Corrosion Science 50 (2008) 24932497

0010-938X/$ - see front matter 2008 Elsevier Ltd. All rights reserved.doi:10.1016/j.corsci.2008.06.034

Contents lists available at ScienceDirect

Corrosion Science

journal homepage: www.elsevier.com/ locate /cor sc i

1. Introduction

Formation and transformation of iron oxides is of interest to a wide variety of industries including steel making, power generat-ing, paint systems, pharmaceutical, and petrochemical, to name a few. The thermal hydraulic performance and integrity of the steam generators in nuclear power plants may be compromised due to the presence of corrosion deposits. The amount of iron transported in a steam generator is dependent on the composition of iron oxide formed in the feed train. Deposits that contain well crystallized magnetite and maghemite are more stable than deposits contain-ing a combination of oxides and oxyhydroxides [1]. Therefore, it is very important to quantitatively understand the composition of the deposits.

Standard methods for identifications and characterizations of iron oxides have traditionally used either X-ray diffraction (XRD) or Mssbauer spectroscopy (MS) [2]. XRD has been widely used in characterizing corrosion products. Although these techniques have served industry well in the past, they suffer from shortcom-ings that could be replaced by Fourier transform infrared spectro-photometry (FTIR). MS is a technique that utilizes a live radioactive source, which makes the technique relatively unsafe from an oper-ational point of view, since it poses a potential health risk to the operator. The main limitations of MS technique include the level

of operator expertise requirements and the complexity of spectral interpretation. Even though XRD is easier to operate and interpret, its spectra is limited in its ability in differentiating magnetite and maghemite.

On the other hand, FTIR instrumentation is simple and spectra interpretation is relatively easy [3]. In the past, it has been estab-lished that the FTIR technique can be routinely used to efficiently differentiate and quantify different iron oxides and oxyhydroxides. To the best knowledge of the authors, no attempts have been made in using FTIR for quantitative analysis of iron oxides formed in the power plants. The main objective of this research is to quantify iron oxide phases formed in the secondary side of the steam gener-ator units at Comanche peak steam electric station (CPSES). Such analysis will allow better interpretation and control of the corro-sion process.

2. Literature review

Magnetite is a well-known form of iron oxide that forms at room temperature in crevices between steel plates and at elevated temperature inside boiler tubes, heat exchangers etc. The oxida-tion product of Fe3O4 is either -Fe2O3 or -Fe2O3 depending on the oxidation temperature and/or possibly the crystallite size of the starting magnetite [45].

Studies performed by Nasrazadani and Raman [4] have shown that transformation of magnetite to hematite goes through the for-mation of maghemite. The production of maghemite begins with

Quantitative analysis of iron oxides using Fourier transform infrared spectrophotometry

H. Namduri, S. Nasrazadani *

College of Engineering, University of North Texas, Denton, TX-76207, USA

a r t i c l e i n f o a b s t r a c t

Article history:

Received 14 August 2007

Accepted 17 June 2008

Available online 4 July 2008

In this study, a systematic approach based on the application of Fourier transform infrared spectropho-

tometry (FTIR) was taken, in order to quantitatively analyze the corrosion products formed in the sec-

ondary cycle of pressurized water reactors (PWR). Binary mixtures of iron oxides were prepared with

known compositions containing pure commercial magnetite (Fe3O4), maghemite ( -Fe2O3), and hematite

( -Fe2O3) for calibration purposes. Calcium oxide (lime) was added to all samples as a standard reference

in obtaining the calibration curves. Using regression analysis, relationships were developed for intensity

versus concentration for absorption bands corresponding to each of the phases in their corresponding

FTIR spectrum. Correlation coefficients, R2, of 0.82, 0.87, and 0.86 were obtained for maghemitemagne-

tite, hematitemagnetite, and hematitemaghemite systems, respectively. The calibration curves gener-

ated were used to quantify phases in multi-component unknown field samples that were obtained from

different components (moisture separators, condensers, and high and low pressure heaters) of the two

units (units 1 and 2) of the secondary cycle of the Comanche Peak PWR.

2008 Elsevier Ltd. All rights reserved.

Keywords:

A. Steel

B. IR spectroscopy

C. Rust

C. Oxidation

C. Passivity

* Corresponding author. Tel.: +1 940 565 4052; fax: +1 940 565 2666.E-mail address: nasr@unt.edu (S. Nasrazadani).

http://www.sciencedirect.com/science/journal/0010938XNancy Prietohttp://www.elsevier.com/locate/corscimailto:nasr@unt.edu%20

2494 H. Namduri, S. Nasrazadani / Corrosion Science 50 (2008) 24932497

nucleation and growth of goethite or lepidocrocite, followed by dehydration to hematite, and then reduction to magnetite.

