University of Nebraska - Lincoln

DigitalCommons@University of Nebraska - Lincoln

Robert Katz Publications Research Papers in Physics and Astronomy

1-1958

Physics, Chapter 14: Temperature

Henry Semat
City College of New York

Robert Katz
University of Nebraska-Lincoln, rkatz2@unl.edu

Follow this and additional works at: https://digitalcommons.unl.edu/physicskatz

Part of the Physics Commons

Semat, Henry and Katz, Robert, "Physics, Chapter 14: Temperature" (1958). Robert Katz Publications. 138.
https://digitalcommons.unl.edu/physicskatz/138

This Article is brought to you for free and open access by the Research Papers in Physics and Astronomy at
DigitalCommons@University of Nebraska - Lincoln. It has been accepted for inclusion in Robert Katz Publications by
an authorized administrator of DigitalCommons@University of Nebraska - Lincoln.
Part Two
HEAT

14
Temperature

14-1 Concept of Temperature

Temperature is one of the fundamental concepts of physics. We are all

able to recognize that some bodies are hotter than others, but our tem-
perature sense is qualitative rather than quantitative and is capable of only
a limited range. The sense of touch can frequently be used to distinguish
between hotter and colder objects, provided that these lie in a temperature
range consistent with the stability of human tissue. Even within this
range the sense of touch is often unreliable as a measure of temperature.
The metal bracket holding a wooden rail may feel much colder to the touch
than the railing itself, even though both are at the same temperature.
It is a matter of common observation that some of the physical prop-
erties of many substances are altered when the temperature is changed.
The volume or the pressure of a gas increases when the temperature is
raised. The length of a copper rod changes with changing temperature.
Some of the electric and magnetic properties of substances vary with
changes in temperature. The changes that take place in these physical
properties can be used to measure the changes in temperature which pro-
duced them.
In order to make a measurement of the temperature of a body, it is
important to be able to decide when two bodies are at the same temperature.
Suppose that the length of a copper rod is measured while it is in the
laboratory and exposed to the air in the room; if the rod is then put into a
mixture of ice and water, its length will first decrease and then reach a new
value which will remain constant as long as it is in the ice-water mixture.
We then say that the temperature of the copper rod is the same as that of
the ice-and-water mixture. The two systems, the copper rod and the ice-
water mixture, are in thermal contact and have reached thermal equilibrium.
Thus two systems are said to be at the same temperature when they are in
267
268 TEMPERATURE §14-2

thermal equilibrium. We may thus consider temperature to be a property

of a system which determines whether or not it is in thermal equilibrium
with any other system that is placed in thermal contact with it.

14-2 Thermometers
A physical device which measures temperature is called a thermometer.
In order to construct a thermometer, it is necessary to choose some thermo-
metric property of a system whose value depends upon its temperature.
Although the length of a solid rod
F C is seldom used as this thermometric
212 100 property, the length of a liquid col-
umn in a glass tube is used very fre-
quently. In this liquid-in-glass ther-
mometer, the difference in volume
expansion between the liquid and the
glass container is visible as a change in

32 a

Fig. 14-1 A mercury-in-glass ther- Fig. 14-2 Constant-volume

mometer with the ice point and the gas thermometer.
steam point marked in both the
Fahrenheit and Celsius scales.

length of the liquid column when it is allowed to expand into a very fine
capillary tube attached to the glass bulb containing most of the liquid, as
shown in Figure 14-1.
The constant-volume gas thermometer, in which the pressure of the gas
is the thermometric property, provides an extremely accurate measurement
of temperature. This thermometer, shown in Figure 14-2, utilizes the
§14-3 TEMPERATURE SCALES 269

change in pressure of a gas with temperature, when the gas is maintained at

constant volume. Gases customarily used are air, hydrogen, or helium.
The color and intensity of the light emitted from a furnace vary with
the temperature. An optical pyrometer is a device in which the temperature
of a furnace is measured by comparing the light emitted from the furnace
with the light emitted by an electrically heated filament which has been
previously calibrated. The character of the light emitted by an incan-
descent body is a thermometric property suitable for use in the determina-
tion of temperature.
The electrical resistance of a wire changes with temperature. As we
shall see in a subsequent chapter, the electrical resistance of a metal is an
easily measured property. This thermometric property provides the basis
for the resistance thermometer.
If the ends of two wires made of different metals or alloys are connected
together and their junctions maintained at different temperatures, it is
found that an electric current flows in the wire loop. The current is pro-
duced by an electromotive force whose value depends upon the difference
in temperature of the junctions. This thermometric property provides the
basis of the thermocouple type of thermometer. When a number of small
thermocouples are connected together, the effect is enhanced, and the
result is called a thermopile.

