ANALYTICAL SCIENCES OCTOBER 2014, VOL. 30 1013

Introduction

Corrosion has a major impact on the efficiency and reliability of electric power plants.1 Tube corrosion is known to be the main cause of boiler failure, and at its worst, a single corrosion failure can make a plant inoperable.1,2 The total cost of boiler tube failures in power plants is estimated to be about 5 billion dollars a year,3 and accidents temporarily interrupt power supply to tens of thousands of consumers. Scale deposits in boiler systems are another main concern since they significantly decrease energy efficiency.2 For thermal power plants, the economic losses caused by a boiler failure are a priority concern. On the other hand, in the case of nuclear power plants, the safety risk of the release of dangerous radionuclides into the environment has a higher priority. The boiler water system undergoes various chemical and physical treatments to control the corrosion processes.4–6 As an indicator of scale and corrosion occurrence in a boiler, the Japan Industry Standard regulates iron concentration at less than 10 μg dm–3 for the all-volatile treatment system and lower than 5 μg dm–3 for the oxygenated treatment system.7 For the purpose of monitoring such a low concentration of iron, a sensitive detection method of which detection limit reaches the sub-μg dm–3 level is needed. The 2,4,6-tris(2-pyridyl)-1,3,5-triazine (TPTZ) solution method (JIS B 8224-2005) is not sensitive enough, therefore the concentration (e.g., 16 times) of the sample solution by heating for a long time is required, which makes the operation complicated and troublesome. To reduce the analytical time, inductively coupled

plasma atomic emission spectrometry (ICP-AES) and inductively coupled plasma mass spectrometry (ICP-MS) are used for the determination of iron. However, besides the effectiveness of the determination method, the cost for the instrument and running is also a major concern for industries.

Solid phase spectrometry (SPS) can overcome the above problems because of its high sensitivity, simple operation and low cost. SPS denotes the direct absorption measurement of solid beads on which the colored analyte was adsorbed.8 Regarding the concentration, the SPS efficiency is significantly greater than that of evaporation concentration. For example, in the solution spectrometric method, an 800-cm3 boiler water sample has to be evaporated/concentrated to 50 cm3 by heating for 4 – 5 h to obtain a measurable iron concentration.9 However, in the present SPS method for iron determination, a 0.08 cm3 cation exchanger is added to a 50-cm3 boiler water sample, and mixed for 30 min to adsorb/concentrate the colored analyte on the solid beads, resulting in a 625 times concentration of the target analyte in 30 min without any other procedure. Therefore, the sensitivity of SPS is several orders of magnitude greater than the solution method, and the procedure is much simpler. In addition, the 30-min mixing can be done in a closed container to avoid any iron contamination from the experimental environment.

In this study, TPTZ was used as the coloring reagent. TPTZ was reported to form an intense violet colored 1:2 complex (Fe(TPTZ)2

2+) with Fe(II),10 and used for the determination of the total iron in various samples.10–12 The complex has an absorption maximum at 593 nm, and is stable for at least 32 h in an aqueous solution. The molar absorptivity of the complex is known to slightly vary in different solvents, and 2.26 × 104 dm3 mol–1 cm–1 in water. The TPTZ method has the advantage of being free from common interferences.10,13 As a preliminary

2014 © The Japan Society for Analytical Chemistry

† To whom correspondence should be addressed.E-mail: sarna2010@yahoo.co.jp

Determination of Trace Iron in the Boiler Water Used in Power Generation Plants by Solid-phase Spectrophotometry

SARENQIQIGE,*† Akihiro MAEDA,** and Kazuhisa YOSHIMURA*

* Department of Chemistry, Faculty of Sciences, Kyushu University, Hakozaki, Higashi, Fukuoka 812–8581, Japan

** Energy and Environment, Mitsubishi Heavy Industries, Ltd., Wadasaki-cho, Hyogo, Kobe 652–8585, Japan

A sensitive, simple and low-cost determination method for the total iron concentration in boiler water systems of power generation plants was developed by solid phase spectrometry (SPS) using 2,4,6-tris(2-pyridyl)-1,3,5-triazine (TPTZ) as a coloring agent. The reagents and 0.08 cm3 of a cation exchanger were added to a 50-cm3 boiler water sample, then mixed for 30 min to adsorb/concentrate the produced Fe(TPTZ)2

2+ colored complex on the solid beads, resulting in a 625 times concentration of the target analyte without any other procedure. The detection limit of 0.1 μg dm–3 was obtained, and the optimum conditions for the digestion procedure and color developing reaction was investigated and reported. According to the application of this method to real samples, the present SPS method is the best one because of the shorter analysis time, simpler operation and use of very low-cost equipment compared to the conventional methods, such as TPTZ solution spectrophotometric method after a 16 times concentration, ICP-MS and AAS.