The deposition rate of hematite is an order of magnitude greater than magnetite. As seen from the Pourbaix diagram of iron, it is important that the reducing conditions be maintained in the steam generators during operation, so as to facilitate formation of magnetite. Turner and Klimas showed that lowering the concentra-tion of hematite relatively to magnetite in the feedwater will signif-icantly lower the rate of tube bindle fouling [6]. Theoretical studies by Jobe showed that hematite has a very low solubility and a much smaller dissolution rate than magnetite and lepidocrocite in the presence of 5 ppb of dissolved oxygen. Formation of thin layer of maghemite/magnetite is known to act as a very good passive film [7].

FTIR spectra of iron oxides are well established. It is been observed that the absorption band at a high wavenumber region is due to OH stretching, and at lower wavenumber as a result of Fe-O lattice vibration. FTIR spectrum of magnetite exhibits two strong infrared absorption bands at 570 cm1 ( 1) and 390 cm

1( 2) [8]. According to Ishii et al, these bands can be assigned to the Fe-O stretching mode of the tetrahedral and octahedral sites for the

1 band at 570 cm1 and the Fe-O stretching mode of the octahe-

dral sites for the 2 band at 390 cm1, provided that Fe3+ ion dis-

placements at tetrahedral sites are negligible [8]. FTIR spectrum of magnetite exhibits two other absorption bands at 268 cm1and 178 cm1 which were beyond the detection limit of our instru-ment. Maghemite, a defective form of magnetite, has absorption bands at 630 cm1, 590 cm1, and 430 cm1. Table 1 summarizes possible FTIR peaks for different iron oxides.

Legodi and his group performed quantitative analysis on cal-cium carbonate present in different cement blends using FTIR [9]. Reig and group performed quantitative FTIR analysis on calcium

carbonate and silica (quartz) using the constant ratio method. The group used potassium ferricynaide as standard and successfully showed the accuracy of quantifying the concentration of silica and quartz in geological samples using FTIR [10]. The same group also successfully showed that FTIR can be used to quantify butyl acetate and toluene in binary and ternary mixtures using constant method ratio. They used valeronitrile as the standard and they also showed that the above method is independent of optical path length [11].

The Xu group showed that FTIR can be efficiently used for quan-tifying minerals. They used a multifunctional analysis, which is based on Beers law to quantify different minerals present in oil wells. In this method, the absorbance at a specific wave number is equal to the sum of the absorbance of all sample components at that wavenumber [12].

3. Experimental procedure

Commercially available powders of magnetite (puratronic 99.999% purity), maghemite (99+% purity), and hematite (99.99% purity) were obtained. Three binary sets of sample mixtures with known concentrations of maghemite and magnetite, hematite and magnetite, and hematite and maghemite were prepared. All the samples were added to KBr powder and compressed into pellets using hydraulic press. Magnitude of compression applied in KBr pellet preparation needs to be kept constant to avoid variance in absorbance intensity from one sample to the next. Nicolet Avatar 370 DTGS FTIR was used to quantify iron oxides. FTIR spectra collec-tion was done for 32 scans with 2 cm1 resolutions. Three equiva-lent runs of each of the three sets of the samples were made on the FTIR spectrometer. The average values of background subtracted peak intensity results were used for obtaining calibration curve. Once all of the spectra for the samples were obtained, three cal-ibration curves were drawn for the three sets of samples. To set the calibration curve for known amounts of iron oxide (magnetite, maghemite and hematite) in each mixture, I/Io ratio was used. The intensity of the iron oxide peak (magnetite 570 cm1, hematite-540 cm1 and maghemite 630 cm1) is represented by I and Io rep-resents, the intensity of the 3640 cm1 peak of CaO.

This calibration curve was used to quantify the amount of iron oxides present in the field samples collected from the secondary side of CPSES. The most readily available samples of the secondary system were obtained from the feedwater heaters (FW HTR-low pressure and high pressure feedwater heaters), condenser, and moisture separator-reheater (MSR), as these components are rou-

Table 1

Infrared bands of different iron oxides [4,13]

Iron oxide/hydroxide Wave numbers of bands (cm1)

Magnetite (Fe3O4) Broad bands at 570 and 400 cm1

Maghemite ( -Fe2O3) 700, 630660, 620 range (Fe-O range)

Hematite ( -Fe2O3) 540, 470 and 352 cm1

Goethite ( -FeOOH) 1124, 890 and 810 cm1 for OH stretchLepidocrocite ( -FeOOH) 1018 cm1 (in plane vibration) and

750 cm1 (out of plane vibration)

Fig. 1. Simplified schematic of secondary system sample locations.

