Steel corrosion in diluted ammoniac solutions studied [601657]

Steel corrosion in diluted ammoniac solutions studied
by Mo ¨ssbauer spectrometry
I. Bibicua,*, A. Samideb, M. Predab
aNational Institute of Materials Physics, P.O. Box MG-6, Bucharest 76900, Romania
bFaculty of Chemistry, University of Craiova, Craiova 1100, Romania
Received 15 August 2003; received in revised form 16 March 2004; accepted 20 March 2004
Available online 26 May 2004
Abstract
The corrosion product films that were formed on industrial, low-carbon concentration Fe–C steel after immersion in diluted ammoniac
solutions were investigated by57Fe Mo ¨ssbauer spectroscopy. The amorphous Fe3+oxyhydroxide was the major constituent of the corrosion
layers. Layer thickness increased as NH 3concentration changed from 10/C01to 10/C04N. All CEMS spectra indicated a magnetic anisotropy on
the surface of the samples.D2004 Elsevier B.V . All rights reserved.
Keywords: Low alloy steel; Corrosion; Mo ¨ssbauer spectroscopy; Thin films
1. Introduction
Researchers have commonly ignored the industrial and
residual waters containing ammonia/ammonium regardingthem as harmless to the environment and posing no dangerin terms of corrosion. Relatively recent researches, however,have showed such waters have a negative impact on theenvironment leading to pipe corrosion in the cooling waterssystems, especially in the ammonium fertilizer industry [1–
3]. Studies have been undertaken on the carbon steel
behavior in ammonia mediums containing (NH
4)2CO 3[4],
as well as on the corrosion products formed on the carbonsteel surface in mediums containing NH
4NO 3[5]and in
diluted ammonia mediums containing NH 3[6–9] .
Fe–C steel corrosion in diluted ammoniac solutions was
investigated in this article mainly by using Mo ¨ssbauer
spectroscopy. Mo ¨ssbauer spectroscopy is particularly well
suited for providing an insight into the degradation of many
materials, because it makes it possible to study the behavior
of iron on a microscopic scale. Weight loss measurementswere used to supplement the Mo ¨ssbauer data.2. Experimental
Industrial Fe–C steel with low (up to 0.10 wt.%) C
concentration was used as samples. The overall sampledimensions were 33 /C222/C21m m .T h es u r f a c eo ft h e
samples was polished with diamond paste, degreased inethylic alcohol and desiccated in warm air. The sampleswere corroded in a standard electrolytic cell using dilutedammoniac solutions with NH
3concentrations of 10/C01,
10/C02,1 0/C03and 10/C04N. Corrosion took place at a
temperature of 25 jC, as the corrosive environment was
kept in contact with air. The working electrode was a 4-cm2
plate made of Fe–C steel with low C concentration. Thereference electrode was made of saturated calomel. Theauxiliary electrode was identical with the working electrode.The samples were kept at corrosion potential for 20 min andthen corroded for 35 min at current densities in the range2.3/C210
/C03to 9/C210/C02mA/cm2. Electrochemical measure-
ments [10] have showed that the final potentials reached by
samples in the corrosion process are located in the activeregion.
Mo¨ssbauer spectroscopy was performed at room temper-
ature in the transmission (TMS) and conversion electronspectroscopy (CEMS) [11] using a conventional constant-
acceleration spectrometer with a
57Co–Rh source. The
0167-577X/$ – see front matter D2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.matlet.2004.03.043* Corresponding author. Tel.: +40-21-4930047; fax: +40-21-4930267.E-mail address: [anonimizat] (I. Bibicu).www.elsevier.com/locate/matletMaterials Letters 58 (2004) 2650–2653

CEMS measurements were conducted with a high degree of
accuracy, ensuring the same geometry of the detection spaceand same gas flow rate for all the samples. The parameters
of the Mo ¨ssbauer spectra were calculated using a computer-
fitting program, which assumed a Lorentzian line shape.The isomer shifts were referred to a-Fe.
The weight loss measurements were carried out using a
Keithley 2420 3A Source Meter and were focused oncalculating the corrosion rate of carbon steel in dilutedammonia solutions. The steel samples were made of thesame industrial Fe–C steel with low C concentration and
prepared in the same way as for the corrosion process. The
samples were kept for 5 days in the respective media in aclosed system at room temperature. The corrosion productswere removed as soon as the samples had been taken out ofthe aggressive solutions, by washing them in a 5% HClsolution, in a warm medium. The samples were thendegreased in alcohol and dried in warm air. The gravimetricindex k
g(g/m2day) was used to express the corrosion rate.