Keywords Iron determination, SPS method, boiler water, power generation plants

(Received April 30, 2014; Accepted August 4, 2014; Published October 10, 2014)

Notes

1014 ANALYTICAL SCIENCES OCTOBER 2014, VOL. 30

study, the applicability of the SPS method was reported without the optimization of the experimental conditions.9 In this paper, the optimization of the measurement experiment was reported in detail. The present SPS method significantly improved the sensitivity of the TPTZ coloring method for total iron determination, and made it possible to detect the sub-μg dm–3 level total iron in water samples without any special and expensive equipment.

Experimental

ReagentsHighly-purified water prepared using a Milli-Q SP system

(Millipore, USA) was used throughout the study. A standard Fe(III) solution (500 μg dm–3) was prepared by diluting a 1000 mg dm–3 atomic absorption standard solution (Wako, Osaka, Japan) with water and acidified to pH 2 with hydrochloric acid. The hydrochloric acid (1+1) was prepared by diluting 36% hydrochloric acid (for trace analysis, Wako) with water. The buffer solution (6 mol dm–3) was made by mixing an ammonia solution (ultrapure, 28.0%, Kanto), formic acid (for trace analysis, 98.0%, Kanto), and highly purified water in the volume ratio of 2:1:2. A coloring reagent TPTZ solution was prepared by dissolving 0.32 g TPTZ (Wako, Tokyo, Japan) in 1 cm3 (1+1) hydrochloric acid, and then diluted to 500 cm3 with water. A reducing agent (hydroxylammonium chloride) solution (10%) was made by dissolving 10 g of hydroxylammonium chloride (Wako) in water and made up to 100 cm3. The Muromac 50W-X2 cation exchanger (100 – 200 mesh) was purchased from Muromachi Technos, Tokyo, Japan.

Devices and equipmentThe equipment used for the light measurement was described

in our previous papers.14,15 A portable single-beam type spectrophotometer for the SPS, Model SP-101 (W250 × D100 × H50 mm), was purchased from Satoda Science, Hiroshima, Japan. It contained a halogen tungsten lamp, a grating (600 lines/mm) as the monochromator, a CCD array (2048 bit) as the light detector (measurable wavelength: 400 – 850 nm, resolution: 2.5 nm), and a laptop PC as the data processor. A black cell having a light-path length of 10 mm and a 2-mm width (M20-B-2, GL Sciences, Tokyo, Japan) was used.15 A constant volume of the resin beads was taken using an aliquotting device; a PTFE tube (1.0-mm i.d. and 70-mm long) was fit to the side by a PP resin filter tip and connected to a 10-cm3 disposable syringe.16,17

A heating bath (ADVENTEC) was used for preparing the boiling water for the sample digestion procedure.

Analytical procedure for measurement by solid-phase spectrophotometry

All bottles used throughout the study were allowed to soak in a surfactant bath and the sulfuric acid bath (v/v 2%) for at least 24 h to eliminate the contaminants. Preparation of all solutions was conducted in a clean bench to prevent iron contamination from the air. To a 50-cm3 boiler water sample, 2.5 cm3 of (1+1) hydrochloric acid was added, and then the solution was allowed to stand in boiling water for 30 min to digest the particulate matter containing iron. The water sample was then allowed to cool to room temperature, and 1 cm3 of the hydroxylammonium chloride solution, 2 cm3 of the TPTZ solution, 3.5 cm3 of the formic acid-ammonia buffer solution, and 0.08 cm3 of Muromac 50W-X2 resin (Muromachi Technos) were added in this order. Next, the mixture was stirred for 30 min at 25°C in the

thermostated bath. The colored resin beads were collected and packed in the black microcell using a pipette. The absorbances at 610 and 750 nm were measured using the single-beam type spectrophotometer. The difference in the absorbances at two wavelengths (ΔA) was used for the iron determination.

