int. j. for the history of eng. & tech., Vol. 84 No.

1, January 2014, 1–29

Design and Development of Exhaust

Valves for Internal Combustion Engines
from the Perspective of Modern
Thinking: Part 2 1930–90
Fred Starr

The changes in high temperature materials for exhaust valves resulted from
the development of the engine fuels. Ricardo and Midgley were the early
pioneers of fuel improvement, but later on Houdry and Pines invented pro-
cesses for upgrading straight run gasoline. None of the early alloys used for
valves was specifically invented for that purpose. Marsh developed nickel-
based alloys for a thermopile; Haynes, the Stellite series of cobalt alloys
as ultrahard and corrosion-resistant materials; Brearley 12Cr martensitic
stainless steels for erosion resistant gun barrels.
In the 1920s, the automotive sector began to use Silchrome, which had
good resistance to non-leaded fuels. However it was barely adequate in aero
engines, and this stimulated the invention of sodium-cooled valves by
Heron. Fuels containing tetraethyl lead, invented by Midgley, became stand-
ard during the 1930s, but lead from the combustion process was highly
corrosive to Silchrome. A British austenitic alloy, KE965, proved to be much
better, but during WWII this had to be given high nickel coatings. The
Americans developed a material similar to KE965, but for high output radial
engines used Inconel M and Stellite coatings.
The post-war era has been dominated by the automotive sector. Low nick-
el austenitics superseded KE965, but Silchrome is still in use for inlet valves.
Zero lead fuels caused problems with ‘valve seat recession’ which has been
overcome by induction hardening of cast iron seats, or by incorporating seat
inserts. High performance cars have adopted Inconel and Nimonic alloys
and sodium cooling, but lightweight titanium alloys are being used to give
the ultimate in capability.

keywords poppet valves, IC engine fuels, Silchrome, KE965, Stellite, nickel-

based alloys, sodium cooled valves, exhaust valve corrosion, valve seat
recession

© The Newcomen Society for the Study of the DOI 10.1179/1758120613Z.00000000030

History of Engineering & Technology 2014
2 FRED STARR

Engine designers and fuel developers

The important people of this period are Ricardo, Heron and Midgley. The work of
Ricardo has been covered in Part 1 of this paper.1 His contribution was establish the
relationship between fuel composition, its propensity to knock and power output,
initially quantifying the fuel in terms of a Toluene rating.2 But Ricardo saw how,
provided some sacrifice was made to power output, the side-valve concept, when
combined with the squish head combustion chamber, could allow car engines to run
successfully on the low-octane fuels available to the motoring sector. Ricardo devel-
oped a single-cylinder test engine with a variable compression ratio which could be
used to evaluate fuels (Figure 1). When fuels began to be classified in terms of octane
rating, this type of equipment made it possible to correlate octane rating with a
‘Performance Number’ (PN) which a direct measure of potential power output from
a given fuel compared to that of isooctane.
Heron was, in some respects, similar to Ricardo, having been at one stage an engine
designer and not just a person who specialized in the development of components
(Figure 2). When he worked at the Royal Aircraft Factory (later to become Establish-
ment) in the UK with Gibson, Heron had formulated a more rational approach to
cylinder head design. When he moved to America, he used it in the Wright J5 Whirl-
wind radial engine. Heron’s early thinking on air-cooled engine design is represented
in a paper he gave on air-cooled automotive engines.3 He pointed out that good
design of the valves themselves could overcome deficiencies in materials. For example,
tulip-shaped valves with thick rims could run successfully, even when made of tung-
sten steel which ‘burn readily if the design is poor’. Thick rims permitted heat to be
conducted away from the sealing surfaces; the tulip shape formed a kind of pocket in
the valve face, which reduced the heat input at this point. However, as will be seen,
his most important contribution was to develop a practical method of cooling valves,
an essential feature of all aircraft engines after 1930.4
The third person who played a key role was Thomas Midgley Jnr, whose name is
now reviled because of his work on TEL (tetraethyl lead, Pb (C2H5)4) as a knock
suppressant, and his development of chlorofluorocarbon compounds for refrigeration
systems. What is forgotten is that TEL extended the world’s oil reserves, and
domestic refrigerators saved money and lives in preventing the spoilage of food.
Midgley was told to work on the project by Charles Kettering at the Dayton
Engineering Laboratories who was looking for something that would improve the
knock resistance of kerosene (Figure 3). The idea that knocking could be suppressed
by small amounts of additives ran counter to the opinion of Ricardo, who was
content to rely upon the beneficial effects of comparatively large amounts of cyclic
hydrocarbons. But in the USA much of the gasoline was of very poor quality and
its improvement by additives was a priority. Much of Midgley’s work had an Ediso-
nian aspect, in which scores of potentially interesting compounds were tried. The
route by which TEL was discovered was reasonably rational, given the thinking at
the time. Knocking was ascribed to the heating of the ‘end gas’, an uncombusted
mixture of air and petrol vapour, by radiation from the part of the mixture that was
already burning. The idea was to put something into the petrol that would absorb
the radiation. Iodine was tried, giving some improvement, leading to a set of organic
compounds being synthesized by Midgley’s research team containing elements
EXHAUST VALVES FOR INTERNAL COMBUSTION ENGINES 3

figure 1a Sir Harry Ricardo (Ricardo Library).

figure 1b The variable compression test engine

(Ricardo Library)
4 FRED STARR

figure 2a Samuel Heron standing by his

air-cooled cylinders. Aircraft Engine Historical
Society Inc.

figure 2b Wright J5 Whirlwind engine which

Heron helped design. Aircraft Engine Historical
Society Inc.
 EXHAUST VALVES FOR INTERNAL COMBUSTION ENGINES 5

figure 3a. Thomas Midgeley,

Jnr. Kennesaw State University.

figure 3b The structural formula

of Tetraethyl Lead. The lead atom
is at the centre and joined to four
sets of C2H5 groups. Kennesaw
State University.

such as selenium, tellurium, tin and lead.5 The breakthrough with TEL was soon
recognized and in 1923 Midgley was awarded the Nichols Medal by the American
Chemical Society.

Inventors, metallurgists and metallurgical developments

To be a successful inventor is to see that there is a social/technological need which is
not fulfilled by the equipment we have at the moment. A good inventor recognizes
that by making use of current technology and engineering insights, plus a limited
6 FRED STARR

amount of development in a specific area, the invention is likely to be, at the very
least, a technical if not a commercial success. Inventors of this type have been Edison
with the light bulb and Sir Frank Whittle with the jet engine.
Scientists have who are at the leading edge of their discipline have a rather different
aim. They come to see that existing explanations of natural phenomena are
inadequate. Improved theories need to be formulated, which should have a strong
predictive element. Newton’s Law of Gravitation predicted that the earth would be
flattened at the poles; Maxwell’s electromagnetic theory suggested that radio waves
could be produced from a high frequency alternating current.
Metallurgists, working as industrial scientists, have a similar attitude to inventors
in that there is a social/technological impetus to their investigations. Their work is
scientific only in the sense that it involves a ‘principal idea’ which it is believed
will improve the strength, corrosion resistance or workability of metallic materials.
Research metallurgists then formulate new alloys on the basis of the principal idea.
There is a need to act in this way since the scientific basis for the work is often quite
undeveloped. Even if there is a good theory, the hard data to enable it to be used is
often absent. It is an ‘engineering science’ approach.
By 1900 it was beginning to be recognized that phase diagrams could, in principle,
help with the formulation of new alloys. How useful was this knowledge is question-
able, as only the simplest phase diagrams were available, but it would have provided
some insights, particularly in terms of heat treatment. However, the prospects
for commercial success in developing new alloys were much more realizable than
when Faraday had begun work on new steels. What changed the situation was the
availability, in a reasonably pure form, of chromium, tungsten, vanadium, silicon and
molybdenum. A key feature was that the carbon content of these ‘reasonably pure’
materials was low, as excessive levels of carbon make alloy steels brittle, and under-
mine corrosion resistance. Nickel and cobalt, also newly available, could also be used
in the development of steels, but of greater significance, especially in terms of materi-
als for exhaust valves, was that both of these elements, themselves, were to be the
basis of new groups of alloys.
The other factor in the development of successful alloys is temperature control
during forging and in annealing and tempering heat treatments. Given that the basic
scientific understanding was lacking, it is amazing that metallurgists had any success
at all. Indeed, the history of the nickel based ‘superalloys’, the earliest of which were
used for exhaust valves, seems to have been a series of near misses. These seemed to
be largely caused by mistakes with the alloy content and problems with formulating
a successful high-temperature treatment, known as ‘ageing’.

