Hot-water corrosion failure of a hard-drawn copper tube

Theofani Tzevelekou
a
, Athanasia Flampouri
a
, Andreas Rikos
a
, Athanasios Vazdirvanidis
a
,
George Pantazopoulos
a,
, Dionysios Skarmoutsos
b
a
ELKEME Hellenic Research Centre for Metals S.A., 252 Pireaus Str., 17778 Athens, Greece
b
Halcor S.A. Metal Works, 57th km Athens-Lamia National Highway, 32011 Oinofyta, Greece
a r t i c l e i n f o
Article history:
Received 5 February 2013
Received in revised form 3 April 2013
Accepted 26 April 2013
Available online 22 May 2013
Keywords:
DHP-copper
Pitting corrosion
Leakage failure
Corrosion deposits
a b s t r a c t
Deoxidized high phosphorus copper (C12200, DHP-Cu) is the principal construction ele-
ment in piping in water distribution systems and HVAC (Heating Ventilation and Air-Con-
ditioning) installations due to its excellent thermal and electrical conductivity, formability,
corrosion resistance and antimicrobial activity. However, design and installation decien-
cies or aggressive environmental parameters, such as improper water quality, may lead to
unexpected failures causing material losses and many times property damage incurring
high costs of repair. A corroded copper water tube caused leakage in a hot water hydraulic
installation. The leakage was reported after approximately 6 years in service. Tubes were
disconnected from a hydraulic installation and samples were taken for failure investiga-
tion. Failure analysis ndings suggest strongly that the leakage was caused by tube perfo-
ration from the inner waterside due to localized pitting corrosion. The main cause of the
accelerated tube pitting corrosion was the improper water chemistry (especially high chlo-
rine content and low pH) inducing chemical attack of copper leading to scale layers of cupr-
ite, chlorides and sulfates.
2013 Elsevier Ltd. All rights reserved.
1. Introduction and background information
Deoxidized-high-phosphorus (DHP) copper tubes are extensively used in water, healthcare, heating and refrigerant appli-
cations due to their high thermal conductivity, good heat transfer characteristics, excellent formability, corrosion resistance
and antibacterial properties. Characteristic failures of copper tubes at various applications ranging from water supply to heat
exchanger systems, may be attributed to pitting corrosion [1], intergranular cracking and corrosion [2,3], stress-corrosion
cracking [47], fatigue [7], erosioncorrosion [8], and microbiological inuenced corrosion (MIC) [9]. However, design and
installation deciencies or aggressive environmental conditions, such as improper water chemistry and stagnant water con-
ditions, may be considered as very important contributors of pipeline unexpected damages.
A 52 1 mm CuDHP hard drawn sample which exhibited leakage during service was received for corrosion failure
investigation. The chemical composition of the DHP tube is within the range: Cu minimum = 99.90 wt.%, P = 0.015
0.040 wt.%, according to the EN 1057 standard and in compliance with other relevant specications for water supply appli-
cations. The damaged area was properly marked. According to case history information, the installation has been completed
since early 2006 and it was a part of a hot water circuit. Service time of the water installation is estimated to be approxi-
mately up to 6 years. According to background information, owing medium is normal tap water which has been subjected
to water treatment, i.e. chlorination. Usual circuit temperature was up to 5060 C while typical operation pressure ranged
within 23 bar. Unfortunately, no other specic information on representative water chemistry used through service
1350-6307/$ - see front matter 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.engfailanal.2013.04.020

