Corrosion and Corrosion Prevention in Heat Exchangers

Corros Rev 2019; aop
Willem Faes*, Steven Lecompte, Zaaquib Yunus Ahmed, Johan Van Bael, Robbe Salenbien,
Kim Verbeken and Michel De Paepe
Corrosion and corrosion prevention in heat

exchangers
https://doi.org/10.1515/corrrev-2018-0054
Received June 22, 2018; accepted December 13, 2018
1 Introduction
Abstract: In many industries and processes, heat exchang- Heat exchangers are devices used to transfer thermal
ers are of vital importance as they are used to transfer heat energy over a heat transfer surface from one fluid to a
from one fluid to another. These fluids can be corrosive second fluid at a lower temperature (Kakaç et al., 2012).
to heat exchangers, which are usually made of metal- Several types of heat exchangers exist, but the most
lic materials. This paper illustrates that corrosion is an common types (the focus of this paper) are shell-and-tube
important problem in the operation of heat exchangers in heat exchangers and plate heat exchangers.
many environments, for which no straightforward answer In a shell-and-tube heat exchanger, a number of tubes
exists. Corrosion failures of heat exchangers are common, are mounted in a cylindrical shell, parallel to the axis of
and corrosion often involves high maintenance or repair the shell (Kakaç et al., 2012). Typically, several vertical
costs. In this review, an overview is given of what is known plates, baffles, are inserted in the shell to redirect the flow
on corrosion in heat exchangers. The different types of over the tubes and to improve the heat transfer. A sche-
corrosion encountered in heat exchangers and the suscep- matic representation is given in Figure 1, where a cold
tible places in the devices are discussed first. This is com- fluid flows through the tubes and a hot fluid flows through
bined with an overview of failure analyses for each type the shell.
of corrosion. Next, the effect of heat transfer on corrosion Plate heat exchangers consist of a number of thin
and the influence of corrosion on the thermohydraulic plates, mounted behind each other in a frame (Kakaç
performances are discussed. Finally, the prevention and et al., 2012). A gasket is placed between two adjacent
control of corrosion is tackled. Prevention goes from gen- plates to determine the flow path and to prevent the fluid
eral design considerations and operation guidelines to the from leaking to the environment. The plates are typically
use of cathodic and anodic protection. corrugated to improve the heat transfer. The working prin-
ciple is illustrated in Figure 2.
Keywords: corrosion; geothermal; heat exchangers; heat
Heat exchangers play a key role in the operation of
transfer.
various industrial processes like the oil and gas industry,
the chemical and pharmaceutical industry, petroleum
refining, and the pulp and paper industry. In each of these
industries, the heat exchanger can be subject to a variety
of environments. Some of them are known to be aggres-
sive, while others are often not considered to be corrosive.
It was calculated in 2001 that the direct cost of all types of
*Corresponding author: Willem Faes, Department of Flow, Heat corrosion in the United States amounts annually to $276
and Combustion Mechanics, Ghent University, Ghent, Belgium; and billion or 3.1% of the gross domestic product of the country
Flemish Institute of Technological Research (VITO), Mol, Belgium, (Koch et al., 2002). Of these costs, $17.6 billion are allocated
e-mail: willem.faes@ugent.be
to production and manufacturing businesses, under which
Steven Lecompte and Zaaquib Yunus Ahmed: Department of Flow,
Heat and Combustion Mechanics, Ghent University, Ghent, Belgium the aforementioned industries are categorised. For electri-
Johan Van Bael and Robbe Salenbien: Flemish Institute of cal power generation, in which heat exchangers are also
Technological Research (VITO), Mol, Belgium; and EnergyVille, frequently used, the annual corrosion cost is estimated
Thor Park, Genk, Belgium in the same study at $6.9 billion. In a more recent study
Kim Verbeken: Department of Materials, Textiles and Chemical
by NACE International (2016), it was calculated that the
Engineering, Ghent University, Ghent, Belgium
Michel De Paepe: Department of Flow, Heat and Combustion
annual global cost of corrosion amounts to $2.5 trillion.
Mechanics, Ghent University, Ghent, Belgium; and Flanders Make, According to other sources (Electric Power Research
Core Lab UGent-EEDT, www.flandersmake.be, Lommel, Belgium Institute, 2001; Syrett & Gorman, 2003; Roberge, 2008),
Authenticated | willem.faes@ugent.be author's copy

Download Date | 3/14/19 10:31 AM
2 W. Faes et al.: Corrosion and corrosion prevention in heat exchangers
In addition to these economic issues, Davis (2000) also

gave some examples of social consequences:
–– Safety (e.g. risk for explosions);
–– Health (e.g. escaping of a poisonous product from a
corroded heat exchanger);
–– Depletion of natural resources;
–– Visually unpleasant appearance;
–– Environmental contamination.
That a failing heat exchanger can pose a potential safety

threat was illustrated by Collier (1983). According to the
Figure 1: Schematic representation of a shell-and-tube heat author, explosion of boilers (a thermal system analogous to
exchanger. a shell-and-tube heat exchanger) used in American steam-
boats occurred at least 233 times between 1816 and 1848,
which in total caused the death of 2560 passengers and left
the total cost of corrosion for the electrical power industry 2100 others injured. One major accident was the explosion
was $15.4 billion in the United States in 1998. This amount of the Sultana steamboat on the Mississippi River in 1865,
is split into categories, where corrosion of heat exchangers resulting in >1250 deaths (Yager, 1976). Jennings (2015)
takes up 5.55% or $855 million annually. These high losses stated that the boiler in this steamboat was made from
are a combination of direct and indirect costs caused by wrought iron that most likely contained inclusions and
corrosion and can be attributed to several economic con- imperfections resulting from the production techniques
sequences (Davis, 2000), as follows: from that period. These are sites where corrosion could
–– Replacement of corroded equipment; initiate, and therefore Jennings claimed that possibly cor-
–– Oversizing to take possible corrosion into account; rosion at these sites was the root of the tragic incident. The
–– Preventive maintenance; explosion of land-based boilers was a common problem
–– Shutdown of equipment; as well. In his publication, Collier (1983) also emphasised
–– Loss of efficiency (e.g. when corrosion products on the fact that between 1890 and 1904, 5000 incidents of
decrease the heat transfer rate of the heat exchanger); such explosions were reported by one insurance company.
–– Loss of a product (e.g. through leakages or contamina- Although the knowledge about heat exchangers
tion of a product). and the existing production techniques have drastically
improved, accidents caused by corrosion still occur. Two
more recent failures of heat exchangers, illustrating the
potentially catastrophic consequences, were reported
on an offshore gas production platform in 2006 and in a
refinery plant in 2010.
In the first incident, a shell-and-tube heat exchanger,
made from carbon steel clad with titanium, was used to
cool hydrocarbon gas with seawater (Health and Safety
Executive, 2006; Parrott, 2014). Galvanic corrosion
between the carbon steel and titanium cladding caused
the escape of the gas into the seawater, overpressuring the
shell. The shell and the tubes got torn from the tube sheet
and the escaping gas exploded. Luckily, there were only
two minor injuries. The second incident was a carbon steel
shell-and-tube heat exchanger, but with a stainless-steel
type 316 liner [U.S. Chemical Safety and Hazard Investiga-
tion Board (CSB), 2010; Marsh, 2016]. The shell ruptured
at the weld seams, releasing the hot hydrogen and naptha.
Figure 2: Plate heat exchanger with indicated flow paths.
Republished with permission of Taylor & Francis Group LLC, from
The mixture autoignited, resulting in seven fatalities. The
Kakaç et al. (2012); permission conveyed through Copyright root cause of the accident was determined to be high-
Clearance Center, Inc. temperature hydrogen attack (HTHA) of the carbon steel.

W. Faes et al.: Corrosion and corrosion prevention in heat exchangers 3
Both galvanic corrosion and hydrogen damage in heat

exchangers will be discussed later.
In the literature, various failure analyses can be found
where heat exchangers suffered corrosion. According to
the authors’ knowledge, no paper gives an overview or
describes the particularities of the corrosion processes
occurring in heat exchangers. Therefore, this study gives
a non-exhaustive overview of the failures, to show which
types of corrosion are common and which places are sus-
ceptible. As the research on either the influence of heat
Figure 3: Schematic representation of the corrosion process of iron-
transfer on corrosion, or vice versa, is scattered, this is
based alloys in aqueous solutions.
described in a next section. Based on recommendations
given in the failure analyses and heat exchanger design
handbooks, guidelines are provided on how to prevent cor- 2.1 Uniform corrosion
rosion in heat exchangers. Additionally, two more special-
ised corrosion protection techniques, cathodic and anodic When corrosion is homogeneously spread over the entire
protection, are discussed. A great deal of the information heat transfer surface, the term “uniform corrosion” is
about these methods was found in the patent literature. used. Uniform corrosion is usually easy to detect visually.
The many failures discussed in this paper demo- The thinning of the tube or plate increases the stresses in
nstrate that corrosion is a challenging problem in the the metal, which might eventually lead to failure (Wang
operation of heat exchangers. Most failures are caused by et al., 2007). One strategy to avoid this type of corrosion
localised forms of corrosion, such as pitting or stress cor- is by selecting a material that passivates in the specific
rosion. To prevent this, careful material selection, design environment. Alternatively, by doing exposure tests prior
phase, and operation are absolute prerequisites to guar- to the use of the heat exchanger, the corrosion rate can
antee a long lifetime of the heat exchanger. Electrochemi- be determined (expressed, e.g. in mm/year), which allows
cal techniques exist to limit corrosion of heat transfer designing the heat exchanger with an increased wall
surfaces (cathodic and anodic protection). Their applica- thickness (Revie & Uhlig, 2008).
tion in heat exchangers is, however, not straightforward. This does not imply that uniform corrosion does not
pose any problems. For example, different uniform cor-
rosion rates in different parts of the exchanger are prob-
2 T
ypes of corrosion encountered lematic for overhauling. In the case described by Xie et al.
(2015), the carbon steel tubes of a vertical shell-and-tube
in heat exchangers heat exchanger, used to cool a stripper overhead vapour,
had to be replaced after only 2.5 years of service (Xie
Corrosion is a term used to describe the (unwanted) deg- et al., 2015). Around the tubes, cooling water flowed,
radation of a metal when exposed to an environment, due which cooled the process vapour inside the tubes. Brown
to interaction with this environment. Most of the exam- and reddish rust was found on both internal and exter-
ples discussed below involve aqueous environments. The nal surfaces, and severe wall losses (>70% for 159 of the
corroding metal in the liquid is called the anode, and the 1410 tubes) were detected. Figure 4 shows a cross section
metal ions go into solution following the anodic oxidation of different tubes. Signs of corrosion can be seen on both
reaction. Corrosion is only possible if the electrons result- external and internal surfaces, and the difference in wall
ing from this reaction are consumed by a cathodic reaction thickness can clearly be observed. The authors discov-
(e.g. the reduction of dissolved oxygen). An electrically ered that the corrosion was most severe just underneath
conductive path must exist between the two locations the top tube sheet. This was explained by the accumula-
for the electrons to pass (Edwards, 2013). This corrosion tion of non-condensable gases at the top of the shell area,
process is illustrated in Figure 3. Many types of corrosion causing the concentration of corrosive elements (e.g. chlo-
are possible, depending on the material, the environment, rides) in the water on the hot tube surfaces at the vapour-
and the working conditions (Roberge, 2000). In the next liquid interface.
sections, the most prevalent forms encountered in heat The uniform corrosion was more severe at the cooling
exchangers are discussed and several examples of failures water side than at the internal surface of the tubes.
of heat exchangers, caused by corrosion, are presented. However, uniform corrosion was not the only type of