Exhaust valve serendipity — Nichrome, Stellite and the first stainless

steels
There was no deliberate effort before 1945 to ‘design’ alloys for exhaust valves. But
once Goldschmidt was able to produce pure carbon-free chromium using the thermite
process, it would only have been a matter of time before corrosion and oxidation
resistant alloys were developed. However Marsh (Figure 4), in the United States, set
out to invent an alloy which had to have good oxidation resistance, as it was to be
used in a thermopile for the generation of electricity. As his patent shows, a more
 EXHAUST VALVES FOR INTERNAL COMBUSTION ENGINES 7

figure 4 Alloy inventors Albert Marsh (Nichrome), Elwood Haynes (Stellite) Harry Brearley
(stainless steels). Author’s collection.

realistic market was the use of the alloy as a low conductivity resistance wire.6 But
it was his co-worker, and the man who provided money and facilities, William
Hopkins, a prolific and successful inventor, who realized that the material would be
ideal for domestic toasters.7 At that time, R&D in the United States seems to have
been geared to finding new products for the domestic market; in Europe the arms
race provided the incentive for materials development. The alloys developed by Marsh
contained about 80% nickel and 20% chromium and were marketed as Chromel,
Nichrome and Brightray resistance wire.
At about the same time as Marsh was working in Chicago, Elwood Haynes, in
Kokomo, Indiana, was trying to develop a new series of alloys. Haynes, even as a
teenager, had been interested in metallurgy and had written a thesis on iron-tungsten
alloys. As a young man he continued his interest in technology, helping to develop
America’s first long-distance gas pipeline and then building one of the first successful
automobiles. For the new alloy, high hardness combined with resistance to rust was
the target.8 Haynes initially worked on both nickel- and cobalt-based alloys, but to
avoid conflict with Marsh, he concentrated on cobalt, after the patent situation
had been resolved. Cobalt when combined with additions of chromium, carbon
and tungsten gave tough, very hard, corrosion-resistant materials. Because of their
brilliant and star-bright lustre, they were sold under the trade name of Stellite. Only
later did it become clear that they had good high temperature resistance. However,
because of the cost, the Stellites were only used for hard facing of the seats and faces
of the valve assembly, where the combination of hardness and resistance to lead
attack was critical.
Brearley, when developing steels for gun barrels at Firth Brown in Sheffield, is
generally held to have discovered stainless steels, if only because he was the first to
recognize the commercial prospects. He noted the good corrosion resistance when he
was trying to etch specimens for metallographic examination.9 The high-temperature
properties would have emerged quite quickly, once materials began to be heat treated,
rolled and forged. Hence, although the initial market was for high quality domestic
knife blades, the British recognized its value as a material for exhaust valves.
8 FRED STARR

The rise of Silchrome — the alloy for the Roaring Twenties

Although the Brearley alloy was used in British aircraft engines in WWI, it was soon
superseded by Silchrome for both automotive and aircraft applications, developed by
Percival E. Armstrong of the Ludlum Steel Corporation.10 Silchrome was first taken
up in the mass production automotive sector where its 3% silicon and 8% chromium
contents gave significant cost savings. At high temperature the silicon forms a very
thin sub-layer of silica (SiO2), which provides an extremely effective second line of
defence beneath an outer layer of chromia (Cr2O3).
Nevertheless, in 1918 oxidation resistance was not considered to be a priority. The
perceived benefit was that the high level of silicon and chromium in Armstrong’s
Silchrome raised the transformation temperature, at which martensite reverts to
austenite, to well above the temperatures at which automotive type valves normally
operate. This reduced the risk of distortion and embrittlement resulting from
metallurgical phase changes.
Silchrome was on the borderline as an aircraft valve alloy. In 1931 at the Sympo-
sium on the Effect of Temperature on Metals, Boegeld (from General Motors) and
Johnson (from the United States Air Corp.) stated that, ideally, the tensile strength
of a valve steel should be 25,000 lb/in2 at 1600°F (172 MPa at 871°C).11 In the discus-
sion, Heron pointed out that this was simply a useful guide to the assessment of new
alloys. He emphasized that engines should not be operating with valves running so
hot. Pre-ignition would be certain. In fact, British work at about that time conducted
by J. R. Handforth showed that even the strongest of Silchrome type alloys were
falling well below this target. Silchrome could offer 172 MPa, but only at 800°C,
seventy degrees less than what was desired.12 It was the shortcomings of Silchrome
which led Heron to develop the sodium cooled valve, a subject which is discussed
later.

Austenitic stainless steel — the 1930s material of choice

The new entrant into the exhaust valve market was austenitic stainless steel. The
earliest commercial austentics were developed, in 1925, by W. H. Hatfield, Brearley’s
successor at Firth Brown, building on earlier work done in Germany. From this came
the standard 18/8 stainless steels containing 18% chromium and 8% nickel, which
are used around the house for sink tops and stainless spoons and forks. The rela-
tively high nickel content has the effect of making austenite stable over the complete
temperature range, so there is no critical temperature at which the material suddenly
weakens. However, Hatfield’s company seem to have specialized in the development
of materials that were resistant to aqueous corrosion and it was left to another Brit-
ish company, Kayser Ellison, to develop an austenitic stainless steel, KE965, with
extremely good high-temperature properties. KE965 became the standard exhaust
valve material in Britain up to the 1950s.
As these alloys are fully austenitic, they do not soften at temperature, as do the
12% chromium and Silchrome types of alloy, whose martensitic microstructure is
basically unstable. To give hardness and strength, tungsten was added as alloying
element and the carbon content was at a much higher value than with stainless steels
with good resistance to aqueous corrosion. These two elements react to form
carbides, strengthening the basic alloy. Some of these reactions will occur when the
valve is in use and running hot.
 EXHAUST VALVES FOR INTERNAL COMBUSTION ENGINES 9

The quest for range or speed

Engine design and the sodium-cooled valve

The use of ‘or’ rather than ‘and’ in the title of this section is deliberate. The attitudes
to aviation were different in America and Europe.13 In the USA, distances between
major centres of population (for example, New York–Chicago, or Chicago–San
Francisco) encouraged the development of reasonably fast but safe and economic
commercial air services. On this side of the Atlantic, countries were small enough for
railways to be adequate. However, national rivalries were intense and were at the
root of competition for speed (evidenced by World Speed Records and the Schneider
Trophy). Later, the fear of war led to the design of fighter aircraft with the highest
possible performance.
Economy, reliability and weight caused the Americans to begin the design of an
outstanding series of air-cooled engines, namely the Pratt and Whitney Wasp and
Wright Cyclone series. In Europe, the perception was that the air-cooled radial engine
produced so much drag it would limit fighter speeds to less than 300 mph (480 km/h).
The liquid-cooled engine was easier to streamline and, with careful design, the
radiator could produce jet thrust.
Heron’s invention of sodium cooling was critical to the success of the American
radials and was hailed by Cowell of TRW as being the ‘greatest single step’ in valve
design.14 Sodium cooling reduced aircraft valve temperatures to below those of truck
and bus engines. According to Cowell these could be running at the incredibly high
temperature of 870°C. The other two developments in valve design and technology
identified by Cowell were (a) forging techniques for manufacturing a one-piece,
hollow-head valve and (b) general improvements in air-cooled cylinder head design.
Figure 5, taken from his paper, shows how the Wright exhaust valves had evolved
over a fifteen-year period. Pratt and Whitney valves changed in a similar way. The
stem of the valve on the top left of the figure was drilled for lightness only; it con-
tained no coolant. The next three were salt-cooled, but Cowell states that the salt, a
mixture of lithium and sodium nitrates, decomposed at temperature and its use was
quickly discontinued. The four valves in the bottom row were all sodium-cooled.
In terms of materials, the valve on the top left was a cobalt-chrome alloy, essentially
a very high carbon, 12% chromium steel, with 4% cobalt. It rusted so much in a
marine environment that it was soon dropped. Any rust resistance benefit from the
chromium would be undermined by the reaction with it of the incredibly high level
of carbon. All of the other valves were fabricated from the TRW alloy, TPA, essen-
tially a Chinese copy of KE965 apart from a reduced silicon content. The last picture
in the series shows a valve in which faces have been Stellite coated. In this context it
is noteworthy that the face temperature was determined by TRW as being in the
range 590–605°C, quite a low value. But judging from some work by NACE in 1943,
engine outputs had risen to the point where, even with sodium cooling, valve life was
becoming critical, with valve crowns becoming badly corroded and collapsing on one
type of engine. This is not too surprising since valve temperatures were recorded as
being up to 725°C.15
10 FRED STARR

figure 5 Evolution of Wright radial engine valves (from Cowell)16.

