 4.94
Ca(SE), 6.42
Ca(FE), 6.42
10000
COUNTS
8000
6000
S(FE), 5.42
4000
2000
4.75
5.75
6.75 Energy (MeV)
7.75
Fig. 2. Prompt g-rays spectra of the concrete samples above 3 MeV energy, containing 2.0, 4.0 and 6.0 wt% sulfur. The sulfur peak intensity increases with sulfur concentration, top most spectrum for 6 wt% sulfur concentration, middle one for 4 wt% and bottom most spectrum for 2.0 wt% sulfur concentration.
40000 5.42 MeV Gamma Ray Experimental Yield
The sulfur contaminated concrete samples used in the study were 14 cm long and 12.5 cm in radius. The procedure used to calculate their size, using the Monte Carlo simulations, is described elsewhere [11,12], The front moderator thickness used here was 5 cm. The sulfurcontaminated concrete samples were prepared by mixing the sulfur powder with the concrete mix in 2, 4, 6 and 8 wt% proportions. The concrete mix was prepared by mixing the concrete ingredients in the following ratio: Portland cement (370 kg), coarse aggregate (1110 kg); fine aggregate (680.4 kg) and water (198 kg) [11]. Then the concrete ingredients and the sulfur powder were mixed together and cast in specially designed 14 cm long plastic moulds with a radius of 11.5 cm. After demoulding, the concrete samples were cured in water for 7 days and then air dried in the laboratory environment (22721 C) for 7 more days. Thereafter, they were dried in an electric oven at 701 C temperatures for about 3 days until a constant weight was obtained [11]. The sulfur contaminated concrete samples were irradiated in the KFUPM PGNAA setup with the rectangular moderator [12]. The g-ray yield from the sulfur mixed concrete samples was measured using pulsed beam of 2.8 MeV neutrons produced via D(d, n) reaction. The deuteron pulse has a width of 5–6 ns and a repetition rate of 31.25 kHz [12]. The experiment was conducted with a typical beam current of 3.5–4.8 mA. The prompt g-rays produced in the concrete samples were detected by a 25 25 cm (diameter length) NaI detector. The prompt g-ray data were acquired for a preset number of charges measured at the electrically isolated neutron-producing target. The prompt g-ray detector signals were acquired in coincidence with a gate signal being derived from the beam pick-up signal.
805
Exp. Yield-5.42 MeV Monte Carlo fit-5.42 MeV
30000
20000
10000
0 0
2
4
6
8
10
Sulfur concentration (wt. %) Fig. 3. Experimental yield of 5.42 MeV prompt g-rays from sulfur in the four concrete samples plotted as a function of sulfur concentration.
of the sulfur peak increases with increasing sulfur concentration in concrete, over 2–6 wt% range. The net counts under each peak of 5.421 MeV g-ray were obtained by subtracting background from the sulfur gamma peak in the pulse height spectra of all four samples. Counts under each peak were integrated and normalized to the same amount of integrated beam charge measured at the target. The experimental yield of 5.421 MeV prompt g-ray from sulfur in the four concrete samples is plotted in Fig. 3 as a function of the sulfur concentration. Within the
ARTICLE IN PRESS A.A. Naqvi et al. / Nuclear Instruments and Methods in Physics Research A 569 (2006) 803–809
806
experimental uncertainties, there is a linear correlation between the 5.417 MeV prompt g-ray yield and the sulfur concentration. Solid lines in Fig. 3 represent the calculated yield of the 5.417 MeV prompt g-ray from the sulfur obtained through the Monte Carlo simulations. Consequently there exists an excellent agreement between the experimental and the calculated yield of 5.421 MeV sulfur g-rays from the four samples. 3.2. Prompt g-analysis of pure concrete samples For a better understanding of the g-ray spectrum of the pure concrete sample, it may be worthwhile to discuss the mutual interference amongst the g-rays emitted by the concrete constituents. The data is taken from ref. [12]. Fig. 4 shows an experimentally measured g-ray spectrum of a pure concrete sample above 3 MeV showing full energy and single escape peaks of calcium and silicon [12]. Also superimposed on the spectrum is the g-ray background spectrum (lower spectrum) taken without the sample. As shown in Table 1, there is interference between the g-rays from calcium and silicon in concrete. To be consistent with experimental data, where only g-rays with energies in excess of 3 MeV were analyzed, we will focus upon those prompt g-rays with energies above 3 MeV. The calcium has two g-rays with energies above 3.0 MeV, namely 4.42 MeV (14.9%) and 6.42 MeV (38.9%) energies and intensities respectively. Silicon has two prompt g-rays with energies 3.54 MeV (68%) and 4.94 MeV (62.7%). The full energy peak of 4.42 MeV (14.9%) from calcium interferes with the single escape peak of 4.94 MeV (62.7%) g-ray from silicon and its intensity is added under the peak resulting in the higher intensity of the single escape peak than the corresponding full energy peak of 4.94 MeV g-ray from silicon, as shown in Fig. 4.