as the application of a coating) are often susceptible to

pitting (Revie & Uhlig, 2008). At mechanical defects of
this protective layer, pits are initiated, and when these
grow sufficiently, they might perforate the heat transfer
surface. The initiation of pitting is influenced by metal-
lurgical and structural factors, environmental factors,
polarisation phenomena, and the presence of corrosion
products. For stainless steel, high chloride content, high
temperatures, and low pH are promoters of pitting corro-
sion. Some metals often used in the construction of heat
exchangers whose protective film might locally break
down, and as such initiate pitting, are other passivating
metals such as aluminium, nickel, titanium, and their
alloys (Kuppan, 2000).
The tendency of a stainless steel to be susceptible to
pitting corrosion can be predicted with the pitting resist-
ance equivalent number (PREN). This number can be
Figure 4: Section of the tube with severe uniform corrosion and wall
calculated with Eq. (1) based on the mass fraction (in per-
thickness reduction. centage) of chromium (Cr), molybdenum (Mo), tungsten
Reprinted with permission from NACE International, from Xie et al. (W), and nitrogen (N) in the alloy (Roberge, 2000). The
(2015). common type of stainless steel, type 304, has a PREN of
19, while highly alloyed types of stainless steel can go to a
PREN of >50. For seawater applications at 20°C, a PREN of
corrosion observed inside the tubes; localised forms of at least 32 is advised (Veldkamp et al., 2016).
corrosion, like pitting corrosion (Section 2.2) and erosion
PREN = Cr + 3.3 × (Mo + 0.5W) + 16N. (1)
corrosion (Section 2.5), were also detected.
The recommendations made by the authors were
The shell-and-tube heat exchanger discussed in the failure
adaptation of the water discharge to avoid the accumu-
analysis by Julian et al. (2015) was used in a geothermal
lation of non-condensable gases and a different type of
plant. Deep pits, of which some perforated the tube, were
steel.
discovered in the middle tubes (the lower and upper tubes
One other example of a failed shell-and-tube heat
were free of pits). Figure 5 shows a cross section of a part
exchanger for which uniform corrosion was one of the main
of the tube where pits are present. It was found that the
causes of failure is an oil cooler described by Eich (2012).
No cases of plate heat exchangers that failed because of
uniform corrosion could be found in the literature. This
low number of incidents, in comparison to the next types
of corrosion, can possibly be attributed to the fact that it
is, in general, easy to account for design or material selec-
tion and to detect uniform corrosion (Roberge, 2008). In
contrast to uniform corrosion, the next types of corrosion
do not include an appreciable weight loss of the metal.
The damage to the material is localised and not always
easy to monitor (Revie & Uhlig, 2008).
2.2 Pitting corrosion
The first form of localised corrosion is pitting corrosion.

Metals covered with a protective film (either reaction Figure 5: Cross section of a tube damaged by pitting corrosion.
products resulting from spontaneous interactions with Reprinted from Julian et al. (2015), with permission from
the environment or created by a surface treatment such International Geothermal Association.

lower and upper tubes were preferentially filled with brine similarity to pitting corrosion and will be discussed in the
and steam, respectively, while the middle tubes received next section.
a mixture of condensing steam and brine. These condi-
tions caused an increasing concentration of acid gases
like CO2 and H2S in the condensing water, making it even 2.3 Crevice corrosion
more acidic. This locally high acidity was suggested as
the reason for the rapid development of pits in only one Crevice corrosion is another intense and dangerous form
row of tubes. Also, elemental arsenic, which precipitated of localised corrosion, caused by a local environment dif-
from the brine, was observed in the pits, which might have ferent from the bulk environment. In crevices or shielded
enhanced corrosion by galvanic effects (galvanic corro- areas with stagnant liquid, a depletion of oxygen can
sion is discussed in Section 2.7). occur. The initiation of crevice corrosion is influenced by
An example of a plate heat exchanger that failed due this differential concentration of oxygen. The presence of
to pitting is described by Deen et al. (2010). In their study, chlorides can also cause the initiation of crevice corrosion.
the failure of a heat exchanger with plates constructed The rate of corrosion increases as the crevice narrows and
with type 316L stainless-steel plates (PREN ≈ 24.2) is the cathode area is increased (Roberge, 2000). In plate
investigated. The device was used in the cooling loop of heat exchangers, such crevices are typically present at the
textile mill. A 316L stainless steel was originally chosen gaskets (metal/non-metal contact), at the plate contact
for its corrosion-resistant properties; however, when the points, and under deposits.
heat exchanger was opened for maintenance after 2000 h The studies by Crum et al. (2007) and Cassagne et al.
of operation, it was observed that several of the plates (2010) illustrate the difficulty of avoiding crevice corrosion
were perforated by pits. An example of a failed plate can in plate heat exchangers for seawater applications. In both
be seen in Figure 6. The authors concluded that a local studies, several metals are tested as plate material at ele-
breakdown of the passive oxide film occurred, which was vated temperatures (up to 70°C) with both laboratory tests
caused by high chloride contents in the cooling water. and exposure to real seawater (with controlled chlorine
Additional examples of failure analyses of shell-and- concentrations, added to suppress marine life) in dummy
tube heat exchangers where pitting was one of the main heat exchangers. The metals tested were super-austenitic
modes of corrosion-related failure can be found in the fol- stainless steels, nickel-chromium-molybdenum alloys, and
lowing references: Xie et al. (2015), Ghayad et al. (2015), titanium. Almost all metals, even titanium, showed signs
Qiankun et al. (2017), Sanders and Iwand (2013), Ifezue of crevice corrosion in the most severe conditions. Only the
and Tobins (2015), Zheng et al. (2014), Song et al. (2013), most recent Ni-Cr-Mo alloys, like alloy 686 [Unified Num-
and Abouswa et al. (2007). Two papers in which pitting bering System (UNS) N06686], could withstand the corro-
corrosion in plate heat exchangers was observed are sive environment. An example of severe crevice corrosion at
described by Crum et al. (2007) and Turissini et al. (1997). the gasket area of a plate made from alloy 31 (UNS N08031)
The main corrosion process in this last case was, however, is shown in Figure 7. More cases of crevice corrosion in plate
crevice corrosion, which is a type of corrosion with some heat exchangers were described by Turissini et al. (1997).
Figure 6: Pitting corrosion damage in a plate heat exchanger.

Reprinted from Deen et al. (2010), with permission from Elsevier.

external, but also residual stresses, e.g. from welding

or cold working (Roberge, 2000). In shell-and-tube heat
exchangers, residual stresses are often found at tube-to-
tube sheet expanded joints and U-bends (Kuppan, 2000).
Residual stresses also often cause SCC in cold-rolled plate
heat exchangers (Wang et al., 2007).
Walker (1990) mentioned that SCC can also be caused
by corrosion products in a constricted or confined space.
The volume of corrosion products is often larger than that
of the base metal, causing a wedging action with high
local stresses. Similar to this wedging is denting, which
occurred in nuclear steam generators (Walker, 1990).
Inconel tubes passed through carbon steel supports that
corroded. The tubes were dented inwards because of
an accumulation of corrosion products in the annulus
between the tubes and the support.
A second kind of corrosion where stresses play a role
Figure 7: Crevice corrosion under the gasket of a plate heat is corrosion fatigue. This is characterised by a fluctuat-
exchanger. ing load substantially below the yield point. The fatigue
Reprinted with permission from NACE International, from Cassagne
behaviour of metals is affected by corrosive media. Figure 8
et al. (2010).
illustrates the influence of corrosion on the fatigue behav-
iour with an S-N curve. It can be seen that failure occurs
In shell-and-tube heat exchangers, the critical areas within fewer cycles for a given stress, and a fatigue limit
for crevice corrosion are clearances between rolled is no longer observed (Revie & Uhlig, 2008). Similar to
tubes and tube sheets, open welds, and bolt holes. Also, SCC, corrosion fatigue is often initiated at various types of
complex geometric designs with sharp edges are prone to stress concentrations (Wang et al., 2007).
crevice corrosion (Kuppan, 2000). The investigations in Tawancy (2009) analysed the failure of a shell-and-
the publications by Yang et al. (2012), Kaul et al. (1995), tube heat exchanger used in an oil refinery. Leaks were
and Allahkaram et al. (2011) give some additional exam- detected in some bends after 48 h of starting the unit after
ples of failure analyses where crevice corrosion was a period of downtime. The conclusion of the author is that
found between the tube and the tube sheet. In the work during the shutdown, polythionic acid was formed from
of Allahkaram et al. (2011), crevice corrosion was also sulphur-bearing deposits and condensed water vapour.
found in gaps between the tubes and the baffles. A last This acid, in combination with internal stresses in the
example is the study on oil coolers by Eich (2012) (already bend, induced the cracks in the metal (type 321 steel),
mentioned in the discussion of uniform corrosion), where shown in the photograph of Figure 9.
crevice corrosion was detected on the tube sheet. The
crevice was, however, not formed between the tube sheet
and the tubes but between the tube sheet and a deflector
in the head of the cooler.
2.4 Stress corrosion
When a corrosive environment is combined with mechan-

ical stresses, the corrosion process is aggravated. A first
type where corrosion and stress join forces is stress cor-
rosion cracking (SCC). This corrosion mechanism involves
a static tensile stress present in a susceptible metal in a
corrosive environment, causing failure where the same
structure would be able to endure the same stress in Figure 8: S-N curve illustrating the influence of corrosion on fatigue
a non-corrosive environment. These stresses can be behaviour.

Figure 9: Stress corrosion cracks induced in a bend by polythionic

acid.
Reprinted from Tawancy (2009), with permission from Elsevier.
Several other examples of stress corrosion causing the Figure 10: Effect of flow velocity on the corrosion rate of carbon
failure of heat exchangers can be found in the literature steel in seawater.
for both shell-and-tube heat exchangers (Otegui & Fazzini,
2004; Guo et al., 2011; Qu et al., 2011; Ravindranath et al.,
prone to erosion corrosion, especially at the tube side,
2012; El-Amoush et al., 2014; Corte et al., 2015; Xu et al.,
where this is often called inlet-tube corrosion (Walker,
2015; Corleto & Argade, 2017) and plate heat exchangers
1990).
(Turissini et al., 1997; Traubert & Jur, 2012; Khodamorad
Two other forms of wear, sometimes classified under
et al., 2016). For two of these cases, it concerns corrosion
erosion corrosion, are “cavitation erosion” and “fret-
fatigue (Otegui & Fazzini, 2004; Ravindranath et al., 2012),
ting corrosion” (Walker, 1990). On the one hand, in the
while in the others, SCC was the main cause of failure.
former, bubbles collapse on the surface. This often occurs
in evaporators, but is also possible in non-boiling condi-
tions at pressure drops, resulting from sudden changes in
2.5 Erosion corrosion velocity. Fretting corrosion, on the other hand, is caused
by a periodic rubbing motion between two metal parts.
Erosion corrosion is the synergistic interaction between This often occurs in shell-and-tube heat exchangers at the
corrosion and erosion, caused by a relative motion between tube-to-tube sheet joints or at the tube-baffle contacts.
a corrosive fluid and a metal wall. The combined effect of This increases the shell-side baffle bypass flow, leading to
the two phenomena is larger than corrosion or erosion a less efficient heat exchanger.
separately. The phenomenon normally occurs under tur- Several examples of erosion corrosion can be found in
bulent flow conditions. Surface material is removed by the literature for shell-and-tube heat exchangers. Erosion
the fluid shear stress, and a protective corrosion film (e.g. corrosion was observed in several studies (Kuznicka,
formed on stainless steel or aluminium) can be destroyed 2009; Eich, 2012; Yang et al., 2012; Li et al., 2013; Klein
by the motion of the fluid. The damage increases with et al., 2014). Also, some examples of cavitation erosion
increasing fluid velocity, and most metals have a criti- (Reitz, 2002; Klenowicz et al., 2003) and fretting corrosion
cal velocity above which the corrosion rate strongly rises (Cassagne et al., 2010; Klein et al., 2014; Shahrani & Al-
due to the destruction of the protective film (Wang et al., Subai, 2014) could be found. An example of a plate heat
2007). In seawater, where there is a high chloride con- exchanger that failed because of fretting corrosion was
centration, stainless steel might not be able to establish described by Wassilkowska et al. (2016).
passivity, causing the corrosion rate to increase with flow In the case described by Kuznicka (2009), pits were
velocity (Revie & Uhlig, 2008). A relationship between the detected on the outside surface of the copper tubes in a
flow velocity of seawater and the corrosion rate of carbon shell-and-tube heat exchanger. Some of these pits were
steel was proposed by LaQue (1948). This relationship was perforating the tube wall. Around the tubes, industrial
fitted to an exponential function by Guedes-Soares et al. water was flowing with an inlet temperature of 16°C. The
(2011), as shown in Figure 10. inside surface of the pits was smooth and free of corrosion
In plate heat exchangers, erosion corrosion can occur products. The pits were positioned in the longitudinal
at the inlets and is more severe when the fluid contains direction of the tube and appeared as craters undercutting
abrasive solids such as sand particles (Wang et al., 2007). the metal. These observations led the author to believe
In shell-and-tube heat exchangers, the inlets are also that the pits were created by erosion corrosion. Also, the