Undoped gasoline fuels and valve materials

Henshaw, Technical Director of Kayser Ellison and Co. who developed KE965, gave
a paper on exhaust valves to the Royal Aeronautical Society in 1926.17 This described
a variety of alloys, many of which contained reasonable levels of chromium, indicat-
ing that the importance of good oxidation resistance had, at last, been recognized.
Henshaw began his paper by stating
It is quite evident that that the valve steels which have been regularly used are not
satisfactorily meeting all the present conditions, and therefore advances will have to be
made.

These comments were made when in Britain aircraft fuels were becoming standard-
ized, consisting of gasoline, plus about 20% Benzole, raising the octane rating to
about 70. The products of combustion would have been fairly innocuous in terms
of corrosivity, but somewhat more aggressive than that of air, on account of the
presence of steam, carbon monoxide and dioxide and, when an engine was running
rich, hydrocarbons. Many tests of oxidation resistance were done in laboratory
furnaces fired by town gas, so the results would have been a fairly good indication
of how materials would behave in practice. Henshaw’s test times were apparently
only up to 144 hours, which, although very short by today’s standards, would have
been adequate, given the expected valve replacement life.
 EXHAUST VALVES FOR INTERNAL COMBUSTION ENGINES 11

Henshaw was able to show that there was a temperature at which scaling became
excessive for each material. Table 1, based on his paper, shows that there is a rough
relationship between chromium content and the scaling temperature. The exception
is a Silchrome type alloy, which in these short-term tests has outstanding resistance,
but is only effective where the corrosion comes from gases rather than molten salts.
Although there was more interest in the oxidation of valve steels than when
Aitcheson had done his work, most of Henshaw’s paper focuses on mechanical
properties. High temperature metallurgists and designers had begun to realize that,
when materials are put under stress at high temperatures, they gradually stretch or
‘creep’ over time. Eventually, they will fail by cracking. The first major paper on
creep had been published by Dickenson just four years previously, the purpose being
to identify steels for ammonia synthesis plants.18 Dickenson’s tests lasted for several
thousand hours and showed conclusively that creep strengths were very much less
than in the standard, short-time, high-temperature tests. Nevertheless, it was hard for
him to get over the point that a short-time tensile test could be quite misleading.
Although the concept of creep was becoming accepted, it is not surprising that
some of the ideas expressed in the discussion following Henshaw’s presentation seem
somewhat strange, given the knowledge of today. Roy Fedden, designer of the Bristol
Jupiter, thought that efforts should be made to establish the stress at which, in an
eight-hour test, no creep would occur, since the duration of most flights was less than
six hours. In practice creep rate varies over the length of the test. Even with pure
metals, the rate changes with time, the various stages being termed primary, second-
ary and tertiary. With complex alloys, the growth and disappearance of precipitates
(some of which can be metastable) affects the inherent strength of an alloy, so the
creep rate is bound to vary during the life of a component. Nevertheless there are
some grounds for supporting Fedden’s view. What he was hoping for was a material
with a short time creep rate that was low enough to preserve tappet clearances,
typically 10–20 thou (0.25–0.5 mm), between take off and touch down.
More serious was the apparent belief that some valve materials could self-harden
and thereby self-repair in service. The basis for this supposition was that the
temperature of the valve head would be so high that, once the engine stopped, the
valve would go through a quenching treatment. Note that one of the claimed
advantages of Silchrome was that it was stable enough to avoid these effects!

TABLE 1
TEMPERATURE AT WHICH EXCESSIVE SCALING BEGINS

Material Main Alloying Elements Scaling Temperature

Silicon Chrome Fe-8Cr-3Si-2W 1000°C

Chrome Steel Fe-10Cr c. 800°C
High Nickel Chrome Fe-13Cr-13Ni-2W 950°C
Cobalt Chrome Fe-13Cr-4Co-1.5C c. 850°C
Stainless Steel Fe-13Cr-1Si 825°C
3% Nickel Chrome Fe-3Ni-1Cr 650–700°C
High Speed Steel Fe-19W-4Cr 725–750°C
12 FRED STARR

By the time of the 1931 presentation by Handforth to the Iron and Steel Institute
(now the Institute of Metals, Minerals and Mining), KE 965 was being marketed.19
Handforth’s paper showed a great many microstructures of other austenitic alloys
in which the samples had been heated for 1000 hours at temperatures between
750–950°C.
In terms of the discussions going on in America about the shortcomings of
Silchrome, KE965 had a tensile strength of 277 MPa at 800°C which was at the limit
of Handforth’s tests. But it can be shown that KE965 would have had a strength of
over 200 MPa at 871°C, easily beating Heron’s target. Hence it is not surprising that
in Great Britain, KE965 became the standard alloy for exhaust valves, in both the
aircraft and the motor car industries. The American austenitic, CNS, being promoted
at the time, seems to have been designed with cost in mind. The nickel content was
lower than KE965 (see Table 2) and this, for reasons connected with the stability of
austenite, precluded the ability to add tungsten as a strengthening element.
Handforth was concerned with the mechanical properties after long periods
of heating since some materials, including CNS, became embrittled. KE965 was
relatively immune and this was probably another reason why it became the standard
alloy in the UK. In actual fact the risks of embrittlement in austenitic alloys are
likely to be more serious after heating for long periods at about 650–700°C, but this
was not known at the time. The risks can be reduced by keeping the nickel at a high
level and by reducing the silicon content. Eventually Thompson Products Inc.
developed the alloy TPA, CNS being dropped (Thompson became TRW Inc. in 1958
— Thompson Ramo Wooldridge Inc. — as a result of expansion of the business).

Octane rating and tetraethyl lead

In the discussion which followed Henshaw’s 1926 paper, there was a question about
the effects of ‘dopes’ containing tetraethyl lead. (The origin of the word dope may
have come from the fatalities at the plants which produce TEL.) It was a question
unanswered, although word must have got through to the UK that TEL, although
benefiting octane rating, was having a very adverse effect on alloys of the Silchrome
type. There was mention of TEL by Handforth in 1931, in which the view was taken
that there was nothing much to worry about, even with Silchrome, providing that the
valves did not run too hot. Banks, who at that time was the Technical Representative

TABLE 2
COMPOSITION OF PRINCIPAL EXHAUST VALVE ALLOYS IN USE IN THE 1930S

Alloy Type Cr % Ni % Si % C% W% Mn %

Silchrome Mart 8.5 – 3.0 0.45 – 0.4

Silchrome XB Mart 21.0 1.5 2.0 0.4 – 0.7
CNS Aus 12.7 7.7 2.5 0.3 – 0.3
TPA Aus 14.0 14.0 0.55 0.45 2.4 0.5
KE965 Aus 14.0 14.0 2.0 0.42 2.5
 EXHAUST VALVES FOR INTERNAL COMBUSTION ENGINES 13

in the UK for the Ethyl Corporation, in a written submission to Handforth’s paper