14000 Si(FE), 3.54
Si(SE), 4.94 Si(FE), 4.94 Ca(SE), 6.42 Ca(FE), 6.42
12000
Electronics sum Peak
COUNTS
10000 8000 6000 4000 2000 0 3.09
4.83 6.56 Energy (MeV)
8.30
Fig. 4. Experimentally measured prompt g-ray spectrum of the concrete sample above 3 MeV showing full energy and single escape peaks of calcium and silicon. Also superimposed on the spectrum is the g-ray background spectrum (lower spectrum) taken without the sample [12].
Table 1 Energies and intensities of capture g-rays of concrete [15] Element
g-rays energy (MeV)
Intensity (%)
Calcium
1.942 4.418 6.420
72.5 14.9 38.9
Silicon
3.539 4.934
68 62.7
Aluminum
7.724
27.4
Iron
7.631 7.646
28.5 24.1
Hydrogen
2.223
Sulfur
5.421
100 59.1
Table 2 Energies and intensities of capture g-rays of chlorine [15] Eg (MeV) (keV)
Intensity (%)
788 1165 1950 2675 2863 3061 4979 5715 6110 6619 7413 7790
15.00 19.93 21.72 2.58 6.93 3.95 4.04 5.50 20.00 8.01 10.42 7.78
3.3. Prompt g-analysis of chlorine in concrete sample Chlorine has several capture g-rays. Their energies and intensities are listed in Table 2 [15]. Although chlorine g-ray energies below 3 MeV are listed in Table 2, but they are not used in the analysis of this study. They were used in a previous study to analyze 1.165 MeV g-ray from chlorine in the concrete [12]. Fig. 5 represents the prompt g-ray experimental pulse height spectra with energies above 3 MeV from four chloride-contaminated concrete samples, containing 0.5–3 wt% chlorine concentration and superimposed upon each other [12]. Full energy and escape peaks of chlorine are quiet prominent amongst the g-rays from the concrete constituents shown in Fig. 5. The highest chlorine peak spectrum corresponds to 3 wt% chlorine concentration, while the lowest peak spectrum corresponds to 0.5 wt%, next-to-the highest peak spectrum corresponds to 2 wt% chlorine concentration, while next-to-the lowest spectrum corresponds to 1.5 wt% chlorine concentration. The main feature of Fig. 5 is the increased intensities of some peaks due to constructive interference. The 6.110 and 6.619 MeV g-rays from chlorine have strong interference due to the full energy and single escape peaks of the 6.42 MeV g-ray from calcium in concrete. The single escape
ARTICLE IN PRESS A.A. Naqvi et al. / Nuclear Instruments and Methods in Physics Research A 569 (2006) 803–809
Si(FE), 3.54
Si(SE), 4.94 Si(FE), 4.94
Electronics Peak
COUNTS
20000
90000
Ca(FE), 6.42
15000 CI(FE),7.41
10000
Exp. Yield -6.110 MeV
Gamma Ray Experimental Yield
25000
807
5000
Monte Carlo fit-6.110
60000
30000
0 CI(FE), 5.72
CI(FE), 6.11
0
CI(FE), 6.62
0.5
1
1.5
2
2.5
3
3.5
Chlorine Concentration (wt. %) 0 4.55
5.35 6.15 Energy (MeV)
6.95
7.75
Fig. 5. Prompt g-rays spectra of the concrete samples above 3 MeV energy, containing 0.5, 1.5, 2.0 and 3 wt% chlorine. The chlorine peak intensity increases with chlorine concentration, top most spectra for 3 wt% chlorine concentration while bottom most spectrum for 0.5 wt% chlorine concentration [12].