fact that the water contained some solids in suspension 2.6 Intergranular corrosion
was thought to be the possible cause of the erosion corro-
sion. Most of the tubes on which the pits were discovered In solid state, metals typically have a crystalline structure.
were laying in the same row: the row closest to the baffle Carbon steel has a body-centred cubic crystal structure;
edge. According to the author, the turbulence was most aluminium, copper, and austenitic stainless steel are
intense at this location, increasing the effect of erosion face-centred cubic; and magnesium has hexagonal close-
corrosion. packed crystal lattice. During the solidification process,
An example of a shell-and-tube heat exchanger that these crystals are formed as individual grains where the
failed because of fretting corrosion is discussed by Klein boundaries between the different grains are chemically
et al. (2014). In a wastewater plant, hot vapours were used more active (Walker, 1990).
on the shell side. These vapours, however, contained Narrow regions along the grain boundaries can be
some debris, fouling the outer surface of the tubes. This corroded with only negligible attack on the grains them-
blocked the flow path in some areas of the heat exchanger, selves. This can cause the metal to disintegrate or lose
causing the fluid velocity to be higher in other regions. An its strength. Intergranular corrosion often occurs on
oscillating behaviour of the tubes resulting from these welded structures of the widely used type 304 stainless
high velocities caused the tubes to be in contact with each steel. Operation at high temperatures, like an improper
other, inducing fretting corrosion on each other. Figures 11 heat treatment or welding, can cause the depletion of
and 12, respectively, show friction marks on the outside of the alloying element chromium near the grain bounda-
the tubes and worn-out holes in the baffle plates, created ries in a process called sensitisation. These chromium-
by the vibrating motion of the tubes. depleted regions have lost their corrosion resistance,
Only one case of erosion corrosion (in the form of fret- and what remains are the characteristics of the base
ting corrosion) could be found for plate heat exchangers. metal (Kuppan, 2000). The chromium forms chromium
The corrosion process was described by Cassagne et al. carbides, because at these temperatures, the carbides
(2010), who investigated materials for plate heat exchang- are insoluble and for them to precipitate at the grain
ers. The fretting corrosion occurred at the contact points boundaries, they must obtain chromium from the sur-
of adjacent plates. rounding metal. When the corroding metal is under a
tensile stress, corrosion cracks can form along the grain
boundaries. This phenomenon is termed intergranular
SCC (Roberge, 2000).
From the examples given in the section on stress
corrosion, the cracks discussed in several publications
were intergranular (Turissini et al., 1997; Guo et al., 2011;
Corleto & Argade, 2017). Intergranular cracks (without the
Figure 11: Friction marks on the outside of heat exchanger tubes.
Reprinted from Klein et al. (2014), with permission from Elsevier. application of stresses) were found in the cases discussed
by Ifezue and Tobins (2015) (together with pitting) and
Chandra et al. (2010). Figure 13, from the work of Ifezue
and Tobins (2015), clearly shows intergranular corrosion
of a section of an aluminium tube, which typically results
Figure 13: Intergranular corrosion of an aluminium tube.

Figure 12: Worn-out holes in a baffle plate. Reprinted by permission from Springer Nature: Journal of Failure
Reprinted from Klein et al. (2014), with permission from Elsevier. Analysis and Prevention, Ifezue and Tobins (2015), Copyright 2015.

from inappropriate heat treatment, e.g. during material 2.8 Selective leaching
production.
Selective leaching or de-alloying is a form of galvanic cor-
rosion on a microscopic scale where one alloying element
2.7 Galvanic corrosion is preferentially attacked over the other alloying elements
in the matrix material. As a result, the remaining structure
When two dissimilar metals are electrically connected and is weakened (Roberge, 2000). A common example is the
are in contact with an electrolyte, a galvanic cell is created removal of zinc from brass (de-zincification) in stagnant
and a potential difference exists over the two metals. In waters. Also, the removal of aluminium from aluminium
this bimetallic couple, the metal with the lowest poten- brass in acidic solutions or nickel from cupronickel alloys
tial, the anode, will start to corrode. The more noble metal under conditions of high heat flux (respectively called de-
will behave as a cathode, and its surface will remain unaf- aluminiumification and de-nickelification) is reported in
fected (Roberge, 2000). In this context, it must be added industry (Kuppan, 2000). Addition of appropriate alloy-
that the potential of a metal depends on the characteris- ing elements is frequently quoted to be a viable strategy
tics of the electrolyte as well. Consequently, it is possible against selective leaching. For example, de-zincification
that for two dissimilar metals, one metal is anodic with of brass can be prevented in most aggressive environ-
respect to the other in one electrolyte, while it behaves ments by adding arsenic as alloying element to the brass
cathodic with respect to the same other metal in a differ- (Ravindranath et al., 2012).
ent electrolyte. Galvanic corrosion can also happen on a According to Wang et al. (2007), failures caused by
single metal, caused by local imperfections, local differ- selective leaching are not common in plate heat exchang-
ences in the chemistry of the electrolyte (e.g. an oxygen ers, as they can be avoided by selecting an appropriate
concentration cell), or metallic deposits on the surface material for each application. One example was found in
(Wang et al., 2007). the literature where cupronickel tubes in a shell-and-tube
Galvanic corrosion is common in shell-and-tube heat heat exchanger experienced de-nickelification (Cincera
exchangers at the contacts between the tubes and the et al., 2012). The failure analyses in publications by Qu
baffles or tube sheets when different metals are used for et al. (2011), Ravindranath et al. (2012), Li et al. (2013), and
those parts. Also, the contacts between the baffle and Ranjbar (2010) mention de-zincification as one of multiple
the shell are sensitive areas (Kuppan, 2000). In plate causes of failure of a shell-and-tube heat exchanger.
heat exchangers, galvanic corrosion can exist at welded, In the example discussed by Ranjbar (2010), tubes
brazed, or soldered joints (Wang et al., 2007). One spe- made of yellow brass were used in a shell-and-tube heat
cific example of galvanic corrosion was already men- exchanger. In the tubes, circulating cooling water flowed
tioned in Section 1. Parrott (2014) described the failure of to condense the steam on the shell side. Several tubes
a shell-and-tube heat exchanger made from carbon steel experienced severe fouling and even plugging, causing
with a titanium cladding on an offshore gas production the velocity to increase in the tubes that were not plugged.
platform. In this case, the carbon steel was anodic to the The flow in these tubes was rapid with a high turbulence
titanium and corroded. Another example of galvanic cor- intensity. Additionally, the metal was weakened, because
rosion where carbon steel was anodic to titanium was zinc was leached from the copper alloy. This weaken-
discussed by Yang et al. (2012) in a failure analysis of a ing, in combination with the turbulent motion, caused
shell-and-tube heat exchanger (caused by a combina- severe erosion. A section that de-zincified can be seen in
tion of corrosion mechanisms) in a nuclear power plant. Figure 14. The original brass had a yellow colour, while
One last illustration of the need to avoid combining dis- the de-zincified area has a red colour, as one typically
similar metals is the failure analysis of Mousavian et al. observes for pure copper.
(2011). In a shell-and-tube heat exchanger, used to cool
oil with cooling water, the tubes, tube sheet, and gaskets
were made of copper, carbon steel, and aluminium, 2.9 Hydrogen damage
respectively. After being used for only 4 months, the oil
and water mixed because both the gaskets and the tube Hydrogen can cause damage of a metal in different ways.
sheets corroded. Galvanic corrosion was determined to be Atomic hydrogen H (not the molecule H2) is the smallest
the main cause of failure. The concept of galvanic corro- atom and is often able to diffuse in the metal lattice inter-
sion is rather well known, so the mistake to bring different stitially at a measurable rate, along grain boundaries and
metals into contact with each other is only seldom made. along dislocations. Dislocations often play a crucial role

steel with a stainless-steel cladding and used to heat

naptha for the removal of sulphur, nitrogen, and oxygen
impurities. The formation of cracks by HTHA occurred
near the welds and was facilitated because the shell did
not receive an appropriate post-weld heat treatment.
2.10 Microbiologically induced corrosion
Microbiologically induced corrosion (MIC) is described in

a paper by Dexter (2003). Microorganisms, including bac-
teria, algae, and fungi, are encountered not only in almost
Figure 14: De-zincification of the inner surface of a yellow brass tube. every natural aqueous environment but also in several
Reprinted from Ranjbar (2010), with permission from Elsevier. industrial fluids. They are, for example, very common in
cooling water systems, where they reduce the heat trans-
fer rates and increase the pressure drop over the con-
in hydrogen-related failure (Depover & Verbeken, 2018). denser (Rao & Nair, 1998). When these organisms attach to
Atomic hydrogen can be produced by the cathodic reac- a surface, a biofilm is formed that will influence corrosion.
tion of the corrosion process, while the formation of H2 The microorganisms can initiate corrosion, which would
during the cathodic reaction can be caused by the pres- not occur in the same environment without microorgan-
ence of certain chemical species (so called “poisons”, like isms. The biofilm can also change the type of corrosion
H2S) (Roberge, 2000). (e.g. from uniform to localised corrosion) or increase or
A first form of corrosion damage caused by atomic decrease the corrosion rate. It is, however, also possible
hydrogen is hydrogen blistering. This is caused by atomic that the biofilm does not influence corrosion. The possi-
hydrogen collecting at internal defects, close to the ble influences of microorganisms on corrosion are sum-
sample surface. In these cavities, it recombines to gaseous marised by Dexter (2003) and schematically represented
H2, building up very high internal pressures. This causes in Figure 15.
blisters to form on the metal surface. Another form of A failure analysis was done by Huttunen-Saarivirta
hydrogen damage is hydrogen-induced cracking (HIC), et al. (2012) on a stainless-steel (type 304) shell-and-tube
where the metal is embrittled by the present hydrogen. heat exchanger that was used for only 36 months. On the
When stressed, the normally ductile metal will fail with surface of the shell, where the cooling water flows, corro-
a significantly reduced elongation. This form of corrosion sion tubercles (Figure 16) were found. Some of them pen-
is called sulphide stress cracking when the environment etrated the metal (with an original thickness of 6.2 mm),
in which the HIC occurred contains H2S (Roberge, 2000). while most of them had a depth of 2–4 mm. The investi-
At elevated temperatures (>232°C) and high pressures (>7 gation showed that the metal had both a major austenite
bar), molecular hydrogen can dissociate to atomic hydro- phase and a minor ferrite phase. There was a micro-scale
gen on steel surfaces and diffuse into the steel (Groysman,
2017). At grain boundaries, dislocations, and inclusions
(such as carbides) in the steel, it can form methane (CH4).
This process, called HTHA, decarburises the steel, reduc-
ing its mechanical properties.
Titanium is susceptible to hydrogen damage due to
hydride formation, as illustrated by the paper by Yang
et al. (2012). In this study, the leakage of titanium tubes of
a heat exchanger in a nuclear power plant is investigated.
Some galvanic corrosion and crevice corrosion produced
hydrogen, leading to HIC.
HTHA was indicated to be the cause of the catastrophic
failure and explosion of a heat exchanger in the refinery
plant mentioned in Section 1 (CSB, 2010; Marsh, 2016).
The shell-and-tube heat exchanger was made from carbon Figure 15: Influence of a biofilm on the corrosion rate.