stated that most problems had occurred in older designs of engines which had been
uprated.20
The question of octane rating was more critical in America than in Europe, as
gasoline from the Eastern US fields had an octane rating in the 40–50 region.
Although it was known that the addition of 20% benzole could greatly improve
octane ratings, the issue for the US Navy was whether this addition would be
available throughout its sphere of operations, so that they were the first users of TEL.
Similar considerations, about fuel availability, caused the UK Air Ministry to request
that aircraft engines should be designed on the basis that the fuel would have
an octane rating of 70–72, which could be made available throughout the British
Empire.
TEL burns to produce lead oxide which melts at 880°C, particles of which stick to
the exhaust valves. To counter this ethyl bromide or chloride was also added to
enable the lead deposits to be volatilised out of the engine. In practice, there was still
a build up of lead rich deposits (PbO and PbSO4) on the valves and seats. With valves,
if the temperature was sufficiently high the lead compounds reacted with the thin
oxide scale that would normally protect the metal. The lead deposits acted like the
flux in brazing operations. The subject is covered in more detail in a paper given by
the author to EuroCorr 2000 Conference.21
The fluxing attack of lead salts on exhaust valves was, perhaps, the first encounter
by metallurgists of this form of corrosion. Given the knowledge of the time,
there was nothing that could be done other than to try different alloys, and this
approach, plus restriction of valve temperatures through sodium cooling, was largely
successful.
Despite what Handforth and Banks had said about the ability of Silchrome to resist
attack by TEL, an iron-based alloy which relied on a combination of chromium and
silicon would not be able to resist attack. The melting or softening point of a lead
oxide and silica mixture can be as low as 550°C. But, given that TEL is no longer
used as an additive, it is unlikely that we shall ever have the full story why some valve
alloys are resistant and others are not.
Fortunately KE956, though initially having been marketed on the basis of its high
temperature strength, turned out to be much more resistant to TEL than Silchrome.
The American alloy TPA, had a much lower level of silicon, perhaps to give better
lead resistance as well as to minimize in-service embrittlement.
In the mid-1930s, the addition of about 1 part in a 1000 of TEL (4 ml/gallon) raised
the octane rating of fuels used by the RAF from 77 to 87 octane. It allowed the ground
boosting of engines (i.e supercharging at sea level), greatly improving take-off per-
formance and climb rate. Test figures indicate that power output could be increased
by more than 40%. For the Hawker Fury biplane fighter, the sea level rate of climb
was increased from 2900 to 4500 ft/min (Figures 6 and 7).22
For the commercial sector, Kessler, of Royal Dutch Shell, highlighted the huge
benefits which could be obtained from 100 octane fuel (in 1938 it was becoming
available, but only at a premium price). It was predicted that after a flight of 12
hours, representing a flight from London to Moscow, payload would be virtually
non-existent when the standard 73 octane aviation fuel was used.23 With 100 octane,
14 FRED STARR

figure 6 Hawker Fury Fighter Biplane (Author).

figure 7 Hawker Fury climb rate graph. (Redrawn by Author from Kessler24).
 EXHAUST VALVES FOR INTERNAL COMBUSTION ENGINES 15

TABLE 3
EFFECT OF OCTANE NUMBER ON PAYLOAD

Octane Rating 73 ON 77 ON 87 ON 100 ON

Payload after 4 Hours Flying Time 4220 kg 4630 kg 5440 kg 6080 kg

Payload after 12 Hours Flying Time 230 kg 780 kg 2170 kg 3190 kg

payload would rise to more than three tonnes (Table 3). These improvements came
from a 70% reduction in engine weight, and from fuel consumption falling by 25%.
For these reasons, Pan American, who were developing transatlantic flights, were the
first to adopt 96 octane fuel. (Their flights were not non-stop, going via Shannon and
Bermuda.)
By the late 1930s the RAF had standardised on fuel with an octane rating of 87.
But even this represents quite poor gasoline by the standards of today. It has a per-
formance number (PN rating) of about 70, meaning the output of an engine designed
to run on 87 octane fuel would only be 70% of that of a similar engine designed for
100 octane. Although much could be done with TEL, the basic properties of the
gasoline had to be increased.
The Houdry catalytic cracking technique was the first process to come on line. The
fuel it produced was limited to an octane rating of about 80. TEL had to be added
to reach a 100 octane standard. The Houdry technique is reputed to have been the
origin of the stocks of 100 octane petrol accumulated by the RAF in preparation for
the Battle of Britain, although the formulation had to be modified to give a rich mix-
ture rating by the incorporation of ring-type hydrocarbons. This fuel was tankered
over from the USA in secret. The Rolls-Royce Merlin needed little modification to
run on 100 octane grade. All that was required was a change in the throttle settings
to allow the supercharger to provide more boost pressure.
The really big improvement in fuel octane numbers came from the alkylation pro-
cess, which had been discovered by H. Pines, almost by accident, when he was doing
a routine chemical analysis, while working at Universal Oil Products in Des Plains
in 1932.25 The reaction takes place at around room temperature in the presence of
sulphuric acid, and involves building up the isooctane molecule from two simpler
constituents. Alkylation plants came on line around 1940 and produced a fuel which,
when the engine was running rich (that is at the highest possible power), had a
performance number up to 145.26

Longer term pre-war developments — high strength nickel-based alloys

and coatings
During the 1930s, when work began in various countries to improve Nichrome,
exhaust valves were the priority. The gas turbine was a potential user, but was a
dream of a few visionaries. Indeed the perception was that practical gas turbine would
require materials to operate at temperatures around 800°C rather than 700°C. This
higher operating temperature and the fact that the gas turbine was just a gleam in the
eye of its protagonists would have been a deterrent to investing in both new alloys
and gas turbine concepts.
16 FRED STARR

Nevertheless it would have been common knowledge that nickel- and cobalt-based
alloys had inherently good resistance to high temperatures. The USA was first in the
field with the Inconel series where the presence of titanium proved to have a dra-
matic effect on high temperature strength. However, as sometimes happens when
efforts are made to improve on an original invention, several players entered the field.
In France, Chevenard made significant strides in understanding the strengthening
mechanism. Work also began at the NPL, which eventually led to the Nimonic series
of alloys which were the basis of the British, post-war, jet-engine programme. A
thoughtful review of the work in the UK and on the Continent on nickel alloys and
other high temperature materials was given by Allen (who had worked at the NPL)
at a major conference on gas turbine materials in 1952.27
All this work made use of the greatest accidental discovery in the history of high
strength materials, the concept of precipitation hardening, a strengthening mechanism
discovered by chance by Wilm in Germany in 1906, when working on aluminium
alloys. After making up some alloys, he left them for a few days and then, to his
surprise, found them to be much stronger than he expected. Much later, in 1930, it
was determined that the improvement was due to a precipitate of a copper-aluminium
compound, which had formed while the alloys had literally been standing around in
the laboratory.
By the early 1930s the NPL and others realized that the same approach might be
used to strengthen nickel-chromium alloys. In this case the precipitate was a complex
nickel-aluminium-titanium intermetallic. Although the precipitation phenomenon in
aluminium alloys occurs at room temperature, nickel alloys have to be held at about
700°C for several hours for this to occur.
The identification of a successful precipitation or ageing treatment was the culmi-
nation of a series of time-consuming trials, including getting the correct aluminium
and titanium content, casting at high temperature, hot forging, and solution treat-
ment at 1080°C. Failure at any one of these steps would have broken the sequence.
The precipitate diameter is in the range 0.01–0.1 microns, a size well below the point
where it can be seen with an optical microscope — its presence could only be inferred.
Heat treatments and modifications to the composition produced no visible effects,
only rather exasperating changes to the hardness, by which it might be possible to
make an intelligent guess at what was going on.
It is not surprising that it took a long time for a successful alloy to emerge. Eventu-
ally Henry Wiggins Ltd in Birmingham was given the responsibility for producing
a commercial material. Dr Pfeil, responsible for R&D at Henry Wiggins, described
how the company had had to work through the extremely confusing literature on
high temperature iron- and nickel-based alloys before a sensible programme could be
devised.28
By the time Nimonic alloys began to be produced in quantity, the jet engine had
arrived and there was little interest in improved valve alloys. Hitchcock has told the
author that, when he worked at Wiggins, he did get Rolls Royce to try out some
Nimonic valves in the Merlin29 but Morgan states that solid valves in this material
did not perform well.30 It was subsequently successfully used for seat inserts in the
cylinder heads.
Over in the USA, the demands on high output, turbocharged, air-cooled radial
engines in World War II were such that their valves were manufactured from Inconel
 EXHAUST VALVES FOR INTERNAL COMBUSTION ENGINES 17