peak of the 6.619 MeV chlorine peak interferes with full energy peak of the 6.110 MeV chlorine peak. This resulted in an increased intensity of the 6.11 MeV peak, as shown in Fig. 5. The 6.619 MeV g-ray peak from chlorine is located between the electronics sum peak and the full energy peak of the 6.42 MeV g-ray from calcium. The full energy peak of the 6.11 MeV g-ray interferes with the escape peak of the 6.42 MeV g-ray from calcium. However, the 7.413 MeV g-ray peak from chlorine is located in an isolated region and this peak was used to extract the chlorine concentration information, described earlier [12]. In order to obtain net counts under each peak, spectra were analyzed using the XSYS based data analysis system of the KFUPM 350 keV accelerator [9]. Background was defined around a peak by choosing two gates on either side of the peak of interest. The integral counts of the two gates were used to interpolate a pre-selected linear or polynomial background under the peak. Then net counts were determined from the difference between the integrated total counts under the peak counts and background counts under the peak determined by the pre-chosen background fit under the peak. Fig. 6 shows the experimental yield (net counts) of the 6.110 MeV prompt g- from chlorine in the five concrete samples as a function of the chlorine concentration. Within the experimental uncertainties, there was a linear correlation between the 6.110 MeV prompt g-ray yield and chlorine concentrations. The solid line in Fig. 6 shows the normalized-calculated yield of the 6.110 MeV prompt g-ray from chlorine obtained through the Monte Carlo simulations. The 5.715 MeV prompt g-ray from chlorine interferes with the double escape peak of 6.42 MeV g-ray from
Fig. 6. Experimental yield of 6.110 MeV prompt g-rays from chlorine in the five concrete samples plotted as a function of chlorine concentration.
60000 Gamma Ray Experimental Yield
3.75
Exp. Yield -5.715 MeV Monte Carlo fit-5.715 MeV
40000
20000
0 0
0.5
1 1.5 2 2.5 Chlorine Concentration (wt.%)
3
3.5
Fig. 7. Experimental yield of 5.715 MeV prompt g-rays from chlorine in the five concrete samples plotted as a function of chlorine concentration.
calcium, resulting in a broad peak. Fig. 7 shows the experimental yield of the 5.715 MeV prompt g-ray from chlorine in the five concrete samples as a function of the chlorine concentration. Within the experimental uncertainties, there is a linear correlation between the 5.715 MeV prompt g-ray yield and chlorine concentrations. The solid line in Fig. 7 shows the calculated yield of the 5.715 MeV prompt g-rays from chlorine obtained through the Monte Carlo simulations. 3.4. Minimum detection limit of chlorine and sulfur in concrete samples Finally, the minimum detection limit of chlorine and sulfur concentration in concrete MDC was calculated for the KFUPM PGNAA setup using the procedure described in Ref. [16]. The detection limit for an elemental concentration MDC measured under a peak with net
ARTICLE IN PRESS A.A. Naqvi et al. / Nuclear Instruments and Methods in Physics Research A 569 (2006) 803–809
808