Results and Discussion

Absorption spectrum of Fe(TPTZ)22+ in solid phase

The solution of the produced complex of Fe(II) and TPTZ was reported to show an absorption maximum at the wavelength of 593 nm in many previous papers.5 However, in this study, the existing medium of the complex was a cation-exchange resin (50W-X2, 100 – 200 mesh), therefore, the absorption maximum shifted to a wavelength of 610 nm. For the absorbance measurement of the solid phase, it is fairly difficult to maintain the same packing conditions for all the samples, which affects the attenuance at the maximum absorption wavelength. To minimize the error caused by the packing state difference of the solid beads in the cell for each measurement, the absorbances of the solid phase were measured at the maximum-absorption wavelength (λ1 nm) and a weak or no-absorption wavelength (λ2 nm), then the difference in the attenuances (ΔA = A(λ1 nm) – A(λ2 nm)) was used for the quantitative analysis of the target analyte.8,18 Therefore, light attenuances at wavelengths λ1 = 610 nm and λ2 = 750 nm were selected for this measurement, then, ΔA was used for the determination of the total iron.

Optimum experimental conditionsEffect of pH on the color development. It was reported that the Fe(TPTZ)2

2+ complex is stable in the pH range of 3.4 – 5.8.10 However, some researchers confirmed that both H+ and OH– accelerate the dissociation of the Fe(TPTZ)2

2+ complex.13,19 Therefore, the developed color fades quickly in an acidic solution at a pH lower than 3; in addition, the precipitation caused by the hydrolysis of Fe(III) in a weak acidic solution (pH higher than 3.8) was reported by some researchers.20 The pH effect on the color development is shown in Fig. 1, indicating that the range of pH 3 – 4 was best for the formation of the Fe(TPTZ)2

2+ complex. The formate buffer solution was used

Fig. 1　pH dependence of the resin beads coloration. Sample solution: 50 cm3; Fe: 5 μg dm–3; Solid phase: 0.08 cm3 of Muromac 50W-X2 cation exchanger (100 – 200 mesh).

ANALYTICAL SCIENCES OCTOBER 2014, VOL. 30 1015

instead of the acetate buffer solution employed in the JIS method, since pKa of formic acid is around 3.7 and it is easy to control the pH around 3.7 in the formate buffer solution. To a 50 cm3 sample containing 2.5 cm3 (1+1) HCl, 3.5 cm3 of the formate buffer solution was added, and then produced a solution pH around 3.5.Effect of stirring time. The effect of the stirring time on the absorbance at 610 nm is shown in Fig. 2. The absorbance increased with the increasing stirring time in the experimental time period (20 – 40 min). However, in order to save time, the stirring time was fixed at 30 min in this study.Effect of temperature on the color development. The effect of the experimental temperature on the absorbance was examined by varying the temperature from 15 to 40°C using the thermostated bath. The absorbance was temperature-dependent, as shown in Fig. 3. The absorbance increased with the increase of the temperature and the effect was especially significant when the temperature was below 20°C, depending on the rates of complexation and adsorption on the resin beads. Therefore, to improve the reproducibility of the method, it is necessary to keep the experimental temperature constant. For reasons of practical analytical measurements, room temperature (25°C) was chosen in the experiment even though heating at the temperature of 40°C had a slightly positive effect on the absorbance.Effect of TPTZ concentration on the color development. The influence of the TPTZ concentration on determination was studied in the range from 20 to 140 μmol dm–3. A very slight change was observed in the experimental TPTZ concentration range, since even the lowest one (20 μmol dm–3) was much greater than the determined iron concentration (e.g., 5 μg dm–3 = 0.1 μmol dm–3). To force the reaction to completion, the TPTZ concentration was fixed at 80 μmol dm–3.

Interferences of coexisting ionsThe water used for the boiler systems in the power plants is

always first deionized, resulting in fairly low concentrations of various ions in the feed water. However, to evaluate the applicability of the TPTZ-SPS method, interferences from coexisting ions were examined, and the results are shown in Table 1. Mn2+, Cr3+, and Cd2+ gave no interferences at the present level of 1 mg dm–3 when the determined Fe concentration

was 2.5 μg dm–3. Compared to these ions, the interferences of Ni2+, Zn2+, Cu2+, and MoO4

2– were slightly more significant; 1 mg dm–3 of these ions resulted in several tens of percent errors for the determination of 2.5 μg dm–3 of total iron. A solution containing 200 μg dm–3 of these ions resulted in tolerable errors (about ±5%). The interference of Co2+ was the most significant among the tested ions. However, the concentrations of Co2+ and the other tested ions in the boiler water systems were not high enough to cause such high errors. The concern about the interferences can be negligible for the SPS-TPTZ total iron determination method.