TABLE 4
SUPERALLOY TYPE NICKEL BASED VALVE ALLOYS

Alloy Fe Mn Ni Cr Si Al Ti C Other

Inconel M 7 2.25 c. 72 16 0.1 0.05 3.0 0.03 –

Inconel 751 7.5 0.1 c. 70 16 0.1 1.2 2.3 0.05 1 Nb
Nimonic 80A <3.0 <1.0 c. 73 20 <1.0 1.4 2.3 0.1 < 2 Co

M. Inconel M, which relied on a fairly simple nickel-titanium intermetallic, has since

been superseded by Inconel 751, in which the intermetallic is one containing nickel,
titanium, aluminium and niobium (Table 4). The market for Nimonic 80A as a valve
material only started well into the post-war era when it began to be used in high-
powered sports cars, but is now being superseded by Nimonic 90.
Neither Inconel M nor KE965 had adequate resistance to lead in the highly boost-
ed aircraft engines used in WW2. Both alloys were basically designed for high tem-
perature strength. Any resistance to lead attack was fortuitous. Coatings were needed,
with the British using Brightray, a high-nickel alloy, and the Americans using cobalt
compositions of the Stellite type. Figure 8 shows how these two types of coating
maintained corrosion resistance to higher temperatures compared to KE965.
The differences behind the British and American choices of coating probably
stemmed from how engine technology had developed in the two countries. As noted

figure 8 TEL corrosion resistance of KE 965, Stellite and Brightray versus temperature
(redrawn by Author from Cowell)31.
18 FRED STARR

earlier, in Britain earlier engines were designed to give its fighters the highest possible
performance, as typified by the in-line, liquid-cooled Merlin and Griffin. In America,
the most powerful engines were air-cooled radials which were developments of
engines built for the passenger transport market.

Post-war automotive developments

Purpose-designed valve alloys

There is no comprehensive account of the history of exhaust valves in the automotive
sector, but since it is still possible to buy replacement valves for cars no longer in
production, one can make an intelligent assessment of how the situation developed.
In post-war Britain, still recovering from WWII and largely reliant on imported
gasoline, octane ratings were still depressed. TEL levels were relatively low, as were
engine outputs. Many manufacturers were producing engines which were basically
obsolescent. Ford were still producing variants of the E93 side-valve, and the engine
in Morris Minor, although an overhead valve design, was given a long stroke-to-bore
ratio, as though the RAC taxation rules were still in force.32
Not too much was demanded from exhaust valves especially as at that time drivers
were willing to do quite major servicing and repair jobs. The Austin/Morris conglom-
erate appears to have opted for an improved version of Silchrome, Silchrome XB,
containing 20% chromium. The alloy had appeared in the late 1930s in the USA. To
ensure that this alloy was heat treatable to form martensite, it contained 2% nickel
and the carbon content was increased from the 0.45% level of Silchrome to 0.8%.
Valve progress at Ford’s can only be described as peculiar, since in a sense the
choice of valve alloy became less sophisticated as time went on. In the original
pre-war E93, Ford appears to have been using KE 965. This seems an odd choice for
a side valve, but exhaust valves in Ford side-valve engines could only be reground
during engine overhaul, so a resistant austenitic alloy was needed. In the post-war
version of the side valve, the exhaust valves were of the composite type, the head
being an austenitic and the upper part of the stem being martensitic. When Ford
introduced the Anglia 105E engine in 1959, the design was completely new, being
of the overhead valve type. For this Ford appears to have specified the lower cost
Silchrome XB alloy. The only conclusion that can be made is that as the engines
developed, although the TEL content of fuels increased, and engine outputs got
higher, better valve design and improved cooling permitted the use of lower-grade
alloys (Figure 9).
In 1955, the Automobile Division of the I.Mech.E. organized a meeting specifically
on exhaust valves. And three years later the Institution held a symposium on ‘The
Protection of Motor Vehicles from Corrosion’, in which one session dealt with the
problems of exhaust valves. Some concern was expressed at both meetings that
the higher octane fuels that were about to come to the UK could cause problems.33
Higher leaded gasolines were being used in the USA and in consequence a new
exhaust-valve material, 21-4N, was used.
This material contained 21% chromium, 9% manganese, 4% nickel, 0.5% carbon
and most importantly, 0.45% nitrogen, differing from the older austenitic valve alloys
in a number of ways. It seems likely that the main incentive for its introduction was
 EXHAUST VALVES FOR INTERNAL COMBUSTION ENGINES 19

figure 9 UK-built Ford car models and exhaust valve materials. Author.

the need for an alloy with a reduced nickel content. Supplies of nickel had been
rationed in the Second World War. The Korean War had produced a new round of
shortages. Eventually, an alloy with even less nickel was developed, 21-2N.
The reduced level of nickel had to be balanced by incorporating a significant
amount of manganese into the alloy, so as to keep the alloy austenitic (Table 5). The
addition of nitrogen also helped to stabilize austenite. However, as with Nimonic 80A
this was another alloy which made use of precipitation hardening. After a twelve-
hour heat treatment at 750°C, very fine chromium carbo-nitrides are formed, greatly
increasing high-temperature strength. Probably because of the high chromium and
low silicon contents, the resistance to oxidation and attack by TEL was quite good.
In this respect Cowley et al. at another I.Mech.E. meeting in 1965 only briefly men-
tion KE965.34 The superiority of the 21-4N and 21-2N class of alloys in terms of
strength and cost is such that the older austenitic alloys have ceased to be marketed.
Cowley worked for Octel, who had taken over the manufacture of TEL. Although
almost fifty years have passed since Cowley and his colleagues set out their work,

TABLE 5
LOW NICKEL/MANGANESE AUSTENITICS FOR EXHAUST VALVES

Alloy Fe Mn Ni Cr Si Mo P N C

21-2N Bal 7.5 1.5 20 0.2 0.4 0.03 0.35 0.55

21-4N Bal 9 4 21 <0.25 – 0.03 0.45 0.55
20 FRED STARR

their paper is, in the opinion of the writer, the best on the corrosion of valve materi-
als. Obviously a paper coming from Octel would focus on lead attack. However, the
paper also covers the effects of sulphur dioxide as this is a constituent of exhaust
gases. Not surprisingly, because of the propensity of nickel-based alloys to form
liquid nickel sulphides, the high nickel alloy, Nimonic 80A, and a valve facing alloy,
EPD 13, did very badly. The main shortcoming of the paper is that the work was
done just before the era of electron probe microanalysis. These techniques are used
to determine which elements in an alloy are preferentially attacked by a corrodant.
The Cowley paper also mentions the importance of good design, but the most
comprehensive paper on this subject is by W. S. Giles, of TRW, which is also from
the mid-sixties.35 An important section in this paper is on Valve Head Shape. Strain
gauging of the valve head showed that, as might be expected, thin section tulip valves
were more able to conform to distorted seats. Unfortunately the tulip shape tended
to raise the stresses and a deep cupped area also raised the temperatures of both stem
and crown. Much of this work was done using finite difference stress analysis and
showed that at least some people in the 1930s had a proper appreciation of a complex
situation.

High output sports- and racing-car engines

Compared to even the best aircraft engines of WWII, the specific output of car
engines is almost twice as great. For example the new Mini Cooper S obtains 163 bhp
from a 1.6 litre engine. There are big differences. Cylinder size in automotive engines
is much lower than in aircraft, which allows high engine speeds. The ultimate is with
F1 racing car engines. These are Vee-8 engines of 2.4 litres capacity where a limit of
18000 rpm results in the engines being constrained to about 300 bhp/litre.
High specific power results in valves having to work hard. One solution, even in
motor vehicles, is the sodium-cooled valve. Mercedes used this in 1969 with their
Vee-8 280SE 3.5 litre model, which corresponds to a power output of 57 bhp/litre,
but not many other manufactures have followed this route. In production cars the
more recent trend is to have four small valves per cylinder and to use higher strength
alloys such as the Inconels and Nimonics. The smaller valve gives a shorter heat path,
but even the Mini Cooper S has recourse to sodium cooling.
Smaller valves can result in reduced overall weight for the valve assembly which
somewhat surprisingly has a beneficial effect on engine output and efficiency,
especially where the actuation of the valves is through push rods and rockers. The
acceleration and deceleration of the valve and other components between the fully-
open and fully-closed positions is equivalent to well over 1000 g. The energy required
amounts to several brake horse power, even in a normal car engine. The high rates
of acceleration and strong valve springs produce high stresses. J. A. Lumley seems to
have push-rod designs in mind when he wrote that:
In modern engines where the stroke-to-bore ratio is below unity, the inertial loading on
the valve train is probably greater than on the connecting rod [. . .] and will limit the top
36
speed of the engine.