Corresponding author. Tel.: +30 210 4898263; fax: +30 210 4898268.
E-mail address: gpantaz@halcor.vionet.gr (G. Pantazopoulos).
Engineering Failure Analysis 33 (2013) 176183
Contents lists available at SciVerse ScienceDirect
Engineering Failure Analysis
j our nal homepage: www. el sevi er . com/ l ocat e/ engf ai l anal
period could be retrieved. Optical and scanning electron microscopy coupled with Energy Dispersive X-ray Spectroscopy for
local elemental analysis were employed for microstructural and morphological examination and were the main analytical
techniques for the present investigation. Moreover X-ray Diffraction Analysis of the corrosion products was also performed
to identify the main crystalline phases participated on the inner tube scale deposits.
2. Experimental procedure
Macroscopic observations were performed with a Nikon SMZ 1500 stereo-microscope. Cross-section of 52 1 tube was
prepared using hot-mounting, wet grinding up to 1200 grit SiC paper and polishing with diamond and silica suspensions.
Metallographic examination was conducted using a Nikon Epiphot 300 inverted metallographic microscope. Microhardness
testing was performed in an Instron-Wolpert 2100 device using 200 gf applied load. Higher magnication microscopic obser-
vations were carried out on as polished sections using an FEI XL40 SFEG scanning electron microscope (SEM) equipped with
an EDAX Energy Dispersive X-ray Spectrometry (EDS) detector for elemental analysis of selected areas. X-ray Diffraction of
corrosion products was performed using Cu Ka monochromatic radiation employing a Philips XPert Pro diffractometer with
an X-ray Tube PW3373/00 Cu LFF DK117025 according to Refs. [10,11]. X Pert Organiser, Data Collector, Graphics & Identify
was used as analysis software, while the determination and identication of crystalline phases was conducted through Crys-
tallographica Search Match software based on standard spectra library (Powder Diffraction Files) PDF2.
3. Investigation ndings
3.1. Visual examination and stereomicroscopy
Macrographs showing outer and inner tube surfaces are presented in Fig. 1, suggesting that corrosion processes initiated
from inner tube surface. Stereomicroscopic observations indicated accumulation of corrosion deposits (green scale) adjacent
to the leakage waterside region (Fig. 2).
3.2. Optical microscopy and SEM/EDS evaluation
Optical microscopy and SEM evaluation showed the complex layered structure of the corrosion pit close to the leakage
(Figs. 37), indicating that the pitting corrosion initiated from waterside area and progressed towards outside surface lead-
ing to wall disruption and nally leakage. EDS microanalysis identied the exaggerated presence of chlorine attributed to the
presence of chloride salts; sulfur, phosphorus and silicon were also detected mainly as upper layers on pit membrane (Figs. 5,
6 and 8). Chloride salts area principally found on the oor of the pit above the cuprite (Cu
2
O) layer. Pit membrane is a
Fig. 1. Macrographs showing (a) the as received samples, (b) outer surface of the tube at the leakage area and (c) inner surface of the tube. The presence of
corrosion products in the inner tube area is noticeable.
T. Tzevelekou et al. / Engineering Failure Analysis 33 (2013) 176183 177
complex sandwich layer consisting principally of cuprite crystals and sulfates, while high silicon content was identied on
the upper scale layer (Figs. 79).
Microstructure of copper tube consists of highly deformed a-grains exhibiting bent twins, demonstrating a hardness level
up to -140 HV, typical of hard temper (Figs. 3b and c and Fig. 4c and d). Microstructure is consistent to signicantly
deformed copper by drawing without the realization of nal recrystallization annealing treatment (hard temper). No other
Fig. 2. Optical stereomicrographs showing details of the corrosion scale adjacent to the leakage region (a and b) outer tube surface, (c and d) inner tube
surface. Note the accumulation of the green corrosion scale at the inner surface suggesting that corrosion likely started from the waterside and progressed
towards the outside surface, inducing staining and post-leakage corrosion.
Fig. 3. (a) Optical micrograph of transverse section depicting the overall pitting development adjacent to the leakage region. Note the heavy accumulation
and various layers of corrosion products. (b) Optical micrograph showing the microstructure close to the corrosion pit interface, (c) detail of (b).
178 T. Tzevelekou et al. / Engineering Failure Analysis 33 (2013) 176183
evidence of microstructural aw or discontinuity that could be related to the damage was retrieved. Microstructural obser-
vations close to the inner pitted interface did not bring any substantial evidence related to localized abnormality that might
be connected to the failure (Fig. 3b and c).
Fig. 4. Optical micrographs showing (a) the sandwich structure of corrosion deposits consisting of successive layers of oxides (cubic crystals) and scale
products and (b) detail of (a), (c and d) the cold worked microstructure consisting of elongated a-crystals and bend twins; microhardness reached
approximately -140 HV typical for hard temper (immersion etching in FeCl
3
solution).
Fig. 5. (a) SEM micrograph showing the complexity of the pitting structure, (b) detail of the metal/corrosion products interface showing the presence of
cubic oxide crystals corresponding mainly to cuprite (Cu
2
O), (c) principal scale layer of thickness -200 lm at the examined section demonstrating brittle
behavior, (d) detail of (c).
T. Tzevelekou et al. / Engineering Failure Analysis 33 (2013) 176183 179
3.3. XRD evaluation
According to the XRD analysis, the corrosion products consisted mainly of cuprite: Cu
2
O, malachite: Cu
2
(OH)
2
CO
3
and
brochantite: Cu
4
SO
4
(OH)
6
. Secondarily, traces of atacamite Cu
2
Cl(OH)
3
were also identied (Fig. 10). These results come
in agreement with the macroscopical green scale of the corroded areas and the ndings of SEMEDS analysis.
4. Discussion and synopsis
According to BS-EN-12502-2 [13] pitting corrosion of copper in heated water is mainly caused by the water composition
and the water temperature and is restricted to acidic waters (pH < 7) with relatively low bicarbonate (<1.5 mmol/l) and high
sulfate levels. In such waters, the corrosion likelihood is increased if the ratio S = [HCO