3. Results and discussion
The TMS and CEMS spectra of a reference sample
before corrosion are shown in Figs. 1 and 2 . The best fitting
of the TMS spectrum indicated the presence of a singlesextet. The parameters of this sextet were practically the
same as for a a-Fe sample. The full width at half-height of
the outermost lines (0.3 F0.01 mm/s) confirmed the low
concentration of an alloying element found by chemicalanalysis. The CEMS spectrum showed the presence of thesame a-Fe but with a lower value for the hyperfine field,
namely 26.308 /C210
6F0.199 /C2106A/m (330.6 F2.5 kOe)
than the 26.38 /C2106F0.159 /C2106A/m (331.5 F2 kOe)
obtained for TMS spectrum. Experimental errors account for
this decrease. In the CEMS spectrum, the intensities of the
second and fifth peaks of the a-Fe spectrum with respect tothe third and fourth peaks showed that the directions of the
g-ray and magnetic moments were nearly perpendicular. A
magnetic anisotropy mainly as a result of polishing was
found on the surface of the samples. By contrast, the TMS
spectra showed that the magnetic moments inside thesample were in a random arrangement. The line width ofthe CEMS spectrum was 0.26 F0.02 mm/s. A smaller line
width was expected in the backscattering geometry due tolack of saturation broadening.
The CEMS spectra of the samples corroded in solutions
with 10
/C01and 10/C04NH 3concentrations are shown in Figs.
3 and 4 . The best fit of the CEMS spectra for the corroded
samples used a Fe3+paramagnetic doublet in addition to the
sextet. The parameters of the sextet (hyperfine magneticfield, quadrupole splitting, isomer shift, and line width)were almost identical to those of the uncorroded sample.We noted, though, an increase in isomer shift, to0.02F0.02 mm/s, as well as a continuous decrease of the
hyperfine magnetic field value as the NH
3concentration
Fig. 1. Transmission Mo ¨ssbauer spectrum of the reference sample from
Fe–C with low C concentration before corrosion ( .data; — fit).Fig. 2. Conversion electron Mo ¨ssbauer spectrum of the reference sample
from Fe–C with low C concentration before corrosion ( .data; — fit).
Fig. 3. Conversion electron Mo ¨ssbauer spectrum of a Fe–C steel with low
C concentration sample after corrosion in a solution with 10/C01NH 3
concentration ( .data; — fit; /C1/C1/C1Fe3+;– a-Fe).I. Bibicu et al. / Materials Letters 58 (2004) 2650–2653 2651

changed from 10/C01to 10/C04N. The value of the hyperfine
magnetic field dropped from 26.308 /C2106F0.199 /C2106
A/m (330.6 F2.5 kOe) to 26.054 /C2106F0.199 /C2106A/
m (327.4 F2.5 kOe). The latter value along with the slight
change in isomer shift can be ascribed to the presence ofcarbon impurities. This suggests that in the process ofcorrosion there may be a certain preference for the positionsof iron, which are not close to the atoms of the alloyingelements. We noted that the width of the sextet lines tendedto decrease from 0.26 mm/s (in the case of uncorrodedsample) to 0.25 mm/s. The preferential orientation of the
magnetic moments in the sample plane persisted even after
the corrosion of the samples. The main difference betweenthe sextets of corroded samples and those of uncorrodedsamples consisted in a decrease of the intensity lines. Thisproved the presence of a superficial layer on the corrodedsamples. The thickness of this layer increased when NH
3
concentration dropped. The relative area of the doubletincreased from 3 F3% up to 8 F3% when NH
3concen-
tration decreased to 10/C04N. This reconfirmed that a
superficial compound without magnetic ordering wasformed as a result of corrosion. The quadrupole splittingvalues were in the range 0.62 F0.03 mm/s, while the
isomer shift values increased in a straight manner as theNH
3concentration logarithm decreased, from 0.3 mm/s for
10/C01N concentration to 0.42 for 10/C04N concentration.
The isomer shift determination error increased to F0.05
mm/s for 10/C01and 10/C02N concentrations. These param-
eters indicated the presence of Fe3+and were similar to
those shown by amorphous Fe3+oxyhydroxides [12],
superparamagnetic a-FeOOH and/or g-FeOOH [13–22] ,
and Fe (OH) 3[23]. This doublet is hard to assign to a
specific chemical species on the sole basis of the Mo ¨ssbauer
data obtained at room temperature. Small relative areas ofthe doublet as well as its parameters showed an initial stage
of corrosion. At this stage, we believe the main product of
corrosion is an amorphous Fe
3+oxyhydroxide with a non-stoichiometric composition. The isomer shift change of the
superficial compound showed obvious alterations of the Fechemical bonding as NH
3concentration changed. There
was also a process of inhibition of carbon–steel corrosionin the ammoniac solutions, which was revealed by an
increase in the superficial layer as NH
3concentration
decreased.
The weight loss measurements that were performed on the
same samples confirmed the inhibition process. The weightlosses are presented as general corrosion rate in Fig. 5 .W e
obtained a good data fitting with a first-order exponentialdecay. The increase in NH
3concentration obviously led to a
decrease of the corrosion rate.