Where the valve is actuated directly with an overhead camshaft, without the interven-
tion of rockers, energy is saved and such designs are used on racing cars. However,
 EXHAUST VALVES FOR INTERNAL COMBUSTION ENGINES 21

in an effort to save weight, alloys based on titanium are being used for inlet valves.
The weight of such valves is just over half that of a conventional steel valve. Sadly,
the strength of titanium alloys deteriorates rapidly above 500°C, which would rule
out such materials for exhaust valves, except where very short-lived ultra-high
performance is desired. Efforts have been made to develop titanium aluminide and
SiAlon ceramic for exhaust valves, but success has been elusive.

Low lead fuels and valve seat recession

The move to low or zero lead fuels, following the introduction of the catalytic
converter in the 1970s and 1980s might have suggested that valve troubles due to
corrosion were a thing of the past. In fact, a new round of problems began, not with
the valves themselves, but with the seats which, in many engines, were cut directly
into cast-iron cylinder heads. The phenomenon was termed valve seat recession since
the seat gradually wore back until a point would occur when the valve would not
close properly.
Not all cars were beset with valve seat recession. Giles and Updike stated that:
The average driver, who seldom exceeds 70 mph (c.115 kph), should not experience
significant engine deterioration when using lead-free gasoline. The salesman, however,
who drives 15000 turnpike miles a year (c.25000 km) at 80 mph (c.130 kph) may well
37
experience valve-train problems.

Laboratory testing showed that, if engines were run on lead free fuel, there could be
very high wear rates, although the scatter was considerable. With leaded fuel the
maximum wear rates that were measured were around 100 µm after an eighty-
hour moderate speed test. With unleaded gasoline the rates could be over ten times
faster.
At moderate speeds, typical of normal use, an initiation period was needed before
serious wear began to take place, but if an engine was driven hard, right from the
start, severe wear would commence immediately.38 Inspection of valve faces and seats
from affected engines showed that the seats were often covered with a reddish
coloured material, which proved to be haematite, Fe2O3. The corresponding faces of
the valves were covered by what were described as warts. The warts were found to
consist of lenticular masses of metal and oxides which originated from the seats and
which had become embedded in the valve faces (Figure 10).
The suggestion was that in the absence of lead compounds, which would otherwise
have acted as a solid lubricant, an adhesive wear process began to take place. Some
cast iron was thereby transferred to the valve faces where, because the temperature
was so much higher, it began to oxidize. The rotation of the valve led to a balling up
of oxide and particles of cast iron, until the point was reached when the balls were
forced into the valve face surfaces. The valve face now began to act as a grindstone,
wearing the cast-iron valve seat away. The particles that resulted from the abrasive
wear process would begin to oxidize, helping accelerate the wear process.
Once the problem was recognized the solution was relatively simple. The cast iron
seats could be induction hardened.39 The alternative was to use specially hardened
seats of a nickel-based cast iron. A typical composition would be Ni-29Cr-25Fe-
8.5Mo-2.2C-1.5Si.40
22 FRED STARR

figure 10 Schematic of the mechanism of valve seat recession (Author).

Conclusions
The original concept in formulating this paper was to show that there was almost a
direct relationship between the introduction of different types of fuel and the types
of materials used in the manufacture of exhaust valves. But because of the constant
effort to get more power out of the internal combustion engine it became clear that
 EXHAUST VALVES FOR INTERNAL COMBUSTION ENGINES 23

the story was more complex. In consequence there were three main threads to the
story, not two, the third being that of design.
Summarizing, we can state the following:
Fuels
It took over twenty years for people like Ricardo to recognize that some fuels were better
than others. Simple technical fixes, like adding benzole and tetraethyl lead, were soon
introduced, but what was really needed was the modification of fuels by oil refiners to
increase the concentration of isooctane in gasoline.

Materials
The development of higher output engines after 1900 forced designers to try out alterna-
tives to cast and carbon steels. Through the efforts of organizations like the Royal Air-
craft Factory/Establishment and the valve manufacturers, better materials were adopted,
but these were simple modifications of alloys which were in use. It was only after the
Second World War that materials were specifically developed for exhaust valves. Even
this was probably more a result of the shortage of nickel. The other materials just
happened to be in the right place at the right time, the classic example being Nichrome,
originally developed for electric toasters. Even today, the factors which cause valves and
seats to fail are still not properly understood at a fundamental level.

Engine and valve design

Almost from the start efforts were made by designers to prevent exhaust valves running
too hot, even though side-valve combustion chambers compromised power output. The
switch to overhead valves meant that the valves had to withstand the full impact of hot
corrosive exhaust gases. To mitigate these effects, there was an effort to reduce valve
stresses and temperatures by changing the shape of valves and by introducing sodium
cooled valves.
As designers took advantage of the new fuels and new insights into engine design,
poppet valve technology became a limiting factor. Eventually, in seeking ever-higher
power from aircraft engines, Bristol and Napier switched to sleeve valve designs, and if
the gas turbine had not come along Rolls-Royce would have been building the Eagle,
another sleeve valve engine. In the automotive sector the sleeve valve is a lost cause.
Engineering progress has given us poppet valves of such reliability that the cylinder head
never comes off, even in vehicles that that have done more than 200,000 kilometres before
they make that last journey to the scrapyard.

Finally, another theme also emerges from this survey, which relates to the difficulty
in ascertaining exactly how small but important technological innovations came
about. This is especially problematic in the case of alloy and engine development.
Much of the work was confidential and even within the organizations responsible it
was only circulated on a need-to-know basis. If the work was published it was then
done in a highly redacted form. Eventually the internal reports relating to the work
would be destroyed. One then has to try to supplement the historical record with
personal reminiscences of those who had some involvement. These are often more
accurate than what has been formally published, but much of this was and is written
to put various organizations (and Nations) in the best possible light. It is here that
organizations like the Newcomen Society can encourage their members to help put
the record straight.
24 FRED STARR
Acknowledgements
To the unknown reviewer of Part 1 who asked pertinent questions about valve
stresses. To Ed Marshal and Bryan Lawton who did an initial informal review of
what was a very long paper. And to the Editor for agreeing that this monograph on
exhaust valve development be published in several parts in the Newcomen Journal.
Notes
1 F. Starr, ‘Development of the Poppet Type
Exhaust Valve in the Internal Combustion
Engine: Part 1’, Journal of the History of
Engineering and Technology, 82 (2012),
283–314.
2 H. R. Ricardo, Memories and Machines
(Constable, 1968), pp. 126–28, 186–87 and
205–08.
3 S. D. Heron, ‘Aircraft-Engine Practice as
Applied to Air-Cooled Passenger Car
Engines’, J. Soc. Auto. Eng., 12 (1923),
31–56.
4 F. Starr, ‘The Key to Record Breaking-
Exhaust Valve Cooling’, in F. Starr, E.
Marshall and B. Lawton, eds, The Piston
Engine Revolution (Newcomen Society,
2011), pp. 509–37.
5 C. Kettering, ‘A Tribute to Thomas Midgley
Jnr’, Ind. Eng. Chem., 36 (1944), 1179–82;
M. Baker, ‘Thomas Midgley and the Law
of Unintended Consequences’, Invention and
Technology Magazine, 17 (Spring 2002),
38–45.
6 A. L. Marsh, ‘Electric Resistance Element’,
US Patent 811859 (February 1906).
7 N. M. Clark, ‘The World’s Most Tragic
Man is the One Who Never Starts’ (an article
based on an interview with William Hoskins),
The American Magazine (May 1927).
8 C. Sponaugle, ‘History of Haynes Interna-
tional’, Pittsburgh Engineer (Winter 2005),
pp. 7–9.
9 The Discovery of Stainless Steel (British
Stainless Steel Association, c. 2005).
10 D. Dyer, TRW-Pioneering Technology and
Innovation since 1900 (Harvard Business
School Press, 1998), pp. 40–42.
11 A. L. Boegeld and J. B. Johnson, ‘Engineer-
ing Requirements in the Automotive
Industry for Metals Operating at High
Temperatures’, Symposium on the Effect
of Temperature on Metals (ASTM, 1932),
pp. 169–200.
12 J. R. Handforth, ‘Metallurgical Problems
Arising from Internal Combustion Engine
Valves’, Journal of the Iron and Steel Inst.,
126 (1932), 93–157.
13 E. W. Constant II, The Origins of the Tur-
bojet Revolution (John Hopkins University
Press, 1980), Chapter 6.
14 A. T. Cowell, ‘Modern Aircraft Valves’, SAE
Transactions, 35 (1940), 147–65.
15 J. C. Sanders, J. C. Wilsted and B. A.
Mulcahy, ‘Operating Temperatures of a
Sodium-Cooled Exhaust Valve as Measured
by a Thermocouple’, NACE Report, 754
(1943).
16 P. B. Henshaw, ‘Valve Steels’, J. Roy. Aero.
Soc., 31 (1926), 187–217.
17 Cowell.
18 J. H. S. Dickenson, ‘Experiments on the
Flow of Metals at Red Heat, With a Note on
the Scaling of Heated Steels’, J. Iron and
Steel Inst., cvi.2 (1922), 103–54.
19 Handforth.
20 Ibid.
21 F. Starr, ‘Exhaust Valve Corrosion in Inter-
nal Combustion Engines: A Brief History
from the Perspective of Modern Thinking’,
EuroCorr 2000 Conference (Queen Mary
College London, 2000). Copies can be
obtained directly from the author.
22 F. R. Banks ‘Ethyl: Some Information on
the Use and Advantages Gained by the
Employment of Tetraethyl Lead in Fuels for
Aviation Engines’, J. Royal Aeronautical
Soc., 38 (1934), 309–72.
23 Kessler, ‘Significance of Octane Numbers for
Aviation’, in R. J. Forbes and D. R. O’Beirne,
eds, The Technical Development of the
Royal Dutch/Shell 1890–1940 (Leiden: E. J.
Brill, 1957), pp. 392–402.
24 Ibid.
25 ‘Eugene Houdry’ and ‘Herman Pines’,
History of Catalysis (The North American
Catalysis Society, <nacatsoc@gmail.com>)
(website regularly updated).
 EXHAUST VALVES FOR INTERNAL COMBUSTION ENGINES 25