3
[=[SO
2
4
[ < 1:5.
Brochantite due to better stability at high temperatures is commonly found in hot water pipes or hot water recirculation
systems. However, its formation is also dependent upon the concentration of bicarbonates and pH. In general, for waters
with high sulfate to bicarbonate ratio, and relatively low pH, pitting is likely in hot waters due to the formation of brochan-
tite [12].
On the other hand Malachite, is comparative less stable at these temperature ranges, however increase in concentration
of bicarbonate at a lower pH (<7.4) may result in its formation [12]. Malachite is mostly found in cold water pitting, referred
also as Type I-Pitting forming a sandwich layer above cuprite crystals. Moreover, hydroxyl-carbonates deposits, alternated
with cuprite layers, display a characteristic spherical nodule morphology, consisting of radiating needle crystals. Although
malachite was detected by XRD, strong evidence of these deposits was not obtained in the perforation area, in the context
of the present microscopic investigation.
The presence of Atacamite indicates a chloride/chlorine source possibly coming from the water treatment. The corrosivity
in cases of chlorinated water is enhanced by low pH because of the greater oxidizing strength of hypochlorous acid (HClO)
(favored at low pH), over the hypochlorite ion (OCl

) [14]. Looking at a semi-quantied evaluation of the spectrum, the fact

that only traces of atacamite could be identied could possibly be attributed to the sampling area of the corrosion products
(upper part of the pit cap while chloride salts are found at the bottom of pits) in conjunction to the propagation behavior of
Fig. 6. EDS spectrum of selected area of the main corrosion layer (depicted schematically by the framed area in Fig. 5), showing exaggerated presence of Cl
attributed to probably to chlorides. Oxides are also present, while minor peaks of Si and C were detected.
Fig. 7. (a) SEM micrograph showing various complex corrosion deposits corresponding probably to phosphates overlying cuprite, (b) EDS spectrum
corresponding to dark-crystal point microanalysis.
180 T. Tzevelekou et al. / Engineering Failure Analysis 33 (2013) 176183
chlorine ions in pitting corrosion mechanisms. Hydrolysis of chlorides has been postulated as acid formation reaction, aggra-
vating the chemical attack conditions (reaction (1)). The Hypochlorous acid ionizes in the second step (reaction (2)), yielding
hydrogen ion [H
+
] and hypochlorite ion [OCl

]. The dissociation rate of Hypochlorous acid decreases at lower pH (<7.5)

values.
Cl
2
H
2
O HClO HCl (1)
Fig. 8. (a) SEM micrograph showing an area of the waterside corrosion layered structure (membrane area), (b) detail of (a) boxed area. A, B and C are
indicative points of point area microanalysis for corrosion deposit identication.
Fig. 9. EDS spectra at the various indicated areas marked in Fig. 8: (a) A-area (high-S and O, low Si), (b) B-area (high-O, moderate Si), (c) C-area (high-Si and
O).
T. Tzevelekou et al. / Engineering Failure Analysis 33 (2013) 176183 181
HClO OCl

(2)
Although chlorine is related to pH reducing tendency, it may attack copper even at higher pH levels [15].
The above ndings indicated that:
v The leakage was resulted from inside-out pitting corrosion that led to tube perforation. High chlorine content likely
stimulated the pit initiation and growth. It is considered essential for pitting attack by aiding the breakdown of pro-
tective surface layers and inducing to the formation of wide, shallow pits [16].
v The composition of owing medium had likely contributed to the formation of fragile and loosely adherent scale
products, primarily mixed (hydroxy)chlorides, such as CuCl
2
3Cu(OH)
2
, and secondarily sulfates, phosphates, that
they can easily be detached and sustain the corrosion process, see also Ref. [17]. Similar ndings concerning pit struc-
ture and contents are reported in Ref. [18].
v XRD ndings suggested also that the corrosion could be correlated to the characteristics of the water, more speci-
cally relatively low pH and presence of elevated chlorine levels, while the bicarbonate concentration factor cannot
be discounted.
v The pit structure and localized chemistry (cuprite and brochantite) likely indicate the occurrence of hot-water pitting
corrosion, termed also as Type II-Pitting (normally under pH < 7 and T > 60 C) and low bicarbonate/sulfate ratio
[13,19].
v The presence of Si could be attributed to water contamination with SiO
2
coming from dust, soil or cement construc-
tions or from water ltration process, while phosphates may be originated from the addition of polyphosphate chem-
icals for corrosion prevention. Sulfates and chlorine are probably coming from water treatment (such as coagulation
and disinfection processes). Chlorides and sulfates are corrosive constituents for copper pipe components, while chlo-
rides inuence is considered much stronger than that of sulfates [20].
5. Conclusions
Copper water tube (52 1 mm) experienced leakage after being 6 years in service, due to Type-II Pitting Corrosion
(Hot Water Pitting). The investigation ndings suggest strongly that the leakage resulted from waterside driven pitting
corrosion. Microanalysis conducted on the pitted area adjacent to the leakage indicated high content of chlorine
attributed to chloride salt deposits as well as various other substances, such as sulfates and phosphates. The
corresponding crystalline phases, such as brochantite and cuprite, that were veried also by XRD, advocated the action
of Type II-Pitting mechanism (hot water pitting corrosion) under low pH and [HCO

