Previously of BetzDearborn, John R Lane, EurChem
C.Chem M.W.M.Soc MRSC B.Sc ARCS is now an independent consultant
specialising in water treatment and Risk Assessment. John has
kindly written the following article to provide the engineer
with a factual unbiased source of reference that clearly identifies
the cause and effect of problems caused by using water in heating
Water properties and sources
Water is cheap and plentiful, and although used by all as
the heat transfer medium in wet central heating systems, beyond
the basic engineering requirements, many equipment manufacturer
and installers give little consideration to its fundamental
characteristics. In many cases, this leads directly to the
creation of waterside problems and the inability to resolve
The water molecule
Amongst all molecules, naturally occurring or man-made, water
is unique. It is present in the environment in all three states
of matter, solid as ice, liquid as water, and gas as vapour
Because of its unique structure, the amount of energy required
to vaporise water and the amount of energy released when water
freezes, is greater than for other simple chemical compounds.
Linked to this, the amount of heat absorbed or released with
change of temperature is greater than many substances, i.e.,
it has a high heat capacity. This makes water a highly effective
heat transfer medium. Despite its high surface tension, the
viscosity of water which can be visualised as its 'thickness'
is low, making it easy to pump.
Water is non-hazardous and non-flammable, but it is sometimes
referred to as the universal solvent, as it will tend to dissolve
anything with which it comes into contact. Water containing
a low concentration of solids is corrosive and often behaves
highly aggressively, as it seeks to increase its solids content.
Hence, the purer, low solids content waters, such as occur
in Ireland and Scotland are far more corrosive than the chalk
rich waters of Southern England.
The water cycle
of the Earth's surface is covered with water, in places to
a depth of 11,000 meters. This tremendous volume of water
is never still. Besides the physical movement caused by tides
and marine currents, there is a continuous process of evaporation
and condensation, which we call the Water Cycle.
The essential process of the Water Cycle is absorption
of heat from the sun by bodies of water, be they sea, lakes
or rivers, which causes water to evaporate. The evaporated water
rises until the temperature is low enough for it to condense
into clouds of water droplets, which, when they become big enough,
start to fall as rain. As the rain falls, it dissolves anything
that is in the atmosphere mainly the gases oxygen, carbon dioxide
and nitrogen. Nitrogen is known as an inert gas because it does
not react with many things. As it does not react with metal,
nitrogen is not an important consideration in water treatment,
but oxygen and carbon dioxide are. They are the principal causes
Acid in rain
have heard of acid rain produced by absorption of sulphur
and nitrogen oxides by rainfall, but the presence of pollutants
like these gases is not the only source of acidity. Carbon
dioxide makes up 0.3% of the Earth's atmosphere and is readily
absorbed by water to form carbonic acid - the acid found in
carbonated beverages. Hence, all rain is acidic to some extent
and therefore has a strong tendency to dissolve minerals and
rocks with which it comes into contact.
By knowing the recent history of your water and the length
of time it has been in contact with soil and rock, it is possible
to predict its likely quality. In the Lake District and most
of Scotland, the water is soft because the extractions are
mainly of surface water and the contact time between water
and earth is short. Also there's not much chalk in the soil
to be dissolved. In the South East of England, there's a lot
of chalk and much of the water is pumped from underground
where there has been a longer contact time, so the water is
hard. The water used in a domestic central heating system
is normally drawn from the towns mains supply.
The quality of water in the British Isles is generally good,
but it must be remembered that the only statutory obligation
on the Water Authorities is to supply water fit for potable
use. Consequently, the water distributed via the mains system
is unlikely to be ideally suited to the central heating system.
British Standard 7593: 1992 categorises towns mains supplies
Hardness Concentration mg/i (as CaCo3)
Most calcium salts are inversely soluble, i.e.,
as the temperature rises they become less soluble. The higher the hardness,
the more likely the water is to produce a substantial lime scale when it
is heated. Comparison with the system water will indicate if scale has been
formed. A decrease of 100 mg/l of hardness means that 10 grams of scale
have been produced per 100 litres of water in the system. It should be remembered
that 100 litres of water, which has leaked out of the system [and been replaced]
will also have left 10 grams of scale behind. This may have remained in
the boiler heat exchanger.
We will discuss the consequences of the different qualities of waters
for the central heating system.
Scale in the system
people are familiar with the build up of scale on the elements of kettles
and immersion heaters in hard water areas in the country, however, scale
formation in closed circuit central beating systems is rarely considered.
In central heating systems, scaling is caused by the deposition of either
lime or black iron oxide corrosion debris. Lime scale consists of calcium
carbonate (calcite) with lesser amounts of other calcium salts such as
the sulphate. Sometimes lime scale deposits contain corrosion debris which
has become trapped.
Cause of lime scale formation
Unlike most substances, most calcium salts become less soluble as the
temperature rises (inverse solubility). Such salts are naturally present
to some degree in all mains water supplies, therefore, any system where
mains water is heated is liable to suffer from lime scale formation. Consider
a mains water containing 300 milligrams per litre (mg/l) of calcium carbonate
hardness. The potential weight of scale produced in a 100 litre central
heating system is 30 grams, and this is from only the initial fill of
water. Once formed, calcium scale tends not to redissolve, so water lost
from the system for whatever reason will leave it behind.