H. Namduri, S. Nasrazadani / Corrosion Science 50 (2008) 24932497 2495

tinely opened during outages and represent major temperature locations of the system as illustrated by Fig. 1.

The CaO absorption band was used primarily as a reference because it does not interfere with any of the iron oxide phases. Even though an absorption band of 3640 cm1 is close to the OH band, it has a very distinct peak and can be easily discerned (Fig. 2). CaCO3 (894865 cm

1 peak) has been previously used as a standard reference in quantifying the amount of limestone in different cement blends [9]. A linear fit was used to obtain the calibration curve.

4. Results and conclusions

Fig. 3 shows FTIR spectra of single phases of hematite, mag-netite, and maghemite. A sharp peak at 3640 cm1 belonging to calcium oxide is shown in all of the spectra of iron oxides. The peak intensity of CaO was fairly constant in all the spectra. Cali-bration curves were obtained for combinations of two phases of iron oxides. Correlation factors of 0.822 (magnetite added to maghemite, Fig. 4), 0.8584 (maghemite added to hematite, Fig. 5), and 0.8708 (magnetite added to hematite, Fig. 6) were obtained.

Fig. 3. FTIR spectra of 100% hematite, maghemite and magnetite.

Fig. 2. FTIR spectra of 100% CaO showing 3640 cm1 peak.

2496 H. Namduri, S. Nasrazadani / Corrosion Science 50 (2008) 24932497

A confidence limit of 95% was used in the calculations. The aver-age values of I/Io for three runs made with mixtures with different concentrations of iron oxides are shown in the Table 2. Hematite peak at 540 cm1 intensity (the most intense peak for hematite) was used for the I value for mixtures containing hematite and mag-netite, and maghemite and hematite. The peak at 630 cm1 (the

most intense peak of maghemite) was used for I values in the case of mixtures containing magnetite and maghemite. No peak inter-ferences of any phases were observed in all of the mixtures since FTIR spectra of all iron oxides are well resolved and spectra resolu-tion of most FTIR instruments is 2 cm1. These calibration curves were then used to quantify the iron oxide phases present in the field samples collected from the secondary side of unit 1 and unit 2 of CPSES.

The percentage concentrations of the iron oxides present in the selected field samples is given in Table 3. The samples from the moisture separator mainly show hematite and magnetite. The main feedwater heater sample showed 96% magnetite and about 4% maghemite. The high-pressure feedwater heater sam-ple showed mostly hematite; whereas, low-pressure feedwater heater sample showed hematite and magnetite. The presence of magnetite and hematite is expected in feedwater systems due to the transformation of hydroxides and other hydrated iron species, which move through the feedwater system into stable iron oxides (Schikorr reaction). The two samples form the main condenser mainly consisted of hematite with traces of magnetite and maghe-mite. Detection limits determination for iron oxides quantification using FTIR was not done in this study and is planned for future work.

5. Summary and conclusions

A quantitative determination of iron oxides can be quickly per-formed relatively accurately using FTIR technique. The technique involves taking mid infrared measurements with samples in the form of KBr pellets. By using a standard reference like CaO, nor-malization can be performed. The peak of 3450 cm1 is free from interference with any of the major iron oxide peaks considered in this study. This method makes it a very suitable method in quickly determining the concentrations of major iron oxides in the power industry.

The FTIR technique was reconfirmed to be a valuable tool to dif-ferentiate between Fe3O4 and -Fe2O3. It is also been shown that this technique can be used in quantifying iron oxides.

It has also been shown that the infrared active mode of calcium oxide can be efficiently used in the quantification process. The FTIR quantification method performed in this study can be further fine-tuned and extended to other major metallic oxides including: chro-mium oxide, nickel oxide, lead oxide, and silicon dioxide. This will prove valuable for studying corrosion deposits formed in nuclear power plants.

Hematite/MagnetiteHematite 540 cm-1

Fig. 5. FTIR calibration for mixture containing hematite and magnetite.

Maghemite/MagnetiteMaghemite 630 cm-1

Fig. 4. FTIR calibration for mixture containing magnetite and maghemite.