4. Conclusions
The corrosion product films that were formed on indus-
trial, low-carbon concentration Fe–C steel after immersionin diluted ammoniac solutions were investigated by
57Fe
Mo¨ssbauer spectroscopy. The amorphous Fe3+oxyhydrox-
ide was the major constituent of the corrosion layers. Layer
thickness increased as NH 3concentration changed from
10/C01to 10/C04N. An inhibition of carbon–steel corrosion
was found to occur in the ammoniac solutions, which washighlighted by an increase in the superficial layer as NH
3
concentration decreased. The inhibition process was con-firmed by weight loss measurements performed on the samesamples. All CEMS spectra indicated a magnetic anisotropy
on the surface of the samples.Fig. 5. Variation of the corrosion rate with NH 3concentration for the
corroded Fe–C steel samples.
Fig. 4. Conversion electron Mo ¨ssbauer spectrum of a Fe–C steel with low
C concentration sample after corrosion in a solution with 10/C04NH 3
concentration ( .data; — fit; /C1/C1/C1Fe3+;– a-Fe).I. Bibicu et al. / Materials Letters 58 (2004) 2650–2653 2652

References
[1] R. Marr, M. Koncar, Chem. Ing. Tech. 62 (3) (1990) 175.
[2] M. Kotter, L. Riehert, Chem. Ing. Tech. 61 (1) (1984) 174.[3] P.J. Ruffier, W.C. Boyle, I. Kleinschmied, J. Water Pollut. 3 (1981)
367.
[4] H.S. Khim, T.T. Kho, K. Osseoasare, H.W. Pickering, Metall. Trans.,
B, Process Metall. 22B (1991) 232.
[5] P.C. Bhat, M.P. Sathyavathiamma, N.G. Mallya, Corros. Sci. 23
(1983) 733.
[6] R. Cabrera-Sierra, I. Garcia, E. Sosa, M.T. Oropeza, I. Gonzalez,
Electrochim. Acta 46 (2000) 487.
[7] I. Bibicu, A. Samide, M. Preda, ‘‘5th Seeheim Workshop on Mo ¨ss-
bauer Spectroscopy’’, Seeheim, Germany, 21–26 May 2002,
Abstracts Volume, Dohannes Gutenberg Universita ¨t, Mainz, Ger-
many, 2002, p. 50.
[8] M. Preda, A. Samide, I. Bibicu, N. Cioatera, An. Univ. Craiova 31
(2002) 25.
[9] E. Sosa, R. Cabrera-Sierra, M.T. Oropeza, F. Hernandez, N. Casillas,
R. Tremont, C. Cabrera, I. Gonzalez, Electrochim. Acta 48 (2003)1665.
[10] A. Samide, M. Preda, N. Cioatera, Rev. Chim. 54 (2003) 404.
[11] I. Bibicu, M.S. Rogalski, G. Nicolescu, Meas. Sci. Technol. 7 (1996)
113.
[12] M. Gracia, J.F. Marco, J.R. Gancedo, W. Exel, W. Meisel, Surf.
Interface Anal. 29 (2000) 82.[13] J.H. Terrell, J.J. Spijkerman, Appl. Phys. Lett. 41 (1968) 11.
[14] A.M. Pritchard, B.T. Mould, Corros. Sci. 11 (1971) 1.
[15] P.O. V oznyuk, V .N. Dubinin, Sov. Phys., Solid State 15 (1973) 1265.[16] H. Kubsch, H.A. Schneider, Investigations of the initial stage of
corrosion by Mo ¨ssbauer spectrometry, in: Z. Hrynkiewicz, J.A.
Sawicki (Eds.), Proceedings International Conference on Mo ¨ssbauer
Spectroscopy, vol. 1, Akademia Gorniczo-Hutnicza, Krakow, Poland,1975, p. 315.
[17] D. Chambaere, A. Govaert, E. de Grave, G. Harts, G. Robbrecht,
J. Phys., Colloq. C2 40 (Suppl. au 3) (1979) C2–C350.
[18] J. Chen, Z. Cai, Z. Wang, C. Wang, H. Zhang, Proceedings of the
Indian National Science Academy, International Conference on the
Applications of the Mo ¨ssbauer Effect 1981 Jaipur India, Edited 1982,
INSDOC, New Delhi, 1982, p. 261.
[19] R.K. Nigam, U.S. Mehrotra, S. Varma, S.N. Pandey, Proceedings of
Indian National Science Academy, International Conference on the
Applications of the Mo ¨ssbauer Effect 1981 Jaipur India, Edited 1982,
INSDOC, New Delhi, 1982, p. 299.
[20] D.C. Cook, Hyperfine Interact. 28 (1986) 891.[21] J. Davalos, J.F. Marco, M. Gracia, J.R. Gancedo, J.M. Greneche,
Hyperfine Interact. 57 (1990) 1809.
[22] J.F. Marco, J. Davalos, M. Gracia, J.R. Gancedo, Hyperfine Inter-
actions 57 (1990) 1991.
[23] P.P. Bakare, M.P. Gupta, A.P.B. Sinha, Indian J. Pure Appl. Phys. 18
(1980) 473.I. Bibicu et al. / Materials Letters 58 (2004) 2650–2653 2653