The fresh make-up water will then bring further calcium salts into the
system to begin the process again producing an ever increasing thickness
of scale. Being a direct result of increased temperature, lime scale will
form in the hottest part of the system, usually the heat transfer surface
in the boiler. In contrast, iron oxide first forms as a sludge at the
point of corrosion and will only become a scale if carried by the water
to the heat exchanger where it can become hardened by baking.
The problem of scale
The formation of scale in a central heating system has three important
The first of these is the problem of boiler noise or kettling.
Correct operation of the boiler relies on the removal of heat
from the transfer surface such that the temperature of the
metal skin and the re circulating water, which are the same,
remain below boiling point. Boiler noise is usually caused
by localised overheating of the re circulating water leading
to steam generation. This can happen in two ways:
a) If the velocity of the water is too low, the
water in contact with the metal will simmer or boil. When the resultant
steam bubbles reach the cooler water away from the surface they collapse
noisily. The violence of the reaction, which determines the noise level,
depends on the temperature of the water. The colder the water is, the greater
is the shock and the associated noise. The phenomenon is called 'kettling'.
b) Scale deposition is rarely uniform. Thicker areas will always exist
where the metal temperatures are higher. This will concentrate the bulk
of the heat transfer over small areas at above design heat transfer rates,
so causing simmering and boiling. Cooling and reheating of scale as the
system cycles will cause cracking and flaking of the surface giving rise
to boiling in crevices in the fractured deposit, either between deposit
and metal, or in the body of the deposit. This problem is dependent on
the amount of scale present, as once the quantity becomes very great,
the fractures will tend to 'heal' and the water become completely insulated
from the metal surface. At this time boiler efficiency will drastically
reduce and failure become a definite possibility.
Reduced efficiency The loss of boiler efficiency due to reduced heat transfer
is becoming increasingly significant in the central heating system.
The insulation afforded by an egg-shell thickness of scale can lead
to as much as a 10% rise in fuel consumption, and further build-up can
rapidly lead to overheating of the metal and potential failure. The
extent to which scale will affect efficiency depends on thermal conductivity,
e.g., a porous lime scale conducts heat 100 times less effectively than
boiler metal and has a very marked effect.
Increases thermal stress in metal In the extreme case, scale can build up to the point where
the restriction of heat flow can cause the insulated metal surface to
overheat, leading to stress failure.
Corrosion in the system Corrosion will take place to a greater or lesser extent in
all domestic central heating systems. Just how severe it is will depend
on many factors including the types of metal in the system; the degree
to which air can be drawn into the system; the nature of the supply
water. Even the quality of the installation work can have an effect.
The strength of the forces involved in the corrosion process can readily
be appreciated by considering the energy necessary to oppose them, e.g.,
to produce metal in a steel works, or in a smelter, from naturally occurring
ores. The purer the metals produced, the more unstable they are and
the greater their tendency is to revert to their natural state. the
more unstable they are and the greater their tendency is to revert to
their natural state.
Corrosion in the system
will take place to a greater or lesser extent in all domestic central
heating systems. Just how severe it is will depend on many factors including
the types of metal in the system; the degree to which air can be drawn
into the system; the nature of the supply water.
Even the quality of the installation work can have an effect. The strength
of the forces involved in the corrosion process can readily be appreciated
by considering the energy necessary to oppose them, e.g., to produce metal
in a steel works, or in a smelter, from naturally occurring ores. The
purer the metals produced, the more unstable they are and the greater
their tendency is to revert to their natural state. the more unstable
they are and the greater their tendency is to revert to their natural
of oxygen in the system is possibly the most important factor. It is often
wrongly assumed that the oxygen present in the initial fill will soon
be consumed by reaction with the metal surfaces and that once this has
occurred no further corrosion will take place. This is not the case, as
in practice systems are found never to be oxygen free.
At room temperature, water contains 8 milligrams of dissolved oxygen
in every litre, so an open vented system using make-up, or pumping over,
will steadily admit oxygen. Even where no make-up is used, the water at
the surface of the feed and expansion cistern will be saturated with oxygen
and an oxygen gradient will extend down through the tank and the cold
feed into the circulating water.
Oxygen ingress will also occur at pumps, joints and valves. At points
under negative pressure, i.e., those between the neutral point
(the point where the cold feed is connected to the re circulating
system) and the suction side of the circulator, depending
on the soundness of the connections, ingress can be rapid.
It should be noted that negative pressure points can also
be generated in sealed systems.
Even where there is a positive pressure, yet no leakage of water, there
may still be contact with the atmosphere the surface tension of the water
may be preventing any flow through capillaries. Under such conditions
oxygen will enter the system.