26 ‘Manufacture of Components of High 5 (1955/56), 147–69; Downs, ‘Exhaust Valve

Octane Gasoline’, The Petroleum Handbook
(Shell Petroleum Company, 1948), Chapter
XVI, pp. 220–37; G. D. Hobson and W.
Pohl, eds, Modern Petroleum Technology
(1973), pp. 278–374 (Chapter 9 ‘Cracking
and Reforming’ and Chapter 10 ‘Alkylation;
Isomerization; Polymerization; Hydrodesul-
phurization; Sulphur Production’).
27 N. P. Allen, ‘A Survey of the Development
of Creep Resistant Alloys’, Symposium of
High Temperature Steels and Alloys for Gas
Turbines (Iron and Steel Inst., July 1952),
pp. 1–17.
28 L. B. Pfeil, N. P. Allen and C. G. Conway,
‘Nickel-Chromium-Titanium Alloys of the
Nimonic 80 Type’, Symposium of High
Temperature Steels and Alloys for Gas
Turbines (Iron and Steel Inst., July 1952),
pp. 37–45.
29 J. O. Hitchcock, Letters to Prof. J. Harris
copied to F. Starr, 29 and 30 March 1999.
30 J. E. Morgan, ‘Aero Engine Exhaust Valve
Development’, Proc. of Auto Division, Inst.
Mech. Engineers, 5 (1955/56), 138–46.
31 Cowell.
32 For a discussion of the effect of car taxation
on engine development, see A.E. Clausager,
‘The Development of the Piston Engine for
Motor Cars’ in F. Starr, E.L. Marshall and
B. Lawton (Eds.) The Piston Engine Revolu-
tion (Newcomen Society, 2012), pp. 66-69.
33 R. D. Mogford and F. A. Ball, ‘Exhaust
Valve Life: Some Metallurgical and Mechan-
ical Considerations’, Proc. Inst. Mech. Eng.,
Life in Small Gasoline Engines’, Proc. Inst.
Mech. Eng., 5 (1955/56), 124–33; R. W.
Long, ‘Performance of Vehicle Engines on
Heavily Leaded Fuels’, Proc. Inst. Mech.
Eng., 5 (1955/56), 133–46; W. G. Stevenson,
‘Corrosion of Cylinders and Moving Parts of
Engines’, The Protection of Motor Vehicles
from Corrosion, Monograph No. 4 (Society
of Chemical Industry, 1958), pp. 3–11.
34 W. E. Cowley, P. J. Robinson and J. Flack,
‘Internal Combustion Engine Poppet Valves:
A Study of Mechanical and Metallurgical
Requirements’, Proc. Inst. Mech. Eng., 179,
Pt 2A, No. 5 (1964–65), pp. 45–182.
35 W. S. Giles, ‘Fundamentals of Valve Design
and Materials Selection’, SAE Paper 660471
(1964).
36 J. A. Lumley, Engines an Introduction
(Cambridge University Press, 1999), p. 197.
37 W. S. Giles and S. H. Updike, ‘Influence
of Low Lead Fuels on Exhaust Valve
Performance’, SAE Paper 710674 (1971).
38 G. A. Schoonveld, R. K. Riley, S. P. Thomas
and S. Schiff, ‘Exhaust Valve Recession with
Low Lead Gasolines’, SAE Paper 861550
(1986).
39 W. S. Giles, ‘Induction Hardening makes
Exhaust Valve Seats Wear Less with Non-
Leaded Fuel’, Automotive Engineering (June
1971), pp. 33–37.
40 S. L. Narasimhan and J. M. Larson, ‘Wear
Characterisation of New Nickel Based
Alloys for Internal Combustion Engine Valve
Seat Applications’, Wear, 74 (1981–82),
213–27.

Notes on contributor
F. Starr graduated in Metallurgy in 1966 and joined British Gas, becoming their
principal investigator of steam reforming plants failures. From 1974 he ran a group
to develop high temperature alloys for advanced coal and oil gasification processes.
Some of this work is covered in his semi-historical survey ‘The Engineering Science
of High Temperature Corrosion’, a lecture given at the University of Surrey. From
the 1980s onwards he initiated programmes dealing with Stirling Engines, Closed
Cycle and Inverted Cycle Gas Turbines, and the Rotary Vee two-stroke. After leaving
British Gas, in 1996, his main work has been on power plant operation. His last
formal job was with the European Commission’s Institute for Energy, where as a
‘Visiting Scientist’ he was responsible for the design of a plant to produce hydrogen
and electricity from coal. In terms of industrial history, his focus is on steam, IC and
gas turbine development, using modern knowledge to highlight the background to
engine designs and breakthroughs. He achieved a life-long ambition in 2007 in gaining
26 FRED STARR

a doctorate with his thesis on ‘The Development of an Expert System for Failure
Analysis of Power Plant Components’.
Correspondence to: Fred Starr. Email: newcomen.starr@yahoo.com