3
[=[SO
2
4
[ ratio. The presence of high
chlorine content is devastating, since it is consistent to the aggressive and acidic nature of the owing medium
coming probably from water treatment, constituting favorable conditions for localized pitting corrosion. No other man-
ifestation of microstructural abnormality or other discontinuity was able to be identied that could be related to the
failure.
Water quality (mainly physicohemical parameters) should be carefully monitored and regularly controlled during pipe-
line operation, as well as sufcient maintenance of the installation and appropriate water treatment. Minimization of water
stagnation periods should be also attempted in order to assist to the extension of the installation lifetime.
Fig. 10. X-ray Diffraction pattern of corrosion products.
182 T. Tzevelekou et al. / Engineering Failure Analysis 33 (2013) 176183
References
[1] Myers JR, Cohen A. Pitting corrosion of copper in cold potable water systems. Mater Perform 1995;34(10):602.
[2] Duffner DH. Air-conditioner failure investigation intergranular cracking in a pure copper condenser tube. J Failure Anal Prevention 2005;5(1):7985.
[3] Pantazopoulos G, Tsinopoulos G. Corrosion of a copper U-shaped heating element: some microstructural and morphological observations. J Failure
Anal Prevention 2006;6(6):812.
[4] McDougall JL, Stevenson ME. Stress-corrosion cracking in copper refrigerant tubing. J Failure Anal Prevention 2005;5(1):137.
[5] Miyamoto H, Saburi D, Fujiwara H. A microstructural aspect of intergranular stress corrosion cracking of semi-hard U-bend tubes of commercially pure
copper in cooling systems. Eng Failure Anal 2012;26:10819.
[6] Pantazopoulos G, Vazdirvanidis A, Tsinopoulos G. Failure analysis of a hard drawn water tube leakage caused by the synergistic actions of pitting
corrosion and stresscorrosion cracking. Eng Failure Anal 2011;18(2):64957.
[7] Stevenson ME, Barkey ME, McDougall JL. Stresses in bent copper tubing: application to fatigue and stresscorrosion cracking failure mechanisms. J
Failure Anal Prevention 2005;5(6):259.
[8] Kuznicka B. Erosioncorrosion of heat exchanger tubes. Eng Failure Anal 2009;16(7):23827.
[9] Olszewski AM. Avoidable MIC-related failures. J Failure Anal Prevention 2007;7(4):23846.
[10] Klug HP, Alexander LE. X-ray procedures for polycrystalline and amorphous materials. 2nd ed. 1954, 1974.
[11] Connolly James R. Introduction to X-ray powder diffraction. Spring; 2007.
[12] Farooqi Owais E. An assessment and modeling of copper plumbing pipe failures due to pinhole leaks. Thesis submitted to the Faculty of the Virginia
Polytechnic Institute and State University in partial fulllment of the requirements for the degree of Master of Science in Civil Engineering. May 2006.
[13] BS EN12502-2:2004. Protection of metallic materials against corrosion guidance on the assessment of corrosion likelihood in water distribution and
storage systems. Part 2: Inuencing factors for copper and copper alloys.
[14] Abigail F Cantor, Jae K Park, Prasit Vaiyavatjamai. The effect of chlorine on corrosion in drinking water systems. Final report, The Midwest Technology
Assistance Center for Small Public Water Systems (MTAC). 2000.
[15] Jones DA. Principles and prevention of corrosion. 2nd ed. NJ: Prentice Hall; 1996.
[16] Fernandes PJ. Type I pitting of copper tubes from a water distribution system. Eng Failure Anal 1998;5(1):3540.
[17] ASM Handbook. Corrosion, vol. 13. 4th ed, ASM International; 1992.
[18] Lytle DA, Nadagouda MN. A comprehensive investigation of copper pitting corrosion in a drinking water distribution system. Corros Sci
2010;52:192738.
[19] Edwards M, Ferguson JF, Reiber SH. The pitting corrosion of copper. J Am Water Works Assoc 1994;87:7489.
[20] Jia Z, Du C, Zhang D, Li X. Analysis of copper pitting failure in drinking water distribution system. J Failure Anal Prevention 2011;11(2):1527.
T. Tzevelekou et al. / Engineering Failure Analysis 33 (2013) 176183 183