In its simplest form, the reaction between iron (steel) and oxygen in
the presence of water produces iron oxides (initially rust) and hydrogen
gas. The effect of hydrogen production is to depress the water level in
the radiators, so producing a cold area at the top. Sufficient pressure
can be produced to split aluminium radiators. The gas will often end up
in one radiator at the point of lowest pressure, furthest from the circulator.
The gas is slightly soluble in water and, although produced elsewhere,
will transfer to the point of least resistance.
Hydrogen is a colourless, odourless gas . It is not corrosive, but it
is lighter than air and flammable (take care when venting). When gas build
up is detected, the first step is to test for the presence of hydrogen
- the gas may just be air. To do this, the radiator should be vented into
an upturned glass container, e.g., a jam jar. Being the lighter gas, hydrogen
will displace air from the jar. The jar should then be removed from the
radiator and a lighted taper inserted into it. If hydrogen is present,
the gas will ignite with a pop (without risk of explosion), and it can
be concluded that corrosion is occurring.
pH is a numerical indication of the intensity
of acidity or alkalinity of a solution. The pH scale is logarithmic and
runs from 0 to 14, with 7 being neutral. Low numbers are acidic and high
numbers, basic (alkaline). A pH of 4 is ten times more acidic than a pH
of 5, and a pH of 3, one hundred times more acidic.
Metals are affected differently by pH conditions, e.g., mild steel is
most stable from pH 10.5 to 11.5, yet aluminium, always coated with a
layer of aluminium oxide, is readily attacked above a pH of 8.7. Copper
may be adversely affected by a pH greater than 9.5. The ideal pH for the
domestic central heating system is between 6.5 and 8.0, when the corrosion
rates for all metals are acceptable.
metals are in contact a galvanic (electrochemical) cell is created. One
metal will go into solution, determined by the metals' relative positions
in the Electrochemical Series. This series is a law of nature and a property
of the metal in the form in which it is used. Copper is the most noble
metal used in the domestic system, so it never corrodes, but all other
metals will sacrifice in its presence. Aluminium is particularly susceptible
being far away from copper in the table, i.e., a very high corrosion current
is generated. (N.B. pH plays an important role in the case of aluminium,
due to the oxide coating). The sacrificial metal wastes away at the closest
contact point. Insulation minimises the damage, but even fibre washers
are not foolproof as they absorb salts from the water and can then conduct
The metal most commonly corroded in a domestic system is mild steel in
radiators. Here, depending on the amount of copper present, the corrosion
can take two forms:
(a) Where a large area of a radiator becomes plated with copper (as can
happen in new systems where flux has been abused), inevitably there will
be points within the area of copper not covered. Galvanic corrosion then
occurs with the current being concentrated at the point where the sacrificial
metal is in contact with the water. The current is intense, and as all
the corroding metal emanates from the same point, perforation is rapid.
The hole on the outside of the radiator is the same size as that on the
inside, as though the metal has been drilled. Such corrosion is termed
(b) Small pieces of copper can adhere to the
radiator surface, e.g., swarf. Galvanic corrosion will also be induced,
but now the current is concentrated at the noble (non-sacrificial) metal
end. This results in wasting of metal across much of the mild steel surface.
If perforation ever does occur, examination of the internal hole will reveal
a pyramidal shaped pit, whose apex just touches the surface of the radiator.
This is 'general' corrosion.
reactions proceed more rapidly as the temperature increases, hence, a
central heating system will corrode faster than a similarly constructed
chilled water system. The effect of temperature is not, however, restricted
to reaction rate. Most installers are aware that galvanised metal should
not be used in central heating systems, but the reason is not widely appreciated.
If the steel pipe, or surface, is well coated with a tightly adherent
layer of zinc, the degree of protection afforded will be good. However,
the surface is often scratched or pitted, allowing the water to contact
the steel below. The Electrochemical Series indicates that steel is more
noble than zinc, so corrosion of the steel will not occur. At temperatures
below 60-65°C this is indeed the case, however, at higher temperatures,
the relative potentials of mild steel and zinc reverse, making steel the
sacrificial metal, so inducing pinhole corrosion. Galvanised feed and
expansion cisterns are particularly susceptible.
Points of stress, e.g., welds, bends, etc., are always more
liable to corrosion. Certain salts such as chloride will attack
stressed areas along the grain boundaries in the metal. Sources
of high chloride include inefficient water softeners, washing
up liquids (misguidedly used to cut the noise from noisy boilers)
and excessive flux application.
Fouling in the central heating system can be biological, organic or silt
based, or more often a combination. Foulants include bacteria,
slimes, scale and corrosion debris. Once on the metal surface
the insulating layer creates a local dissolved oxygen gradient
which promotes under-deposit corrosion - usually in the form
of pits. The very low level or complete absence of oxygen
under the foulant also encourages the growth of anaerobic
(oxygen hating) bacteria such as sulphate reducing bacteria
(SRB's). Such bacteria produce sulphuric acid as part of their
metabolism, with obvious consequences for metal in the vicinity.
John R Lane, EurChem C.Chem M.W.M.Soc MRSC B.Sc ARCS