T w,h T w,c
T b,h T b,c
Rconv,h Rcond Rconv,c
Figure 17: Thermal resistance network for heat transfer from a hot

to a cold fluid over a heat transfer surface.
This figure shows the bulk (Tb) and wall (Tw) temperature
on the hot (h) and the cold (c) sides, the two convective
resistances (Rconv), and the conductive resistance (Rcond).
The kinetics of electrochemical reactions (such as
corrosion of a metal surface) normally increase with
increasing temperature. An exponential relationship,
which is reflected in the Arrhenius equation, usually
Figure 16: Tubercles observed on the shell surface of a heat
exists between temperature and reaction rates (Silverman,
exchanger. 2003). This implies that when a corrosive fluid is heated,
Reprinted from Huttunen-Saarivirta et al. (2012), with permission the temperature in the boundary layer near the wall is
from Elsevier. higher and corrosion kinetics will be faster than kinetics
driven by the bulk temperature (Tw,c > Tb,c). However, when
a corrosive fluid is cooled, the metal temperature will be
surface roughness at this ferrite phase, giving bacteria the
lower than the bulk fluid temperature (Tw,h < Tb,h) and the
possibility of attachment, subsequently leading to oxygen
corrosion kinetics are reduced (Moore & Smith, 1967). A
concentration cells. These cells locally induced corrosion
schematic illustration of the temperature profile over the
and created the tubercles.
fluids, the boundary layers, and the surface is given in
In the literature, more examples can be found of
Figure 18.
heat exchangers that failed due to MIC (Rao & Nair, 1998;
The local differences in temperature also cause differ-
Abraham et al., 2009; Sharma, 2014; Rizk et al., 2017).
ent solubility products of substances in the fluid. For most
species, the solubility increases with increasing tempera-
ture, while some have inverted characteristics (Stumm &
3 Corrosion and heat transfer Morgan, 2012). Corrosion products are typically among
the first category. As explained, the temperature of the
In this section, the interaction between corrosion and bulk fluid in a heat exchanger is different than the metal
heat transfer is discussed. The first part explains how a temperature or the temperature of the fluid in a boundary
metal surface can experience differences in the corrosion layer near the wall. Heat transfer to the liquid can cause
process under heat transfer conditions, compared to when the corrosion products to dissolve instead of forming a
the temperature distribution is uniform. In the second protective layer, because their solubility near the wall is
part, it is demonstrated how corrosion can affect heat high. This can increase the corrosion rate. When a liquid
transfer and the performance of a heat exchanger.
3.1 I nfluence of heat transfer on corrosion
When heat is transferred from a solid wall to a fluid (or

vice versa), a temperature gradient exists, causing the
fluid near the surface to have a different temperature
than the bulk fluid (Welty et al., 2009). The temperature
of the boundary layers near the wall depends on the tem-
peratures of the bulk fluid, the conductive thermal resist-
ance of the heat transfer surface, and the relative value
of the convection coefficients. This can be illustrated with Figure 18: Temperature profile when a corrosive fluid is being
a thermal resistance network, as illustrated in Figure 17. cooled (left) and heated (right).

is cooled, the corrosion products do not have the tendency then occurs at the point where the solution reaches a criti-
to dissolve, but form an adherent scale, protecting the cal concentration.
surface from further corrosion (Ross, 1967). In addition to
the corrosion products, the corrosive species causing the
degradation of the metal also experience this change in 3.2 Influence of corrosion on heat transfer
solubility. Dissolved oxygen, for example, has an inverted
solubility characteristic, so it is less soluble at higher tem- Not only does heat transfer have an influence on the cor-
peratures. The mass transport of oxygen from the bulk rosion process, but corrosion can also negatively affect
fluid to the hotter metal is difficult, so this might counter- the thermohydraulic performance of the heat exchanger
act the higher corrosion rates expected for a heated fluid (Edwards, 2008; Mousavian et al., 2011).
(Ross, 1967). This is illustrated by the study of Zhao et al. (2012),
Not only a temperature gradient between the fluid in where the effect of corrosion on the performance of fin-
the boundary layer and the bulk fluid can affect corrosion. and-tube heat exchangers was investigated. The heat
On the heat transfer surface itself, different areas can have exchangers were subject to accelerated corrosion by
another temperature. According to Roberge (2000), hot keeping them in a standardised salt spray chamber for 48
metal is anodic to the same metal at a lower temperature. or 96 h. The performance after being corroded was evalu-
This is a cause of thermogalvanic corrosion, which was ated based on the cooling capacity, the heat transfer coef-
mentioned by Ross (1967), Green (1967), and Maylor (1967). ficient, the thermal resistance, and the air-side pressure
Ross (1967) and Green (1967) indicated this phenomenon drop, and compared to the performance before the test.
with the term “hot-spot corrosion”. In heat exchang- All heat exchangers had copper tubes, while the fins were
ers, local overheating typically occurs under deposits or either copper or aluminium. The influence on the cooling
because of damaging behaviour of the flow, like e.g. the capacity can be seen in Figure 19. The cooling capacity
impingement of steam, an acceleration at blockages, or of the heat exchangers with the copper fins is highest for
the stagnation in enlargements. This hot-spot corrosion is all flow rates, both before and after corrosion, due to the
discussed in more detail by Francis (1987). According to higher conductivity of copper than aluminium. The reduc-
this author, it is common in copper alloy heat exchanger tion in performance after the corrosion is larger for the
tubes, and its process depends on the temperatures and samples with the aluminium fins (maximally 19%) than
the chemicals in the water. Several examples can be for the samples with the copper fins (maximally 2%).
found of hot-spot corrosion in copper tubes in desalina- In the study by Zhao et al. (2012), the decrease in
tion plants (Todd, 1967; Powell & Michels, 2000; Abouswa heat transfer is attributed to a layer of corrosion prod-
et al., 2007). ucts that formed on the fins (mainly Al2O3, aggravated by
Additional complications occur during a change of
phase of the fluid. When a liquid starts boiling, small
bubbles of vapour are produced on the surface (nucleate Sample with Cu fins #1 #2 #3
Sample with Al fins
boiling). These bubbles can collapse and cause cavita- 6000 #1 #2 #3
tion erosion of the surface, possibly removing a protective

#1 #2 #3
layer (Green, 1967; Ross, 1967). 5000
Cooling capacity (W)
During the condensing of a vapour, a layer of liquid #1 #2 #3

4000
forms on the metal. Gases in the vapour may dissolve in
the condensate, causing localised corrosion on the metal 3000
surface (Moore & Smith, 1967; Ross, 1967). This process is
sometimes called condensate corrosion (Boffardi, 2003) 2000
and is the cause of the failure of the heat exchanger dis- 1000
cussed by Julian et al. (2015), described in the part about
pitting corrosion. 0
300 400 550 720
A last type of corrosion induced by heat transfer is Air flow rate (m3/h)
“liquid-line corrosion”, described by Green (1967). When #1 before salt spray test, #2 after 48 hours salt spray test,
a diluted acid, flowing through a tube, is heated, the water #3 after 96 hours salt spray test
can evaporate. A more concentrated solution is remaining, Figure 19: Influence of corrosion on the cooling capacity of a fin-
which might be corrosive to the metal, although the origi- and-tube heat exchanger.
nal diluted acid was not a problem. Localised corrosion Reprinted from Zhao et al. (2012), with permission from Elsevier.

Table 1: Thermal conductivity of various metals and metal oxides.
Metal Thermal conductivity Oxide Thermal conductivity

(W m–1 K–1) (W m–1 K–1)
Aluminium 237 (Kakaç et al., 2012) Aluminium oxide (Al2O3) 25 (Weast, 1981)–40 (Kakaç et al., 2012)
Copper 401 (Kakaç et al., 2012) Copper(II) oxide (CuO) 20 (Kakaç et al., 2012)
Carbon steel 52 (Weast, 1981) Iron oxide 0.58 (The Engineering ToolBox, 2017)
Stainless steel 17 (Weast, 1981) Chromium(III) oxide (Cr2O3) 10–33 (Shackelford & Alexander, 2001)
Titanium 20 (Weast, 1981) Titanium dioxide (TiO2) 11.7 (AZoM)
the galvanic cell that exists between the aluminium fins successfully been used to reduce weight and improve
and the copper tubes). This additional thermal resistance corrosion resistance in heat exchangers (Knezevic et al.,
caused by corrosion products also occurs in other heat 2014). When using different metals, one should, however,
exchangers constructed from different materials (Turissini be careful to avoid galvanic corrosion (see Section 2.7),
et al., 1997; Ranjbar, 2010; Yang et al., 2012). Table 1 lists not only with bimetallic tubes but also for tube and tube
the thermal conductivity of some metals commonly used sheet combinations or with smaller elements like bolts
in the construction of heat exchangers and some of the (Kuppan, 2000). When the use of two different metals
oxides that typically result from corrosion of these metals. cannot be avoided, they should be electrically insulated.
It can be seen that the thermal conductivity of the corro- The anodic (corroding) part should be readily replaceable
sion products is often significantly lower than that of the and have a larger surface than the cathode (area effect)
base metal. This could be expected as oxides are ceramics, (Walker, 1990). To protect the metal from the aggressive
which usually are thermal insulators because of their lack environment, coatings can be applied on the surface. Coat-
of large numbers of free electrons (Callister & Rethwisch, ings generally do not provide any mechanical strength,
2011). but only isolate the material from the fluid to preserve its
strength and integrity (Roberge, 2000). They are typically
classified into three groups: metallic, organic, and inor-
4 P
revention of corrosion ganic (non-metallic) coatings.
–– Metallic coatings can be applied to a substrate in
several ways, for example hot dipping, electroplat-
4.1 G
eneral considerations and guidelines ing, spraying, cementation, and diffusion (Revie &
Uhlig, 2008). The choice of the type of coating process
Many problems with heat exchangers for corrosive appli- depends on several factors such as the required corro-
cations can be avoided by using good practice in the mate- sion resistance or the anticipated lifetime of the coat-
rial selection, during the design phase, and during the ing. Nickel, zinc (e.g. galvanising of steel), lead, tin,
operation of the device. aluminium, cadmium, and chromium are some exam-
Material selection is one of the most important ples of metals that are often applied as coating on a
aspects in avoiding corrosion of a heat exchanger in a spe- substrate made of another material (Carter, 1977). The
cific fluid. Some metals are better suited for a certain envi- coating can be both cathodic or anodic with respect to
ronment than others, and the most appropriate material the substrate.
should be chosen (Tuthill, 1990). If only expensive, highly –– Paints, resins, lacquers, varnishes, and plastic lin-
alloyed materials can resist the environment, a plate heat ings are all categorised as organic coatings (Roberge,
exchanger is preferred, because these require less mate- 2008). They are typically designed to have a high
rial than a shell-and-tube heat exchanger. If, however, a impermeability and an inhibitive function to the cor-
shell-and-tube heat exchanger would be preferred, the rosion process while they can also contain cathodi-
corrosive fluid has to flow in the tubes, as then the shell cally protective pigments.
can be made of a cheaper metal (Kakaç et al., 2012). For –– Inorganic (non-metallic) coatings can be divided into
shell-and-tube heat exchangers with different corrosive two categories (Roberge, 2008). Hydraulic cements,
fluids on the shell side and the tube side, bimetallic tubes ceramics, carbon, silicates, and glass are applied on
are sometimes used. These tubes are made from two dif- the surface of the protected substrate. Other treat-
ferent metals, which are usually co-extruded and have ments change the natural characteristics of the oxide