counts P and associated background counts B (under the peak) is defined by [16] pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi MDC ¼ 3:29 Cf ½ð1 þ ZP =ZB Þ= t0 ðP=BÞðP=tÞg (1) where C is the element’s concentration in the peak, t0 is the counting time, P/t is net count rate, and ZP and ZB are the number of channels used to integrate the peak and background areas to calculate P and B counts. If ZP and ZB are equal and t0 and t are equal then the equation reduces to: pffiffiffiffi MDC ¼ 4:653 ðC=PÞ B (2) where C/P is concentration (wt%)/counts, i.e. the calibration constant of the setup for a specific g-ray peak. This is the Curie equation of Minimum Detection Limit (MDL) of counts given by Knoll [14], with counts converted into element concentration. The error in MDC i.e., pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi sMDC ¼ ðC=PÞ ½ ð2 BÞ. Table 3 shows the Minimum Detectable Concentration (MDC) limit of chlorine and sulfur in concrete samples calculated for the KFUPM PGNAA setup. For chlorine, MDC has been calculated for 1.165 and 7.413 MeV chlorine g-rays data published earlier [12]. The MDC of sulfur in concrete for the KFUPM setup using 5.421 MeV g-rays was calculated to be 0.6007 0.187 wt%. The minimum limit of sulfur in concrete set by British Standards is 0.60 wt% [4]. It shows that despite the small neutron intensity of 5 105 n/s of the KFUPM accelerator, the KFUPM PGNAA setup can detect sulfur in concrete within the limit set by the British Standards [4]. The MDC of chlorine in concrete for the KFUPM PGNAA setup was calculated using 1.165, 5.715, 6.111 and 7.413 MeV g-rays from chlorine in concrete. The MDC limit of chlorine in concrete obtained for less interfering 1.165 and 7.413 MeV g-rays is better than that obtained for strong interfering 5.715 and 6.111 MeV g-rays. For the 1.165 and 7.413 MeV g-rays from chlorine, the MDC limit was calculated to be 0.07570.025 and 0.07370.025 wt% respectively while for the 6.111 and 5.72 MeV g-rays, the MDC limit was higher and was found to be 0.1470.07 and 0.2670.05 wt%, respectively. The maximum permissible limit of chlorine concentration in concrete set by the American Concrete Institute is 0.03 wt%. The MDC limit of chlorine in the concrete for
the KFUPM neutron source strength of 3 105 n/s [11] was found to be 0.07370.025 wt%. Within a statistical uncertainty of 70.025 wt%, the MDC lower bound of 0.048 wt% is 60% higher than the maximum permissible limit of 0.03 wt% of chlorine set by ACI Standards [2]. If the KFUPM neutron source intensity is increased by a factor of 100, MDC limit of chlorine can be improved approximately by a factor of 10 [16]. It is desired to compare the MDC limits of chlorine and sulfur in concrete for radioisotope source-based PGNAA setups [5,6] and the KFUPM accelerator-based PGNAA setup. The MDC limit of chlorine in concrete of the KFUPM PGNAA setup is 30% lower than the value (0.1070.20 wt%) reported by Savio et al. [6] for an 241 Am–Be source with 3 Ci activity (3 107 n/s source strength). The MDC limit of chlorine for the KFUPM PGNAA setup is about two times larger than the value of 0.03 wt% reported by Saleh and Livingston [5] for a 5 Ci 241 Am–Be source (5 107 n/s source intensity). Although the KFUPM neutron source has 200 times smaller intensity than that used by Saleh and Livingston [5], the MDC limit of chlorine in concrete for the KFUPM setup is larger by a factor of two only. It can be concluded that after normalization of neutron source intensities amongst various radioisotope-based and accelerator-based PGNAA setups [5,6], KFUPM accelerator-based PGNAA setup has better performance, than the reported performance of the radioisotope source-based PGNAA setups [5,6]. 