Table 2

FTIR intensities for different known concentrations of iron oxides used in calibra-

tion curves (I = intensity of iron oxide mixture, and Io = intensity of 3640 cm1 peak

of CaO)

Hematite (%) Magnetite (%) Maghemite (%) I/Io

100 1.30

100 1.13

100 1.02

80 20 1.01

60 40 0.98

40 60 0.82

20 80 0.81

80 20 1.13

60 40 1.02

40 60 0.91

20 80 0.95

20 80 0.86

40 60 0.81

60 40 0.75

80 20 0.76

Fig. 6. FTIR calibration for mixture containing maghemite and hematite.

H. Namduri, S. Nasrazadani / Corrosion Science 50 (2008) 24932497 2497

Acknowledgement

Authors would like to thank both Mr. Jim Stevens and Mr. Rob-ert Theimer for providing field samples.

References

[1] Domingo, Clementa, The pathways to spinel iron oxides by oxidation of iron (II) in basic media, Materials Research Bulletin 26 (1991) 4755.

[2] Blesa, A.J.G. Maroto, S.I. Passaggio, F. Labenski, C. Saragovi-Badler, Moessbauer study of the behaviour of synthetic corrosion products of nuclear power plants, Radiation Physical Chemistry 11 (1978) 321326.

[3] Brundle Richard, Charles Evans, Wilson, Encyclopedia of Materials Character-ization, ButterworthHeinemann, 1992, ISBN 07506-9168-9.

[4] S. Nasrazadani, A. Raman, Application of IR spectra to study the rust systems, Corrosion Science 34 (8) (1993) 13351365.

[5] Nasrazadani, Namduri, Steven, Theimer, Fellers, Application of FTIR in the Analysis of Iron Oxides and Oxyhydroxides Formed in PWR Secondary Sys-tem, 2003 Steam Generator Secondary Side Management Conference, Febru-ary 1012, 2003.

[6] Turner, Klimas, The Effect of Alternative Amines on the Rate of Boiler Tube Fouling, Final Report, TR-108004, EPRI Report, September 1997.

[7] David Jobe, The calculated solubilities of hematite, magnetite and lepidocro-cite in steam generator feed trains, AECL, 1997.

[8] M. Ishii, M. Nakahira, Infrared absorption spectra and cation distribution in (Mn,Fe)3O4, Solid State Communications 11 (1972) 209212.

[9] Legodi, D. De Waal, J.H. Potgieter, Quantitative determination of CaCO3 in cement blends by FT-IR, Society for Applied Spectroscopy 55 (3) (2001) 361365.

[10] Reig, J.V.G. Adelantado, M.C.M. Moya Moreno, FTIR quantitative analysis of calcium carbonate(calcite) and silica (quartz) mixtures using the con-stant ratio method. Application to geological samples, Talanta 58 (2002) 811821.

[11] Reig, J.V. Gimeno Adelantado, V. Peris Martinez, M.C.M. Moya Moreno, M.T. Domenech Cerbo, FT-IR quantitative analysis of solvent mixtures by the constant ratio method, Journal of Molecular Structure 480481 (1999) 529534.

[12] Xu, B.C. Cornilsen, D.C. Popko, B. Wei, W.D. Pennington, J.R. Wood, Quantita-tive mineral analysis by FTIR spectroscopy, The Internet Journal of Vibrational Spectroscopy 5 (4) (2001) 112.

[13] R.M. Cornell, U. Schwertmann, The Iron Oxides, Weinheim, New York, 1996.

Table 3

Concentration of field sample collected from different components of secondary side of unit 1 and unit 2 steam generator system of CPSES

Field samples Sample description Oxides present Io Maghemite (I) Hematite (I) I/Io Final concentrations

1 Moisture separator Magnetite and hematite 3.02 2.85 0.94 52% Hematite and 48% magnetite

2 Main feedwater heater Magnetite and maghemite 4.29 2.81 0.66 96% magnetite and 4% maghemite

3 Main condenser, hotwell Magnetite and hematite 3.08 4.01 1.3 Mostly hematitea

4 Main condenser Maghemite and hematite 0.28 0.49 1.74 Mostly hematiteb

5 High pressure feedwater heater Maghemite and hematite 0.75 1.03 1.37 Mostly hematiteb

6 Low pressure feedwater heater Magnetite and hematite 5.81 5.37 0.92 52% magnetite and 48% hematite

a Amount of magnetite was below the detection limit. b Amount of maghemite was below the detection limit.

Quantitative analysis of iron oxides using Fourier transform infrared spectrophotometry Introduction Literature review Experimental procedure Results and conclusions Summary and conclusionsAcknowledgementReferences