film, with a low corrosion resistance, into a different the holes in the tube sheets before rolling should be as
protective film or metallic oxide with a higher cor- small as possible. Otherwise, residual stresses after rolling
rosion resistance in the specific operational envi- could induce SCC (Kuppan, 2000). To completely rule out
ronment. Some examples here include anodising, leakages from one fluid to the other, special designs with
nitriding, and phosphatising. double tube sheets in shell-and-tube heat exchangers or
double-wall plates in plate heat exchangers can be used
Each of these types of coatings has been applied in heat (Kuppan, 2000). With these designs, any leakage at the
exchangers to protect the heat transfer surface from a cor- tube-tube sheet interface or through a pit in a plate will be
rosive environment (Chang & You, 1997; Fedrizzi et al., to the environment instead of to the other fluid.
2008; Jong-Soon et al., 2009). In addition to the protec- Special care should also be taken when designing the
tion from corrosion, coatings can also fulfil a different flow path. Dead spaces or areas with low velocities have
role. Wang and Chang (1998) tested hydrophilic coatings to be avoided, because they can induce pitting or crevice
to improve the condensate draining in fin-and-tube heat corrosion. Dead spaces can, for example, exist in verti-
exchangers. This better draining could reduce the pres- cal shell-and-tube heat exchangers above the upper tube
sure drop up to 40%. Also in dehumidifying heat exchang- sheet as a result of a wrong rolling procedure where the
ers, hydrophilic coatings have successfully been applied tubes protrude the tube sheet and a crown is produced
(Hong & Webb, 2000). Hydrophobic coatings, however, in the tube sheet (Collins, 1955). An incorrectly and a cor-
have also been tested in heat exchangers. Das et al. (2000) rectly rolled tube bundle are shown in Figure 20. The fluid
applied such coatings on copper tubes to promote drop- velocity should also not be too high, as this can cause
wise condensation in a condenser. The achieved heat erosion corrosion. Erosion corrosion at the first centime-
transfer coefficients were, in some cases, 14 times higher tres of the tubes is common and often called tube-inlet
than with filmwise condensation. A disadvantage of coat- corrosion. It can be prevented by inserting a short, wear-
ings is that they usually have a significantly lower thermal resistant tube in the wearing inlets, similar to the princi-
conductivity – lower than that of the base metal. In gas ple of bimetallic tubing (Kuppan, 2000). The outer surface
turbines, for example, inorganic coatings are applied of the tubes located near the inlet of the shell is also sen-
because of their lower thermal conductivity to reduce sitive to erosion corrosion. Protection can be provided
the temperature of the airfoils (Strangman, 1985). In heat under the form of wear-resistant impingement plates or
exchangers, their thickness is normally only in the order angle tube protectors, shown in Figures 21 and 22, respec-
of micrometres (Roberge, 2000), so their effect on the heat tively (Kuppan, 2000). Another type of corrosion that can
transfer is limited. be avoided by carefully designing the flow path is fretting
Apart from the material selection, some other meas- corrosion caused by flow-induced vibrations of the tubes.
ures can be taken during the design phase to reduce the Kuppan (2000) proposed changing the fluid velocity or
risk of corrosion-based failures. When the conditions stiffening the tubes with tube support plates, while Walker
to which the heat exchanger will be exposed can cause (1990) also mentioned the use of coatings to roughen the
uniform corrosion, the construction material is typically surface and dampen vibrations.
made thicker than required to guarantee the necessary Pakhomov and Parshin (1993) gave some recommen-
strength over the entire lifetime. This extra material is dations for the thermal design. They defined the term
called the corrosion allowance. This strategy is, however,
not possible with plate heat exchangers, because the plates
need to be very thin, both to maintain a compact form and
for production purposes (Kakaç et al., 2012). Also during
the design phase, sharp edges should be avoided, because
these can induce corrosion fatigue or crevice corrosion. In
general, crevices should always be avoided. When they
cannot be avoided, they should either be filled or be wide
and shallow. The connection between the tubes and the
tube sheets in shell-and-tube heat exchangers is a sensi-
tive location for crevice corrosion. One possibility to avoid
these crevices is to weld the connection instead of rolling.
When using a rolled connection, the difference between Figure 20: Dead space in a vertical heat exchanger (left) and a
the outer diameter of the tubes and the inner diameter of correctly rolled heat exchanger (right).

Figure 21: Impingement plate in a shell-and-tube heat exchanger. Figure 23: Effect of heat transfer coefficient on the effective
temperature near a heat transfer surface.
temperatures or concentrations, non-complete removal

of oxygen, and non-uniform distribution of corrosion
inhibitors (Kuppan, 2000). A good practice is to filter out
abrasive particles, stringy debris, and organisms to avoid
erosion corrosion and MIC. Cleaning of the heat transfer
surface should be done regularly, because deposits on
the metal can induce pitting, crevice corrosion, MIC, and
de-alloying. When MIC is a potential problem, organisms
attached to the surface can be killed by temporarily recir-
culating the cooling water, increasing its temperature to a
level where the organisms cannot survive. Finally, when
Figure 22: Angle tube protectors in a shell-and-tube heat performing maintenance, care should be taken to avoid
exchanger. residual stagnant pools of fluid in the device. Of course,
Republished with permission of Taylor & Francis Group LLC, from this is only possible when the design allows complete
Kuppan (2000); permission conveyed through Copyright Clearance drainage. As mentioned before, low velocities can induce
Center, Inc.
pitting corrosion. Furthermore, some of the remaining
liquid can dry up, leaving a more concentrated and corro-
sive solution. Figure 24 illustrates a bad and a good design
“effective temperature” as the temperature of the fluid in
(Elliot, 1987).
the boundary layer near the wall, which is, in most cases,
Of the many considerations described above, several
closer to the wall temperature than to the bulk fluid tem-
recur as recommendations in failure analyses. In the ref-
perature. Pakhomov and Parshin advised to maximise the
erences made in the discussion on the different types of
heat transfer coefficient on the side of the heat exchanger
corrosion, the recommendations encountered most often
where the coolest fluid is present. This will result in the
are a change of the environment (e.g. by the addition of
lowest possible effective temperature and reduced cor-
rosion kinetics. In the thermal resistance network of
Figure 17, this would mean minimising Rconv,c to reduce Tw,c.
Figure 23 shows the temperature profile for heat transfer
over a heat transfer surface for the case where a corrosive
fluid is being cooled (left) or heated (right). For both cases,
the temperature profile is illustrated with the highest heat
transfer on both the hot and cold sides. It can be seen that
in both cases, the lowest effective temperature occurs
when the highest heat transfer is on the side of the cold
fluid.
In addition to a careful design, attention should be
paid to the operation of the heat exchanger. Possible Figure 24: Design of a vessel allowing incomplete (left) and
problems potentially inducing corrosion are off-design complete (right) drainage.

corrosion inhibitors) and the choice of a different construc- hydrogen-induced damage happen at the cathode. Cor-
tion material. The advice to change the flow configuration rosion of a structure can be avoided by altering the cir-
is also often given. It might, in some cases, be necessary to cumstances such that the protected surface becomes a
either increase or decrease the flow velocity or to reduce cathode. This cathodic protection can be done by reduc-
the level of turbulence in the flow. Preventive measures ing the corrosion potential and is in practice achieved
like periodic cleaning or regular non-destructive testing by two possible strategies (Roberge, 2008). A “sacrificial
of the heat exchanger are also repeatedly recommended. anode” can be connected to the structure to be protected.
Finally, it is sometimes proposed to change the mechani- This part is made from a metal that is less noble in the
cal design or production process. An overview of which galvanic series than the protected structure and will, by
preventive strategy is given in the failure analyses can be consequence, suffer corrosion instead of the protected
found in Table 2. metal. Zinc and magnesium are often used in combination
with steel. When this sacrificial anode has been exten-
sively consumed, it is simply replaced by a new one. This
4.2 Cathodic protection method is easy and relatively cheap; however, the pro-
vided protection is uncontrollable and frequent replace-
As explained earlier, corrosion processes consist of ment of the anodes might be necessary. The second option
a cathodic and an anodic reaction. The anode is the is to use “impressed current cathodic protection” (ICCP).
reactive metal and will corrode. Only features such as The protected structure is connected to an anode through
Table 2: Recommendations given in failure analyses to prevent corrosion.
Change Different Change flow Periodic Regular Change mechanical

environment material configuration cleaning testing design
Abouswa et al. (2007) X

Abraham et al. (2009) X X
Allahkaram et al. (2011) X X
Chandra et al. (2010) X X X X
Cincera et al. (2012) X X
Corleto and Argade (2017) X
Corte et al. (2015) X X X
Deen et al. (2010) X
Eich (2012) X X X X
Ghayad et al. (2015) X
Guo et al. (2011) X
Ifezue and Tobins (2015) X X X
Julian et al. (2015) X
Kaul et al. (1995) X
Khodamorad et al. (2016) X
Klein et al. (2014) X X X
Klenowicz et al. (2003) X
Kuznicka (2009) X
Mousavian et al. (2011) X X
Ranjbar (2010) X X
Ravindranath et al. (2012) X X
Reitz (2002) X X
Sanders and Iwand (2013) X X
Shahrani and Al-Subai (2014) X X
Sharma (2014) X
Tawancy (2009) X
Traubert and Jur (2012) X X X X
Turissini et al. (1997) X
Xie et al. (2015) X X
Yang et al. (2012) X
Zheng et al. (2014) X X X

an external power source. Because of the electric current, Several patents exist disclosing the use of cathodic
impressed by this power source, the protected structure protection in heat exchangers. Patents by Deivasigamani
is cathodically polarised. Because of the anodic polarisa- and Preiser (2013), Noboru et al. (1989), Inagaki et al.
tion, anode materials with a passive surface in the elec- (2002), and Ivusic (2004) discuss the protection of shell-
trolyte are preferred to limit the metal dissolution and the and-tube heat exchangers. In the patent by Deivasigamani
anode consumption rate. and Preiser (2013), both a zinc sacrificial anode and an
Al-Hashem et al. (1997) investigated the possibility of impressed current system with an iron anode are installed
cathodically protecting heat exchanger tubes in seawater in the end chambers. In Figure 25, the two anodes are indi-
at 45°C. The tubes had an internal diameter of 1.27 cm and a cated with 82 and 90, respectively.
length of 6 m, and were made of a copper nickel alloy (UNS A different location, instead of the end chambers,
C71500). The cathodic protection was done with sacrificial where the anode (either as sacrificial anode or for ICCP)
anodes made from carbon steel and attached to both ends can be placed is in the piping just before the shell-and-
of the tubes. The potential of the tubes was recorded at six tube heat exchanger. This was done in the patent by
different places, and the current between the anodes and Noboru et al. (1989) and can be seen in Figure 26.
the tubes was measured with a zero resistance ammeter. In the patent by Inagaki et al. (2002), an invention
The results indicated that the tubes could be protected is described where the goals were to both cathodically
over the entire length. In another study by Wigen et al. protect the tubes and to prevent biofouling of the tube
(2007), the possibility to use cathodic protection to sheet. In Figure 27, the brass tubes are indicated with 1b
prevent corrosion in small-bore stainless-steel tubes was and the tube sheet with 1a. On the tube sheet, a titanium
investigated. In this research, a resistor-controlled sacri- sheet 4 is deposited. With controller 7, the potentials of
ficial anode was used to protect the tubes. Although rec- the cathode (the tubes) and the anode (the titanium sheet)
ommended practices advise a maximum length/diameter are controlled so that oxygen evolves at the anode. This
(L/D) ratio of 5 for carbon steel tubes, their calculations oxygen evolution prevents organisms in the seawater from
showed that this ratio depends on the electrical conduc- fouling the surface. An additional controller 40 is used to
tivity of the fluid, the diameter of the tube, the potential
necessary for protection, and the required current density.
It was found that for stainless-steel tubes, cathodic pro-
tection is possible for tubes with L/D ratios up to 225. The
authors discussed a case where a heat exchanger with
10-m-long tubes was used for cooling with seawater in a
plant. The original heat exchanger with tubes made from
a copper-nickel alloy failed after only a couple of months.
The new heat exchanger had super duplex tubes that were
cathodically protected. After several years, no signs of cor-
rosion were detected.
Care should, however, be taken when designing the
cathodic protection system. In the condenser described
in the failure analysis by Shalaby (2006), titanium tubes
were coupled with Muntz tube sheets and carbon steel
waterboxes. Aluminium sacrificial anodes were attached
to the waterboxes to cathodically protect these waterboxes
and the tube sheets from galvanic corrosion. Despite the
sacrificial anodes (that were found consumed), the carbon
steel, Muntz tube sheets, and titanium all suffered cor-
rosion. Their research indicated that the main corrosion
process was hydrogen embrittlement caused by the over-
protection provided by the sacrificial anodes. The authors
advised the plant operators to remove the aluminium
anodes and to apply a coating to the waterboxes and the Figure 25: Cathodic protection of heat exchanger tubes by placing
tube sheets. Similar findings were published by Fulford the anodes 82 and 90 in the end chamber (Deivasigamani & Preiser,
et al. (1987) and Tsai et al. (1987). 2013).

control the potential of the tubes so that they are cathodi-

cally protected.
The invention in the patent by Ivusic (2004) makes
use of mobile anodes to protect the heat exchanger tubes.
Several spheres with a diameter slightly smaller than the
tube diameter are injected in the flow before the heat
exchanger and are extracted from the flow after the device.
These spheres are either made of metal or of polymer but
with metallic elements on the surface. This metal needs to
be less noble than the protected tubes, so they will behave
as an anode when they come in contact with the tubes.
The principle is illustrated in Figure 28.
Cathodic protection is not limited to shell-and-tube heat
exchangers. The patents by Suzuki et al. (2015) and Terai
et al. (1980) discuss the production of aluminium tube-fin
heat exchangers where the fins have a different composition
than the tubes. The fins act as sacrificial anodes to the tubes,
preventing leakages caused by corrosion of the tubes.
Finally, the use of cathodic protection for plate heat
exchangers is discussed in the patent by Graff et al. (2017).
This patent discloses the placement of either sacrificial
anodes or anodes for ICCP in the inlet channel.
Figure 26: Cathodic protection of heat exchanger tubes by placing
the anodes 9 in the piping 10 before the heat exchanger (Noboru
et al., 1989). 4.3 Anodic protection
Another electrochemical method of reducing corrosion is

anodic protection. As with ICCP, an external power source
is required. It serves, however, to shift the potential of the
protected metal in the anodic direction. For some metals,
like e.g. stainless steels or titanium, this does not increase
the corrosion rate because the metals are passivated (Revie
& Uhlig, 2008). As can be seen in Figure 29, for metals
Figure 28: Cathodic protection of heat exchanger tubes with mobile

sacrificial anodes (SA).
Figure 27: Cathodic protection system to prevent biofouling of the These balls can be made of a polymer with metallic elements (A) or
tube sheet and corrosion of the tubes (Inagaki et al., 2002). can be completely metallic (B) (Ivusic, 2004).