4. Conclusion Sulfur and chlorine concentrations in concrete samples were analyzed using the KFUPM accelerator-based PGNAA facility. The MDC limit of sulfur in concrete for the KFUPM accelerator-based PGNAA facility was found to be 0.6070.19 wt%. Similarly, the MDC limit of chlorine in concrete for the KFUPM PGNAA facility was found to be 0.07370.023 wt%. This study has demonstrated that chlorine analysis in concrete can be carried out using an accelerator-based PGNAA setup. With the accelerator-based PGNAA setups, better MDC limits of sulfur and chlorine in concrete can be achieved than the radioisotope-based PGNAA setup utilizing the neutron sources of the same strengths. Acknowledgement
Table 3 Minimum Detection Limit (MDL) of the KFUPM PGNAA setup for chlorine and sulfur in concrete samples Element
Eg (MeV)
MDL (wt%)
Limits in concrete (wt%)
Sulfur Cl Cl Cl Cl
5.421 1.165 7.413 6.111 5.72
0.60070.187 0.07770.025 0.07570.023 0.14070.068 0.25570.050
0.60 0.03 0.03 0.03 0.03
The authors acknowledge the support of King Fahd University of Petroleum and Minerals, Dhahran, Saudi Arabia, for this study through Project No. SABIC/2002-03. References [1] Portland Cement Association website: www.cement.org. [2] Protection of Metals in Concrete Against Corrosion, Report ACI222R-01, page 222R-11, American Concrete Institute Manual
ARTICLE IN PRESS A.A. Naqvi et al. / Nuclear Instruments and Methods in Physics Research A 569 (2006) 803–809
[3]
[4] [5]
[6]
[7]
[8]
of Concrete Practise 2004(MCP2004), CD-ROM edition (Farmington Hill, Michgan, USA: American Concrete Institute, 2004). O.S.B. Al-Amoudi, M. Maslehuddin, The effect of chloride and sulfate ions on reinforcement corrosion, Cement Concrete Res 23 (1993) 139. British Standards : BS 1881 Part 124. H.H. Saleh, R.A. Livingston, Experimental evaluation of a portable neutron-based gamma-spectroscopy system for chloride measurements in reinforced concretem, J. Radioanal. Nucl. Chem. 244 (2000) 367. D.L.C. Savio, M.A.J. Mariscotti, S.R. Guevara, Elemental analysis of a concrete sample by capture gamma rays with a radioisotope neutron source, Nucl. Instr. and Meth. 95 (1995) 379. R. Khelifi, Z. Idiri, L. Omari, M. Seghir, Prompt gamma neutron activation analysis of bulk concrete samples with an am-be neutron sources, Appl. Radiat. Isot. 51 (1999) 9. C.S. Lim, J.R. Tickner, B.D. Sowerby, D.A. Abernethy, A.J. McEwan, S. Rainey, R. Stevans, C. Manias, D. Retallack, An onbelt elemental analyzer for the cement industry, Appl. Radiat. Isot. 54 (2001) 11.
809
[9] A.A. Naqvi, M.M. Nagadi, Khateeb-ur-Rehman, M. Maslehuddin, S. Kidwai, Monte Carlo simulations for the design of the KFUPM PGNAA facility, Rad. Phys. Chem. 66 (2003) 89. [10] A.A. Naqvi, Fazul-ur-Rehman, M.I. Al-Jarallah, F. Abujarad, M. Maslehuddin, Performance tests of external moderators of a PGNAA setup, Appl. Radiat. Isot. 58 (2003) 27. [11] A.A. Naqvi, M.M. Nagadi, O.S. B Al-Amoudi, Elemental analysis of concrete samples using an accelerator-based PGNAA setup, Nucl. Instr. and Meth. 225 (2004) 331. [12] A.A. Naqvi, M.M. Nagadi, O.S. Baghaabra Al-Amoudi, Prompt gamma analysis of chlorine in concrete for corrosion study, Appl. Radiat. Isot. 64 (2005) 283. [13] Thermo Neutron Generators; Thermo Electron Corporation, 81 Wyman Street, Waltham, MA 02454; USA; website: www.thermo.com. [14] G.F. Knoll. Radiation Detection and Measurements, 2000. [15] M.A. Lone, R.A. Leavitt, D.A. Harrison, Prompt gamma rays from thermal neutron capture, Atomic Data Nucl. Data Tables 26 (1981) 511. [16] D.A. Gedcke, How counting statistics controls detection limits and peak precession. ORTEC Application Notes AN59.website: www. ortec-online.com.