Digestion of the sampleFor the purpose of inhibiting the acid attack corrosion of a

boiler system, the pH of the feed water is generally adjusted to 8 – 9 (even to 11) in the power plants.2,7 Under the weak alkaline condition, iron easily forms hydroxides and/or oxides, and exists as a colloid and precipitation. This study needs to determine the total concentration of iron including the ionic state, colloidal and particulate iron in the boiler system. Therefore, the colloidal and particulate iron in the boiler water should be dissolved to ionic iron before the quantitative analysis

Fig. 2　Time dependence of absorbance at 610 nm. Sample solution: 50 cm3; Fe: 3 μg dm–3; Solid phase: 0.08 cm3 of Muromac 50W-X2 cation exchanger (100 – 200 mesh).

Fig. 3　Temperature dependence of the coloration. Sample solution: 50 cm3; Fe: 3 μg dm–3; Solid phase: 0.08 cm3 of Muromac 50W-X2 cation exchanger (100 – 200 mesh).

Table 1　Interferences of coexisting ions for total iron determination by SPS method

ElementsConcentration/μg dm–3

Added Fe/μg dm–3

Found Fe/μg dm–3 Error, %

Mn2+ 1000 2.52 2.64 +4.8Ni2+ 1000 2.50 1.33 –46.7

200 2.51 2.62 +4.5Cr3+ 1000 2.51 2.62 +4.4Zn2+ 1000 2.50 1.69 –32.5

200 2.50 2.36 –5.5Cu2+ 1000 2.54 1.81 –28.8

200 2.52 2.39 –5.2Cd2+ 1000 2.54 2.58 +1.5MoO4

2– 1000 2.49 2.89 +16.0 200 2.55 2.70 +5.8

Co2+ 1000 2.50 4.96 +98.2 200 2.49 3.98 +59.8 10 2.55 2.95 +15.6 5 2.53 2.67 +5.6

1016 ANALYTICAL SCIENCES OCTOBER 2014, VOL. 30

of the total amount of iron in the water.To obtain the complete digestion of the colloidal and

precipitated iron in the sample, an effective digestion method was needed. Heating the sample solution with a strong acid in boiling water was chosen as the digestion method in this study. In Fig. 4, the time dependence of the digestion in boiling water is shown, and the results of two samples indicated that 30 min was sufficient to digest the colloidal and particulate forms of iron in the sample. A high-pressure digestion treatment of the sample using a microwave oven will improve the digestion time.

Application to real sample analysesAn example of the calibration curve plotted between the

absorbance of the complex versus the Fe concentration could be expressed as ΔA = 0.285 Fe (μg dm–3) + 0.034 (R2 = 0.997). For the total iron determination, the detection limit was defined as the concentration that gave an absorbance corresponding to 3σ for the standard deviation of the fluctuation of the blank, and was 0.11 μg dm–3 (n = 5) when a 50 cm3 sample solution was used.

To check the effectiveness of the present SPS method, the total iron concentrations in the water samples collected from three different sites in the Reihoku thermal power plant of the Kyushu Electric Power Co., Inc. were determined by the present method from 11:30 to 14:30 on December 20, 2013 (Table 2). The concentrations were at fairly low levels and just near the detection limit. The results of the A and C sites showed that there was a fluctuation in the iron concentration according to the sample collection time, indicating that there might be particulate iron in the sample. The fact was confirmed by the ICP-MS

measurement of the dissolved Fe in the samples (on-site filtered through 0.45-μm membrane filters), since almost no dissolved iron was detected. In addition, to check the precision of the method, recovery tests were conducted for the water samples taken from the same power plant, and the results are shown in Table 3. Almost all of the added iron was recovered.

To certify the analytical results, the concentrations were cross-checked using the samples collected from two boiler systems. In the case of the Genkai nuclear power plant of the Kyushu Electric Power Co., Inc. in Saga, the analytical results have already been reported.9 Although the SPS procedure was not optimized at that time, the results obtained were in fairly good agreement with those obtained by the JIS method after a 16-times concentration, as shown in Fig. 5. The analytical results of eight samples from the thermal power plant obtained by both ICP-MS and the SPS method were also plotted in the same figure. It can be seen that the points for all of the sample

Fig. 4　Time dependence of digestion of colloidal and particulate Fe. Sample: ♦ Site A; ■ Site C.

Table 2　Iron concentration of boiler water taken from a thermal power plant (concentrations in μg dm–3)

Site A B C

Collecting time 11:35 13:15 14:30 11:35 14:30 11:35 13:15 14:30

SPS 0.3 ± 0.1(n = 3)

0.4 ± 0.1(n = 3)

0.3 ±0.1(n = 3)

0.1 ± 0.1(n = 2)

0.1 ± 0.1(n = 2)

0.3 ± 0.1(n = 3)

0.3 ± 0.1(n = 3)

0.1(n = 1)

ICP-MSb Total 0.00 0.61 0.28 0.12 0.00 0.17 0.17 0.07Dissovled 0.00 —a 0.02 0.00 0.01 0.00 —a 0.04

a. No ICP-MS result.b. Agilent Model 7500 cx at Kyushu Environmental Evaluation Association; measured against the internal standard of 89Y.