anodic corrosion protection of a sulphuric acid cooler is

discussed in a paper by Paulekat et al. (1982). The acid,
with concentrations up to 99% and temperatures up to
160°C, flows through austenitic stainless-steel pipes.
Without protection, this environment would cause severe
damage to the metal. After performing laboratory studies
to determine the passive range and the current densities
for several conditions, the system was installed and the
corrosion rates were reduced up to two orders of magni-
tude. According to the authors, the system can be used on
complicated designs and, although it has a high invest-
ment cost (a power source, a controller, a cathode, and a
reference electrode are required), the system can reduce
Figure 29: Polarisation curve for a metal with an active-passive the total cost because a cheaper material can be chosen
transition. and less shutdowns occur.
Reprinted from Roberge (2008), with permission from McGraw-Hill
A lot of information on the use of anodic protec-
Education.
tion of heat exchangers is disclosed in patents. Patents
discussing the anodic protection of shell-and-tube heat
that exhibit an active-passive behaviour, an increase in exchangers were assigned to McAlister (1987), Cameron
potential first increases the corrosion rate (indicated on (1996), and Sanz (1986, 1996). In all these documents, the
the figure by the current density), but then decreases corrosive fluid, e.g. sulphuric acid, flows over the tubes
with several orders of magnitude. In the transpassive at the shell side. To protect the heat exchanger, both the
region, the corrosion rate sharply increases again. Unlike shell and the tubes are connected to a cathode through
cathodic protection, anodic protection can only be used an external power source. These cathodes are placed in
with metals that have this active-passive transition in the the shell as long cylindrical objects, parallel to the tubes.
specific environment of the heat exchanger. This kind of set-up is shown in Figure 30. In this figure, the
An anodic protection system requires a power source, shell 10, the cathode 15, the power source 25, and the two
a potential controller, a cathode, and a reference electrode reference electrodes 20 and 30 are displayed. The tubes
(Roberge, 2000). The cathode should be as large as pos- are not shown.
sible. For a reduction of the electrical resistance of the The patents by Bennett and Elliot (1981, 1982) also use
circuit, multiple cathodes can be used in parallel. The ref- anodic protection in heat exchangers. In their invention,
erence electrode is required to carefully control the poten- however, the goal is to prevent biofouling and scaling of
tial. This electrode should have a potential that is constant the tubes. The tubes are anodically polarised to a potential
and should not be influenced by temperature. Failure of where they start evolving oxygen. This prevents organisms
the reference electrode can cause a faulty operation of the in seawater from attaching to the surface. This technique
system with large corrosion problems as a consequence. is similar to the patent by Inagaki et al. (2002), but there,
The advantages of anodic protection are that, once
the metal has passivated, only a small current is required.
Moreover, the current gives a direct indication of the cor-
rosion rate. However, it only reduces the corrosion rate
but never entirely stops it, in contrast to cathodic pro-
tection. When the electrical supply stops, the metal can
de-passivate and be subject to serious damage (Revie &
Uhlig, 2008). If this happens when the heat exchanger is
expected to be protected, the consequences can be disas-
trous. Anodic protection can only be used in well-known
and stable environments, as changes in the environment
can chance the active-passive behaviour (Roberge, 2000).
One of the main applications of anodic protection is
to prevent corrosion in heat exchangers and other equip- Figure 30: Anodic protection of a shell-and-tube heat exchanger
ment handling sulphuric acid (Roberge, 2000). The (McAlister, 1987).

the tubes were cathodically polarised and the tube sheet

was used as anode.
Plate heat exchangers can be anodically protected
as well, as discussed in the patents assigned to Ana-
stasijevic et al. (2004), Karlsson and Berthagen (1989),
and Carlson et al. (1986). For both patents by Anastasi-
jevic et al. (2004) and Karlsson and Berthagen (1989),
the cathodes are placed in the inlet and outlet channels.
The reference electrode is placed in the outlet channel
in the patent by Anastasijevic et al. (2004) and in a sepa-
rate flow path, further away from the cathodes, in the
patent by Karlsson and Berthagen (1989). Karlsson and
Berthagen (1989) claimed that this secures the anodic
protection in all parts of the heat exchanger. The inven-
tion in the patent by Carlson et al. (1986) is based on
Figure 32: Anodic polarisation of plates to control biofouling in
the patents assigned to Bennett and Elliot (1981, 1982), plate heat exchangers (Carlson et al., 1986).
and serves again to prevent biofouling. Pairs of plates
act either as cathode or anode. In the channel with the
fouling fluid, e.g. seawater, oxygen is evolved from the
anode, thus preventing the deposition of organisms 5 Conclusion
on the plates. The corrosion protection and biofoul-
ing control of plate heat exchangers are illustrated in In this paper, all main aspects about corrosion in heat
Figures 31 and 32, respectively. The power source 20 in exchangers were reviewed. The first part provided a dis-
Figure 31 is connected to the cathodes 16 and 17 in the cussion on the different types of corrosion occurring in
inlet and outlet channel and to the plates 7. The refer- heat exchangers, together with an overview of failure
ence electrode 27 is also located in the outlet channel. analyses where the respective corrosion processes were
In Figure 32, oxygen is evolved in the seawater channels the main cause of failure. Next, the influence of heat
34 from the plates 12 connected to the positive pole of transfer on corrosion and the influence of corrosion on the
the power source. thermohydraulic performance of heat exchangers were
discussed. Finally, several methods used for the preven-
tion of corrosion were considered. The following observa-
tions and conclusions could be made:
1. Ample examples of heat exchangers failed by cor-
rosion can be found, indicating the severity of the
problem. Only a minority of the failures are caused by
uniform corrosion. Localised forms, especially pitting
and SCC, frequently occur. These are followed by crev-
ice corrosion and MIC.
2. The corrosion rate is influenced by the heat transfer
in different ways. The main factor is the temperature
gradient that exists over the boundary layer near the
wall. This causes differences in the solubility of ele-
ments in the fluid. Depending on the corrosive species
and on whether the fluid is being heated or cooled,
this gradient can increase or decrease the corrosion
rate.
3. Corrosion has an adverse effect on the thermohydrau-
lic performance of heat exchangers. The heat transfer
rate will typically decrease. This is mainly attributed
Figure 31: Anodic protection of a plate heat exchanger to the low thermal conductivity of the corrosion
(Anastasijevic et al., 2004). products, which can be several orders of magnitude

smaller than that of the base metal. Furthermore, the corrosion rate (like cathodic or anodic protection) is still
corrosion products, which are usually more volumi- limited and should be expanded. An alternative approach
nous than the base metal, can obstruct the flow and might be to change the heat exchanger design so that
increase the pressure drop. corrosion can be tolerated or that corroded parts can be
4. The most important methods to prevent corrosion of replaced easily.
heat exchangers are an adequate material selection
with a combination of proper control of the environ-
ment. During the design, attention should be paid to
avoiding crevices and sharp edges, and for demand-
References
ing applications, special construction elements can
Abouswa K, Elshawesh F, Elragei O, Elhood A. Corrosion
be used (e.g. impingement plates in shell-and-tube investigation of Cu-Ni tube desalination plant. Desalination
heat exchangers or double-wall plate heat exchang- 2007; 205: 140–146.
ers). Regarding the thermal design, it is important to Abraham JG, Kain V, Dey GK. MIC failure of cupronickel condenser
have the highest heat transfer coefficient on the cold tube in fresh water application. Eng Fail Anal 2009; 16:
934–943.
side as to reduce the temperature of the heat transfer
Al-Hashem A, Carew J, Al-Sayegh A. The sacrificial cathodic
surfaces. protection of UNS C71500 heat exchanger tubes in Arabian Gulf
5. Although cathodic and anodic protection have been sea water. Technical report. Houston, TX: NACE International,
proven to be effective in laboratory studies and sev- 1997.
eral examples of the application of such systems in Allahkaram S, Zakersafaee P, Haghgoo S. Failure analysis of heat
heat exchangers can be found in the patent litera- exchanger tubes of four gas coolers. Eng Fail Anal 2011; 18:
1108–1114.
ture, only a small amount of scientific literature on
Anastasijevic N, Daum KH, Schalk W, Laibach S. Plate-type heat
this topic exists. The results of the existing scientific exchanger with anodic corrosion protection, 2004. Patent
research on these systems in heating or cooling appli- application: US 2004/0251005 A1.
cations are mostly positive. Some failure analyses AZoM. Titanium dioxide-titania (TiO2). URL https://www.azom.com/
exist, however, where improper use of cathodic pro- article.aspx?ArticleID=1179. Accessed 16 March, 2018.
Bennett JE, Elliot JE. Anodically polarized surface for biofouling and
tection is at the basis of the failure.
scale control, 1981. US Patent 4,256,556.
Bennett JE, Elliot JE. Anodically polarized surface for biofouling and
This paper illustrates that the presence of a corrosive scale control, 1982. US Patent 4,345,981.
environment, flow behaviour, high temperatures, and Boffardi BP. Corrosion inhibitors in the water treatment industry.
heat transfer over a surface make the corrosion of heat In: ASM handbook, vol. 13A, chapter Corrosion: fundamentals,
exchangers a specific problem. Nonetheless, failures testing, and protection. Materials Park, OH: ASM International,
2003: 891–906.
occur in many different domains. Applications with sea-
Callister WD, Rethwisch DG. Materials science and engineering, vol.
water, geothermal brines, or in (petro)chemical plants are 5. New York: John Wiley & Sons, 2011.
only some examples of industries that are regularly faced Cameron GM. Anodis protection method and system, 1996. US
with these corrosion-related issues. Patent 5,513,694.
It is clear from the high number of failure analyses Carlson RC, DeRespiris DL, Sacco AR, Groszek DJ, Ivancic BM. Plate
and frame heat exchanger assembly with anodic protection,
found in the recent literature that there is a need for a
1986. US Patent 4,586,562.
continued research on how to reduce the risk of corrosion Carter V. Metallic coatings for corrosion control. London: Newnes-
in heat transfer applications. Compared to the amount of Butterworths, 1977.
failures found, the number of publications on the influ- Cassagne T, Houlle P, Zuili D, Bluzat P, Corrieu JM, Larché N.
ence of corrosion on the performance of heat exchangers Replacing titanium in sea water plate heat exchangers. In:
and the effect of heat transfer on the corrosion process are CORROSION 2010. Houston, TX: NACE International, 2010.
Chandra K, Kain V, Dey G, Shetty P, Kishan R. Failure analysis of
scarce. Also regarding the prevention of corrosion, there
cupronickel evaporator tubes of a chilling plant. Eng Fail Anal
still is room for improvement. 2010; 17: 587–593.
To solve the problem and increase the knowledge on Chang J, You S. Enhanced boiling heat transfer from microporous
the topic, new research could be done on several aspects. surfaces: effects of a coating composition and method. Int J
First of all, continued investigations on the development Heat Mass Transfer 1997; 40: 4449–4460.
Cincera S, Barella S, Bellogini M, Bresciani E. De-nickelification of
of corrosion-resistant materials for different environments
70/30 cupronickel tubing in a cooling heat exchanger. J Fail
are required. The separation of the metal from the solution Anal Prev 2012; 12: 300–304.
by the application of a suitable coating is another option. Collier J. Reliability problems of heat transfer equipment. Heat
The research on electrochemical techniques to reduce the Transf Eng 1983; 4: 51–62.