Discussion
Dan Hayton: What is the difference between internal combustion engines and jet
engines in terms of the temperatures and what the alloys are trying to deal with?
Author: There is definitely a connection between the development of valve materials
and turbine blade alloys. Because of preignition the aim was to produce valve materi-
als which would have the highest possible strength at a relatively low temperature,
say something like 600 to 700°C. That is why it took such a long time for the nickel-
based alloys to be used, since they come into their own in the 750 to 900°C range. In
a jet engine, because of the need to minimize the cooling requirements, you need
turbine blade material capable of running at 750° and ideally 800–850°C.
Sir Frank Whittle knew that his father (who had run a small jobbing machine shop)
had been involved with people building racing and sports cars, and had made valves
using the best high temperature alloys. So Sir Frank, when being told that a jet engine
would need ‘impossible’ materials, could say ‘my father is using these materials’. So
there is a very nice, although indirect, engineering connection between valves and
turbine alloys.
Member: Did they experiment at all with exhaust valves which are larger than the
inlet valves, the point being that it would have more thermal capacity?
Author: The bigger the valve, the more heat is flowing into the valve, and the more
difficult it becomes to disperse the heat. In large spark-ignition stationary engines (not
diesel engines), valve temperature limits engine output.
Edward Field: You handed some valves around with a magnet. In some cases the
valve head is non magnetic, but the stems are magnetic Are these made of different
materials?
Author: The upper part of the stem material is cheaper, and is magnetic because it
has a martensitic structure. In this region of the valve the main requirement is for
high hardness to resist wear as the valve moves up and down in the guide. The
temperature resistant, austenitic alloy, which is non-magnetic, is used in the lower
stem and crown.
Neil Barton: I have three questions. You said that the Ford Anglia was built in 1955
but one of the valve samples you handed round was for a 1959 Ford Anglia. Also at
the beginning of your talk you showed a big American valve from a Wright 3350
radial, and a British valve from the Merlin, and you joked that ‘there were no prizes
for guessing which was which’. Is there a cultural difference, like that with Mobile
Phones, where Motorola tended to produce large clunky phones, and Europe
designed small things? The third question how did German aircraft engines differ
from those of the British?
Author: Ford’s produced various cars, all using the same type of side-valve engine,
with different bores and strokes, and each of the model names Popular, Prefect, and
Anglia were used for a progression of cars, which makes it confusing. However the
engine in the 1959 Anglia was an entirely new design, and used overhead valves.
 EXHAUST VALVES FOR INTERNAL COMBUSTION ENGINES 27

The Allies had developed methods of producing intrinsically better fuels of a high
octane rating using sophisticated techniques for modifying the chemical structure of
poor quality petroleum feedstocks. TEL was added to improve octane rating still
further. The British were able to use these very high quality fuels, so Allied engines
ran at a much higher power output. [At this point a member pointed out that the
Germans made much of their fuel from coal, which if done by hydrogenation route
should give fuel of quite a high octane rating.] However, even so, the German fuels
were not that good (c. 96 octane), and got high power output by nitrous oxide and
oxygen injection.
[Since the meeting, the author has obtained a wartime book Metallurgical Study of
German and Italian Aircraft Engines and Airframe Parts published in 1943. The
exhaust valve alloys were similar to those used by the British, and the designs were
also similar. The exhaust valve from the BMW 132K single row radial, which was
apparently based on the Pratt and Whitney Hornet engine, was of the sodium cooled,
hollow head type, with Stellite on the seats.]
Turning now to the cultural question, Dr Barton is quite right, there is a ‘cultural
difference’, but not in the way one might think. In America, by 1925, they had
realized that the aeroplane was here to stay. Instead of doing a three-day train trip
between New York and San Francisco, it could be done in twelve hours, provided air
travel was reliable. The US Government and various charitable organizations poured
money into aircraft design and engine development. What they did not want, if you
needed a reliable commercial service, were engines which need a lot of maintenance
and valves needing regrinding every two or three days. So they went for the air-cooled
radial engine, eliminating the need for water cooling. For valves, the low maintenance
choice was two valves per cylinder, one inlet and one exhaust, initially relying on salt
or sodium cooling to keep down valve temperatures.
Bryan Price: There was something which distinguished American from European
designs. In the 1920s people who were designing aircraft engines were also designing
for automotive use. Around 1923, the RAC introduced the taxation calculation that
biased engines towards long stroke, small bore designs, which dominated UK auto-
motive engine design for over forty years. Because of this there was a drive towards
high-specific-output, liquid-cooled designs and this would push the use of four valves
per cylinder for better breathing.
Something you have not mentioned is the mass of the valve and valve train dynam-
ics, which was a big issue in the UK and Europe. But in the US, there was much more
emphasis on low cost manufacturing, which was one reason why the Americans stuck
with air cooling for so long and two valves. In addition they had cheaper fuels and
were less concerned about having more fuel-efficient engines. In consequence the
Americans went for big bore, large capacity engines of limited output.
Author: My impression is that in terms of bhp per litre, American car engines used
to be very poor. However, US car manufacturers introduced overhead valves much
earlier than UK engine constructors. The reason for this comes back to the manufac-
turing issue, and here my opinion differs. In Britain we had to build cheap engines
and the side valve was ideal. So we will have to differ on this point.
28 FRED STARR

Bryan Price: Okay. But the first four-valve overhead-cam engines were those from
Peugeot in 1914, and engines of this type were being used in Grand Prix and higher
output engines.
On the German question the biggest competitor to the Merlin was the DB 601. This
differed from the Merlin in having fuel injection.
Dave Andrews: Aircraft engines used to have water injection for take-off. What was
the thinking behind this? Was the injection into the cylinders to keep the valves
cool?
Author: I would ask Clive Ellam, who is from the aircraft industry, if he could
comment on the water injection issue.
Clive Ellam: There is a mystery about the effects of water injection, but on Rolls-
Royce engines the water was injected into the eye of the supercharger impeller. It then
vaporized and the two theories were either that the cooling increased the mass of the
charge, or it cooled the mass of the charge so that the volumetric efficiency was
increased. I do not know any one who injected water directly into the cylinders.
John Russell: There is a discussion in technical press about how much Formula One
contributes to engine design. Have you got any evidence, whether what is going on
at Goodwood and elsewhere, that competition was improving our knowledge and
performance of engines and valves?
Bryan Price: With respect to Formula One, the main developments have been made
in the area of materials, where they have been given the opportunity to prove them-
selves. For example, the ‘turbo era’ of racing cars had encouraged the turbocharging
of production cars. Today, there are efforts to develop KERT (Kinetic Energy
Recovery Technology), which will be incorporated into hybrid vehicles.
Author: At present the F1 rules seem to be hindering development. Engine speed is
limited to 18,000 rpm and there is a minimum weight limit. The rules therefore are
inhibiting the use of light-weight/high-strength materials. For example, if we con-
sider a valve being pushed up and down at 18,000 times a minute, plus the rest of the
valve train, one is expending a huge amount of energy, so lightweight valves could
make a real contribution to engine power. Even with normal vehicles, on the motor-
way, where the engine speed is moderate, the power required to drive the valve train
is a significant fraction of the engine output.
Ed Marshall: Fred, you have given us an extremely entertaining evening of just
short of a hundred years of exhaust valves. It seems that it was really a case of the
need to survive (improvements in engine output) which was at the back of valve
development.
I liked your description of Ricardo, Heron and Midgley as representing ‘The Good,
The Bad and the Ugly’. But in 1920, ‘The Good’, Harry Ricardo, said that the exhaust
valve had reached the limit of its development. Although Ricardo was not often
wrong, he was really wrong about that! But I am not sure that Heron was that bad.
[Author: Heron had a bad temper.] The oil companies had thrown high-octane fuels
and tetraethyl lead at the valves, and Heron did a lot to get round that problem.
 EXHAUST VALVES FOR INTERNAL COMBUSTION ENGINES 29

With respect to the German fuel situation, they hydrogenated brown coal, which
gave them very a aromatic fuel, which has a high octane. They also had tetraethyl
lead, so they had a fuel which was of the 90/130 grade, which was fairly good.
There is an interesting story about lead. When TEL was first manufactured, it was
only made in the USA and then shipped to Europe. By the mid thirties, lead became
quite common and was stipulated in aviation fuel specifications. The Air Ministry did
recognize that in the event of hostilities, should America sign a neutrality agreement,
there would be a problem. So after discussion with the Ethyl Corporation, and then
with ICI, they set up a manufacturing base. To keep things on a level playing field,
the Americans did the same in France and Germany.
Whilst the German plant was being constructed, the British found out, purely
by chance, that the Germans were secretly building a second one! It was a case of a
letter being put in a wrong bag. So at the beginning of the Second World War, the
British knew the location of the two German plants. Now I would have thought that
these would have been prime targets for the RAF, but they were never attacked at all.
The plants went right through the War. At the end, the Russians dismantled them,
toilets and all, and shipped them back to Russia.
I have one complaint! I did not hear the words ‘sleeve valve’, which of course ‘Mr
Good’ put his money on in 1920.
I remember that back in the late 1960s, when we started taking lead out of gasoline,
the first engine we tried was a Jaguar. On lead-free fuel, it ate the valves on just one
tank of fuel. It was then a case of ‘Thank God for metallurgists’, since today we are
now all using lead-free fuel.
Author: We should thank inventors, or God, since metallurgists still did not have a
good theory to predict how materials wear.