Table 3　Recovery test for boiler water taken from a thermal power plant

Sample Added/μg dm–3 Found/μg dm–3 Recovery, %

Site C

Site B

02.42.404.64.7

0.22.92.80.14.54.8

—112108—

96100

Fig. 5　Relationship between the SPS results and cross-check results. ● Results of SPS and ICP-MS determination; ■ Results of SPS and JIS method (16 times concentration); —— Equivalent line Y = X.

ANALYTICAL SCIENCES OCTOBER 2014, VOL. 30 1017

examined lie just near the equivalent line in Fig. 5, confirming the accuracy and the effectiveness of the present method.

Conclusions

The present SPS method significantly improved the sensitivity of the TPTZ spectrophotometric method for the total iron determination, and made it possible to determine the sub-μg dm–3 level of total iron in water samples without any special treatment and expensive equipment. The present method has several advantages compared to the already proposed methods.9 Compared to the TPTZ solution method (JIS method), the required analytical time of the present SPS method is much shorter because the time-wasting concentration of the sample by heating is reduced; compared to the AAS and ICP-MS method, the equipment of the present SPS method is less expensive, decreasing the cost of the equipment and operating cost for the industries. The present method is useful for the monitoring of iron at μg dm–3 levels in boiler systems of both nuclear and thermal power plants as an indicator of the scale and corrosion occurrence.

Acknowledgements

The authors would like to thank the Reihoku thermal power plant of Kyushu Electric Power Co., Inc., Japan.

References

1. G. J. Bignold, “1.10-Corrosion in the Power Industry, Comprehensive Structural Integrity, Volume 1”, 2003, Elsevier, Swindon.

2. D. F. Farhad, M. Amir, H. T. Reza, and N. Farzad. Eng. Failure Anal., 2013, 28, 69.

3. M. M. Rahman, J. Purbolaksono, and J. Ahmad, Eng.

Failure Anal., 2010, 17, 1490. 4. M. Ameri and S. R. Shamshirgaran, Appl. Therm. Eng.,

2008, 28, 955. 5. S. Noori, A. Price, and W. H. John, Nucl. Eng. Des., 2006,

236, 405. 6. H. Kido, T. Ichihara, and S. Tsubakizaki, PowerPlant

Chem., 2012, 14, 568. 7. Japanese Standards Association, “Water Conditioning for

Boiler Feed Water and Boiler Water” (JIS B 8223-2006), 2006, Tokyo.

8. K. Yoshimura, H. Waki, and S. Ohashi, Talanta, 1976, 23, 449.

9. A. Maeda, N. Ishihara, S. Shoda, K. Yoshimura, A. Takahashi, and T. Matsuda, Therm. Nucl. Power, 2013, 64, 506.

10. P. F. Collins, H. Diehl, and G. F. Smith, Anal. Chem., 1959, 31, 1862.

11. J. Dahlen and S. Karlsson, J. Chromatogr., 1999, 848, 491. 12. S. M. Abdel-Azeen, N. R. Bader, H. M. Kuss, and M. F.

El-Shahat, Food Chem., 2013, 138, 1641. 13. E. B. Buchanan, Jr., D. Crichton, and J. R. Bacon, Talanla,

1966, 13, 903. 14. Sarenqiqige, M. Koga, M. Satoda, S. Matsuoka, and K.

Yoshimura, Kyushu University Global COE “Science for Future Molecular System” Journal, 2012, 5, 106.

15. Sarenqiqige, S. Saputro, S. Kai, M. Satoda, S. Matsuoka, and K. Yoshimura. Anal. Sci., 2013, 29, 677.

16. S. Saputro, K. Yoshimura, S. Matsuoka, K. Takehara, and Narsito, Anal. Sci., 2009, 25, 1445.

17. M. Koga, S. Matsuoka, and K. Yoshimura, Anal. Sci., 2010, 26, 963.

18. S. Matsuoka and K. Yoshimura, Anal. Chim. Acta, 2010, 664, 1.

19. G. K. Pagenkope and D. W. Margerum, Inorg. Chem., 1968, 7, 2514.

20. L. Kukoc-Modun and N. Radic, Croat. Chem. Acta, 2010, 83, 189.