Collins JA. Effect of design, fabrication and installation on the Guedes Soares C, Garbatov Y, Zayed A. Effect of environmental
performance of stainless steel equipment. Corrosion 1955; 11: factors on steel plate corrosion under marine immersion
27–34. conditions. Corros Eng Sci Technol 2011; 46: 524–541.
Corleto CR, Argade GR. Failure analysis of dissimilar weld in heat Guo C, Han C, Tang Y, Zuo Y, Lin S. Failure analysis of welded 0Cr13Al
exchanger. Case Studies Eng Fail Anal 2017; 9: 27–34. tube bundle in a heat exchanger. Eng Fail Anal 2011; 18:
Corte J, Rebello J, Areiza M, Tavares S, Araujo M. Failure analysis 890–894.
of AISI 321 tubes of heat exchanger. Eng Fail Anal 2015; 56: Health and Safety Executive. Catastrophic failure of shell and
170–176. tube production cooler, 2006. URL: http://www.hse.gov.uk/
Crum J, Hazeldine P, Shoemaker L, Peschel J. Evaluation of materials offshore/alerts/sa_01_06.htm. Accessed 23 April, 2018.
for seawater plate heat exchanger applications. In: CORROSION Hong K, Webb RL. Wetting coatings for dehumidifying heat
2007. Houston, TX: NACE International, 2007. exchangers. HVAC&R Res 2000; 6: 229–242.
Das A, Kilty H, Marto P, Andeen G, Kumar A. The use of an organic Huttunen-Saarivirta E, Honkanen M, Lepistö T, Kuokkala VT, Koivisto
self-assembled monolayer coating to promote dropwise L, Berg CG. Microbiologically influenced corrosion (MIC)
condensation of steam on horizontal tubes. J Heat Transf 2000; in stainless steel heat exchanger. Appl Surf Sci 2012; 258:
122: 278–286. 6512–6526.
Davis JR. Corrosion: Understanding the basics. Materials Park, OH: Ifezue D, Tobins F. Corrosion failure of aluminum heat exchanger
ASM International, 2000. tubes. J Fail Anal Prev 2015; 15: 541–547.
Deen K, Virk M, Haque C, Ahmad R, Khan I. Failure investigation of Inagaki S, Sakurada S, Nakashima S, Ooba T. Antifouling system for
heat exchanger plates due to pitting corrosion. Eng Fail Anal structure exposed to seawater and heat exchanger, 2002. US
2010; 17: 886–893. Patent 6,579,429.
Deivasigamani S, Preiser HS. Heat exchanger including selectively Ivusic V. Cathodic protection of heat exchanger tubes by means of
activated cathodic protection useful in sulfide contaminated mobile sacrificial anodes, 2004. Patent WO 2004/104270 A1.
environments, 2013. US Patent 8,511,370. Jennings P. A new look at the explosion that caused the Sultana
Depover T, Verbeken K. The detrimental effect of hydrogen at disaster. The Hartford Steam Boiler Inspection and Insurance
dislocations on the hydrogen embrittlement susceptibility of Company, 2015.
Fe-C-X alloys: an experimental proof of the HELP mechanism. Jong-Soon K, Tae-Ho K, In-Kwan K. Surface treatment to improve
Int J Hydrogen Energy 2018; 43: 3050–3061. corrosion resistance of Al plate heat exchangers. Trans
Dexter S. Microbiologically influenced corrosion. In: ASM handbook, Nonferrous Metals Soc China 2009; 19: s28–s31.
vol. 13A, chapter Corrosion: Fundamentals, Testing, and Julian R, Lichti K, Mroczek E, Mountain B. Heavy metal scaling and
Protection. Materials Park, OH: ASM International, 2003: corrosion in a geothermal heat exchanger. In: Proceedings
398–416. World Geothermal Congress 2015, 2015.
Edwards JE. Design and rating shell and tube heat exchangers. Kakaç S, Liu H, Pramuanjaroenkij A. Heat exchangers: selection,
Teesside, UK: P & I Design Ltd., 2008. rating, and thermal design. Boca Raton, FL: CRC Press, 2012.
Edwards R. Aqueous corrosion of steel, 2013. Karlsson B, Berthagen L. Corrosion protection for heat exchangers,
Eich G. Erosion corrosion on oil coolers for navy ships. J Fail Anal 1989. US Patent 4,800,007.
Prev 2012; 12: 490–493. Kaul R, Muralidharan N, Jayakumar T, Kasiviswanathan K, Raj B,
El-Amoush AS, Zamil A, Jaber D, Ismail N. Stress corrosion cracking Dayal R, Raju V, Gnanamoorthy J, Pattu S. Failure analysis of
of the pre-immersed tin brass heat exchanger tube in an carbonate reboiler heat exchangers. Eng Fail Anal 1995; 2:
ammoniacal solution. Mater Des (1980–2015) 2014; 56: 842–847. 165–174.
Electric Power Research Institute. Cost of corrosion in the electric Khodamorad S, Alinezhad N, Fatmehsari DH, Ghahtan K. Stress
power industry. EPRI 1004662, 2001. corrosion cracking in type 316 plates of a heat exchanger. Case
Elliot P. Design details to minimize corrosion. ASM Handb 1987; 13: Studies Eng Fail Anal 2016; 5: 59–66.
338–343. Klein U, Zunkel A, Eberle A. Breakdown of heat exchangers due
Fedrizzi L, Andreatta F, Paussa L, Deflorian F, Maschio S. Heat to erosion corrosion and fretting caused by inappropriate
exchangers corrosion protection by using organic coatings. operating conditions. Eng Fail Anal 2014; 43: 271–280.
Progr Org Coat 2008; 63: 299–306. Klenowicz Z, Darowicki K, Krakowiak S, Krakowiak A. Corrosion-
Francis R. Hot spot corrosion in condenser tubes: its causes and erosion damage of heat exchanger tubes by desalted crude oil
prevention. Br Corros J 1987; 22: 199–201. flowing at shell side. Mater Corros 2003; 54: 181–187.
Fulford J, Schutz R, Lisenbey R. Characterization of titanium Knezevic M, Jahedi M, Korkolis YP, Beyerlein IJ. Material-based
condenser tube hydriding at two Florida power and light design of the extrusion of bimetallic tubes. Comput Mater Sci
company plants. In: Joint ASME/IEEE Power Generation 2014; 95: 63–73.
Conference, Miami Beach, Florida, 1987. Koch GH, Brongers MP, Thompson NG, Virmani YP, Payer JH.
Ghayad IM, Hamid ZA, Gomaa N. A case study, corrosion failure of Corrosion cost and preventive strategies in the United States.
tube heat exchanger. J Metallurg Eng 2015; 4: 57–61. Technical report, 2002.
Graff M, Mazej M, Ogorevc T. Cathodic protection of a heat Kuppan T. Heat exchanger design handbook. New York: Marcel
exchanger, 2017. Patent WO 2017/072177 A1. Dekker, Inc., 2000.
Green M. Corrosion under heat transfer: service failures and Kuznicka B. Erosion-corrosion of heat exchanger tubes. Eng Fail Anal
laboratory testing. Br Corros J 1967; 2: 153–158. 2009; 16: 2382–2387.
Groysman A. Corrosion problems and solutions in oil refining and LaQue F. The corrosion handbook. New York: John Wiley & Sons,
petrochemical industry. Dordrecht: Springer, 2017. Inc., 1948.

Li YD, Xu N, Wu XF, Guo WM, Shi JB, Zang QS. Failure analysis of the Sanders S, Iwand H. Failure analysis of ethanol vaporizer heat
condenser brass tube in 150 MW thermal power units. Eng Fail exchanger tubes. J Fail Anal Prev 2013; 13: 328–332.
Anal 2013; 33: 75–82. Sanz D. Anodic protection system and method, 1986. US Patent
Marsh P. Process-safety case studies – catastrophic failures of 4,588,022.
heat exchangers, 2016. URL http://www.engineersjournal. Sanz D. Anodically protected heat exchanger, 1996. US Patent
ie/2016/11/01/process-safety-failures-heat-exchangers/. 5,515,913.
Accessed 12 April, 2018. Shackelford JF, Alexander W. Thermal properties of materials.
Maylor J. Heat transfer and corrosion of high-nickel alloys. Br Corros Materials science and engineering handbook, vol. 5. Boca
J 1967; 2: 168–171. Raton, FL: CRC Press, 2001.
McAlister DR. Control of anodic passivation systems, 1987. US Shahrani S, Al-Subai S. Failure analysis of heat exchanger tubes. J
Patent 4,689,127. Fail Anal Prev 2014; 14: 790–800.
Moore P, Smith S. Practical applications of corrosion-testing Shalaby H. Failure investigation of Muntz tubesheet and Ti tubes of
apparatus under heat-transfer conditions. Br Corros J 1967; 2: surface condenser. Eng Fail Anal 2006; 13: 780–788.
143–149. Sharma P. Microbiological-influenced corrosion failure of a heat
Mousavian RT, Hajjari E, Ghasemi D, Manesh MK, Ranjbar K. Failure exchanger tube of a fertilizer plant. J Fail Anal Prev 2014; 14:
analysis of a shell and tube oil cooler. Eng Fail Anal 2011; 18: 314–317.
202–211. Silverman D. Aqueous corrosion. In: ASM handbook, vol. 13A,
NACE International. International measures of prevention, chapter Corrosion: fundamentals, testing, and protection.
application, and economics of corrosion technologies study. Materials Park, OH: ASM International, 2003: 190–195.
Houston, TX: NACE International, 2016. Song WM, Yang GR, Hao Y, Zhang YF. The corrosion failure analysis
Noboru K, Yuuj N, Katsunari O, Katsuaki T, Shuichi I, Sunao W. of plate pipe for plate heat exchanger. In: Advanced materials
Cathodic protection apparatus in systems for the circulation of research, vol. 750. Switzerland: Trans Tech Publications, 2013:
corrosive liquids, 1989. European Patent EP 0 324 440 A1. 2252–2257.
Otegui JL, Fazzini PG. Failure analysis of tube-tubesheet welds in Strangman TE. Thermal barrier coatings for turbine airfoils. Thin
cracked gas heat exchangers. Eng Fail Anal 2004; 11: 903–913. Solid Films 1985; 127: 93–106.
Pakhomov V, Parshin A. Corrosion protection of heat-exchanger Stumm W, Morgan JJ. Aquatic chemistry: chemical equilibria and rates
equipment at the design stage. Chem Petrol Eng 1993; 29: in natural waters, vol. 126. New York: John Wiley & Sons, 2012.
181–187. Suzuki H, Sasaki T, Nagasawa M, Takahashi N. Aluminum alloy fin
Parrott R. Potential hazards from undetected corrosion in complex material for heat exchanger and method of production of same
equipment: a case study of the destructive separation of an and method of production of heat exchanger by brazing fin
offshore heat exchanger. Eng Fail Anal 2014; 44: 424–440. material, 2015. US Patent US2015/0107731 A1.
Paulekat F, Graefen H, Kuron D. Anodic corrosion protection of Syrett BC, Gorman JA. Cost of corrosion in the electric power
sulphuric acid plants with regard to the recovery of heat. Mater industry: an update. Mater Perform 2003; 42: 32–38.
Corros 1982; 33: 254–262. Tawancy H. Failure of hydrocracker heat exchanger tubes in an oil
Powell C, Michels H. Copper-nickel for seawater corrosion resistance refinery by polythionic acid-stress corrosion cracking. Eng Fail
and antifouling: a state of the art review. Rem 2000; 90: 10. Anal 2009; 16: 2091–2097.
Qiankun Z, Yafei S, Sixian R, Huifeng L, Xingjiang Z. Corrosion failure Terai S, Tanabe Z, Fukui T. Heat exchanger core having fin members
analysis on heat exchanger pipes. J Fail Anal Prev 2017; 17: serving as sacrificial anodes, 1980. US Patent 4,203,490.
349–353. The Engineering ToolBox. Thermal conductivity of common materials
Qu S, Yao G, Tian J, Zhang Z. Failure analysis of the brass tubes in a and gases, 2017. URL http://www.engineeringtoolbox.com/
lubricating oil cooler. Eng Fail Anal 2011; 18: 2232–2239. thermal-conductivity-d_429.html. Accessed 16 March, 2018.
Ranjbar K. Effect of flow induced corrosion and erosion on failure of Todd B. The corrosion of materials in desalination plants.
a tubular heat exchanger. Mater Des 2010; 31: 613–619. Desalination 1967; 3: 106–117.
Rao T, Nair K. Microbiologically influenced stress corrosion cracking Traubert TD, Jur TA. Metallurgical analysis to evaluate cracking in
failure of admiralty brass condenser tubes in a nuclear power a 316L grade stainless steel spiral heat exchanger. J Fail Anal
plant cooled by freshwater. Corros Sci 1998; 40: 1821–1836. Prev 2012; 12: 198–203.
Ravindranath K, Tanoli N, Gopal H. Failure investigation of brass Tsai WT, Ju CP, Wen YN, Lee JT. Hydride formation during the cathodic
heat exchanger tube. Eng Fail Anal 2012; 26: 332–336. polarization of titanium in artificial sea water. Surf Coat Technol
Reitz W. SO2 heat exchanger failure. Pract Fail Anal 2002; 2: 45–49. 1987; 31: 401–408.
Revie RW, Uhlig HH. Corrosion and corrosion control: an introduction Turissini R, Bruno T, Dahlberg E, Setterlund R. Corrosion failures in
to corrosion science and engineering. New York: John Wiley & plate heat exchangers. Technical report. Houston, TX: NACE
Sons, Inc., 2008. International, 1997.
Rizk TY, Al-Nabulsi KM, Cho MH. Microbially induced rupture of a Tuthill AH. The right metal for heat exchanger tubes. Chem Eng
heat exchanger shell. Eng Fail Anal 2017; 76: 1–9. 1990; 97: 120.
Roberge PR. Handbook of corrosion engineering. New York: U.S. Chemical Safety and Hazard Investigation Board. Catastrophic
McGraw-Hill, 2000. rupture of heat exchanger. U.S. Chemical Safety and Hazard
Roberge PR. Corrosion engineering: principles and practice. New Investigation Board, 2010.
York: McGraw-Hill, 2008. Veldkamp J, Goldberg T, Bressers P, Wilschut F. TNO report:
Ross T. Corrosion and heat transfer – a review. Br Corros J 1967; 2: corrosion in Dutch geothermal systems. Energy/Geological
131–142. Survey of the Netherlands, 2016.

Walker G. Industrial heat exchangers: a basic guide. Washington: Steven Lecompte

Hemisphere Pub. Corp., 1990. Department of Flow, Heat and Combustion
Wang CC, Chang CT. Heat and mass transfer for plate fin-and-tube Mechanics, Ghent University, Ghent,
heat exchangers, with and without hydrophilic coating. Int J Belgium
Heat Mass Transf 1998; 41: 3109–3120.
Wang L, Sundén B, Manglik RM. Plate heat exchangers: design,
applications and performance, vol. 11. Southampton, UK: Wit
Press, 2007.
Wassilkowska A, Skowronek T, Rybicki S. Microstructure Steven Lecompte is a research fellow at the Department of Flow,
investigation of premature corroded heat exchanger plates. Heat and Combustion Mechanics, Ghent University, Belgium. He
Mater Test 2016; 58: 218–223. obtained his master’s and PhD degrees in electromechanical engi-
Weast RC. CRC handbook of chemistry and physics, 61st ed, neering at Ghent University in 2012 and 2016, respectively. His PhD
1980–1981. Cleveland, OH: Chemical Rubber, 1981. thesis was titled “Performance evaluation of organic Rankine cycle
Welty JR, Wicks CE, Rorrer G, Wilson RE. Fundamentals of architectures: Application to waste heat valorisation”. For his mas-
momentum, heat, and mass transfer. New York: John Wiley & ter’s thesis, he received the ArcelorMittal prize. His research is on
Sons, 2009. thermomechanical energy conversion including thermo-economic
Wigen SM, Osvoll H, Gartland PO, Huang W. Efficient cathodic optimisation of thermal systems and multi-phase processes.
protection of stainless steel small bore tubing. In: CORROSION
2007. Houston, TX: NACE International, 2007. Zaaquib Yunus Ahmed
Xie J, Yazdanfar K, Anning G, Rockwell T, Ahmed A. Corrosion of a Department of Flow, Heat and Combustion
vertical shell and tube heat exchanger. In: CORROSION 2015. Mechanics, Ghent University, Ghent,
Houston, TX: NACE International, 2015. Belgium
Xu S, Wang C, Wang W. Failure analysis of stress corrosion cracking
in heat exchanger tubes during start-up operation. Eng Fail
Anal 2015; 51: 1–8.
Yager WM. The sultana disaster. Tenn Hist Q 1976; 35: 306–325.
Yang ZG, Gong Y, Yuan JZ. Failure analysis of leakage on titanium
tubes within heat exchangers in a nuclear power plant. Part I:
Electrochemical corrosion. Mater Corros 2012; 63: 7–17. Zaaquib Yunus Ahmed obtained his master’s degree in computa-
Zhao Y, Qi Z, Wang Q, Chen J, Shen J. Effect of corrosion tional mechanics from the University of Stuttgart and UPC Barce-
on performance of fin-and-tube heat exchangers with lona. The title of his master’s thesis was “Numerical simulation of
different fin materials. Exp Thermal Fluid Sci 2012; 37: tip leakage flows in a compressor cascade”. He later started his PhD
98–103. at the University of Ghent on “3-D thermal modelling of a slab in a
Zheng Y, Liu G, Zhang Y, Hu H, Song Q. Corrosion failure analysis of continuous walking beam furnace”, in cooperation with ArcelorMit-
a condenser on the top of benzene tower in styrene unit. J Fail tal, Belgium. The focus of his research is to model the heat transfer
Anal Prev 2014; 14: 286–295. into the slabs in a furnace, taking into account different thermody-
namic processes that affect the slab heat transfer.
Johan Van Bael

Flemish Institute of Technological Research
Bionotes (VITO), Mol, Belgium; and EnergyVille,
Thor Park, Genk, Belgium
Willem Faes
Department of Flow, Heat and Combustion
Mechanics, Ghent University, Ghent,
Belgium; and Flemish Institute of
Technological Research (VITO), Mol,
Belgium,
willem.faes@ugent.be Johan Van Bael has a master’s degree in electromechanical engi-
neering. As a project manager since 1995, he worked on several
regional, national, and European projects both for the government
and for the industry on cooling and heating systems, conversion
Willem Faes obtained his MSc in mechanical engineering at Ghent systems [organic Rankine cycle (ORC) and heat pumps], co-genera-
University in 2016. The title of his master’s thesis was “Useful tion, thermal energy storage, and renewable energy sources. He is
energy transfer in air-to-air heat recovery units in low energy currently the project manager for the research group on conversion
buildings”. Afterwards, he started his PhD in the same university at systems (heat pumps and ORC), thermal energy storage. Johan is
the Department of Flow, Heat and Combustion Mechanics, together expert in different European projects like FP7 Merits, FP7 E-Hub,
with the Flemish Institute of Technological Research (VITO). His H2020 Story, H2020 MATChING, and H2020 FHP, and different KIC
research is focussed on the design of heat exchangers in (corrosive) EIT projects like KIC Energy Storage, KIC Estore, and KIC Flex2. He is
geothermal environments. also an inventor of several VITO patents.

Robbe Salenbien Michel De Paepe

Flemish Institute of Technological Research Department of Flow, Heat and Combustion
(VITO), Mol, Belgium; and EnergyVille, Mechanics, Ghent University, Ghent,
Thor Park, Genk, Belgium Belgium; and Flanders Make, Core Lab
UGent-EEDT, www.flandersmake.be, Lommel,
Belgium
Robbe Salenbien obtained his MSc (cum laude) and PhD in physics Michel De Paepe (°1972) is professor of thermodynamics and heat
at the KU Leuven, in 2007 and 2012, respectively. He was involved in transfer and is currently the head of the research group Applied
photoacoustic and photothermal studies of ultrathin structures and Thermodynamics and Heat Transfer at the Faculty of Engineering
viscoelastic materials. In 2012, he joined the Flemish Institute of and Architecture of Ghent University. Research in this group focuses
Technological Research (VITO) as a researcher in energy technology, on thermodynamics of new energy systems, performance of HVAC
focussing on smart district applications such as optimal technology (heating, ventilation, and air conditioning) systems and energy
integration. He is engaged in different national and international rel- in buildings, and complex heat transfer phenomena in industrial
evant research projects, e.g. FP7 Resilient, IWT-VIS Smart Geotherm, applications, as in compact heat exchangers, combustion engines,
and EFRO SALK GeoWatt. Currently, he is active as domain responsi- refrigerant two-phase flow, and electronics cooling. Michel De
ble for the thermal energy systems research group in EnergyVille. Paepe was supervisor/promotor of 18 PhDs defended at Ghent
University. Michel De Paepe is (co)author of 117 papers published in
Kim Verbeken international peer-reviewed journals and >350 conference papers.
Department of Materials, Textiles and
Chemical Engineering, Ghent University,
Ghent, Belgium
Kim Verbeken, after obtaining his PhD from Ghent University

in 2004, obtained a postdoctoral fellowship from the Research
Foundation-Flanders (FWO). As a postdoc, he also worked for 1
year at the Max-Planck Institute for Iron Research (Düsseldorf). In
2010, Kim Verbeken became research professor at Ghent University
and initiated the research group Sustainable Materials Science
(https://www.ugent.be/ea/match/sms/en), which is embedded
in the Department of Materials, Textiles and Chemical Engineering
(MaTCh). His research focuses on three topics:
– The study of the interaction of materials with hydrogen;
– The study of degradation of materials by interaction with their
environment;
– The study of high-temperature metallurgical processes by both a
modelling and experimental approach.


Corrosion and Corrosion Prevention in Heat Exchangers

Uploaded by

Copyright:

Available Formats

Corrosion and Corrosion Prevention in Heat Exchangers

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Corrosion and Corrosion Prevention in Heat Exchangers

Uploaded by

Copyright:

Available Formats

Corros Rev 2019; aop

Corrosion and corrosion prevention in heat

Authenticated | willem.faes@ugent.be author's copy

In addition to these economic issues, Davis (2000) also

That a failing heat exchanger can pose a potential safety

Authenticated | willem.faes@ugent.be author's copy

Both galvanic corrosion and hydrogen damage in heat

Authenticated | willem.faes@ugent.be author's copy

as the application of a coating) are often susceptible to

2.2 Pitting corrosion

The first form of localised corrosion is pitting corrosion.

Authenticated | willem.faes@ugent.be author's copy

Figure 6: Pitting corrosion damage in a plate heat exchanger.

Authenticated | willem.faes@ugent.be author's copy

external, but also residual stresses, e.g. from welding

2.4 Stress corrosion

When a corrosive environment is combined with mechan-

Authenticated | willem.faes@ugent.be author's copy

Figure 9: Stress corrosion cracks induced in a bend by polythionic

Authenticated | willem.faes@ugent.be author's copy

Figure 13: Intergranular corrosion of an aluminium tube.

Authenticated | willem.faes@ugent.be author's copy

Authenticated | willem.faes@ugent.be author's copy

steel with a stainless-steel cladding and used to heat

2.10 Microbiologically induced corrosion

Microbiologically induced corrosion (MIC) is described in

Authenticated | willem.faes@ugent.be author's copy

Figure 17: Thermal resistance network for heat transfer from a hot

3.1 I nfluence of heat transfer on corrosion

When heat is transferred from a solid wall to a fluid (or

Authenticated | willem.faes@ugent.be author's copy

tion erosion of the surface, possibly removing a protective

During the condensing of a vapour, a layer of liquid #1 #2 #3

Authenticated | willem.faes@ugent.be author's copy

Table 1: Thermal conductivity of various metals and metal oxides.

Metal Thermal conductivity Oxide Thermal conductivity

Authenticated | willem.faes@ugent.be author's copy

Authenticated | willem.faes@ugent.be author's copy

temperatures or concentrations, non-complete removal

Authenticated | willem.faes@ugent.be author's copy

Table 2: Recommendations given in failure analyses to prevent corrosion.

Change Different Change flow Periodic Regular Change mechanical

Abouswa et al. (2007) X

Authenticated | willem.faes@ugent.be author's copy

Authenticated | willem.faes@ugent.be author's copy

control the potential of the tubes so that they are cathodi-

Another electrochemical method of reducing corrosion is

Figure 28: Cathodic protection of heat exchanger tubes with mobile

Authenticated | willem.faes@ugent.be author's copy

anodic corrosion protection of a sulphuric acid cooler is

Authenticated | willem.faes@ugent.be author's copy

the tubes were cathodically polarised and the tube sheet

Authenticated | willem.faes@ugent.be author's copy

Authenticated | willem.faes@ugent.be author's copy

Authenticated | willem.faes@ugent.be author's copy

Authenticated | willem.faes@ugent.be author's copy

Walker G. Industrial heat exchangers: a basic guide. Washington: Steven Lecompte

Johan Van Bael

Authenticated | willem.faes@ugent.be author's copy

2.2 Pitting corrosion

2.4 Stress corrosion

2.10 Microbiologically induced corrosion

3.1 I nfluence of heat transfer on corrosion