14-3 Temperature Scales

In order to be able to assign a number to the temperature of a body, it is

necessary to agree upon a method for setting up a temperature scale. We
must first choose some thermometric property of a system, for example,
the length of a mercury column in a glass capillary tube, or the electrical
resistance of a platinum wire, or the pressure of a gas kept at a constant
volume. Let us call the value of the chosen thermometric property X and
let T be the temperature of a system that surrounds this thermometer and
is in thermal contact with it. When thermal equilibrium is reached, the
thermometer and the surrounding system are at the same temperature. If
the temperature of the system is changed, the value of the thermometric
property of the thermometer is also changed. The temperature, as indi-
cated by the thermometer, will be some function of X; let us call this
function T(X). Let us assume that T(X) is a linear function of X, thus

T(X) = aX,

where a is a constant of proportionality.

Thus, if at one temperature the value of the thermometric property is
Xl, and if at a higher temperature the value of this thermometric property
270 TEMPERA'l'URE §14-3

is X 2 , then we can write

T(X I ) Xl
--=-. (14-1 )
T(X 2 ) X2
Before 1954 two fixed points were used to assign numbers in calibrating
a thermometer. Since 1954, by international agreement, only one fixed
point is being used. Since the older scales are still in common use, we shall
describe the earlier method first.
A fixed point on a temperature scale is a number that is assigned to the
temperature of an easily reproducible state of a system. Before 1954 the
two fixed points were (a) the temperature of a system consisting of a mixture
of ice in equilibrium with water open to the air at standard atmospheric
pressure with the water saturated with air (called the ice point) and (b) the
temperature of steam in equilibrium with pure water at standard atmos-
pheric pressure (called the steam point).
To measure the temperature of a system, the thermometer is put in
thermal contact with it; the value X of the thermometric property at this
temperature T(X) is measured. Calling Xi the value of this property at
the ice point, we have
T(X i ) Xi
--=-. (14-2)
T(X) X
Similarly, if X. is the value of the thermometric property at the steam
point, then
T(X.) X.
---=-. (14-3)
T(X) X
From the above equations we get
T(X.) - T(X i )
(14-4)
T(X) X

so that T(X)
T(X.) - T(Xi ) x. (14-5)
X. - Xi
On the Celsius scale of temperature, also called the centigrade
scale of temperature, first devised by A. Celsius (1701-1744), the tempera-
ture interval T(X.)-· T(X i ) is assigned the value of WO°C, the ice point
is called O°C, and the steam point WO°C, the interval being divided into
100 equal divisions called degrees Celsius.
On the Fahrenheit scale, first devised by G. Fahrenheit (1686-1736),
the interval between the ice point and the steam point is assigned the
value 180°F, the ice point is called 32°F, and the steam point 212°F, the
interval being divided into 180 equal divisions called degrees Fahrenheit.
§14-3 TEMPERATURE SCALES 271

As the accuracy of temperature memmrements increa:,;ed, discrepan-

cies arose principally because of the difficulty of reproducing the ice point.
It will be recalled that the ice point is the temperature of an equilibrium
mixture of ice and air-saturated water at atmospheric pressure. The dis-
crepancies between measurements made at standardizing laboratories
amounted to as much as 0.04°C, whereas the accuracy of measurement in
this temperature region was good to about O.OOl°C. In 1954 it was de-
cided to change the method of calibrating thermometers and to use only
one fixed point; this fixed point is the temperature of an equilibrium mixture
of ice, water, and water vapor; this state is called the triple point of water
and exists at only one definite temperature and pressure (see Chapter 17).
The number chosen for this fixed point is 273.16 degrees Kelvin, written as
273.16°K. (This is in honor of Lord Kelvin who developed the thermo-
dynamic scale of temperature. See Chapter 19.) Thus if X a is the value
of the thermometric property of the thermometer at the triple point, then
from Equation (14-1)
T(X) X
--=-, (14-6)
T(Xa) Xa
and since T(X a) = 273.16°K, (14-7)

X
we have T(X) = 273.16°K - . (14-8)
Xa
The triple-point cell used by the K ational Bureau of Standards is
illustrated in Figure 14-3. Very pure, air-free water is introduced into the
cell which is then sealed. The cell is then cooled in a thermos flask by an
ice bath until ice, water, and water vapor are present simultaneously in the
cell, indicating that the triple point of water has been attained.
If the constant-volume gas thermometer is used for measuring tem-
perature, the thermometric property that is measured is the pressure P
of the gas. For this case Equation (14-8) becomes
° -P ,
T(P) = 273. 16K (14-9)
Pa
where P a is the pressure of the gas at 273.16°K or the triple-point tempera-
ture. As the result of many careful experiments, we find that the value of
the temperature of a particular system depends upon the nature of the gas
that is used in the gas thermometer. However, as smaller and smaller
amounts of gas are used in the thermometers, so that the pressure of the
gas gets smaller and the density gets smaller, all gas thermometers give the
same result for the temperature T of the system. Thus

T = 273.16 °K
(denRity-.O)
(P)
-
P"
. (14-10)
272 TEMPERATURE §14-3

We shall consider some of the interesting properties of gases in this and

later chapters.
The relationship between the Celsius scale and the Kelvin scale can
now be stated in terms of the single fixed point. On the Celsius scale the
triple point is defined as 0.01 °C; any temperature t on the Celsius scale can
be defined in terms of the tempera-
ture T on the Kelvin scale as

t = T - 273.15. (14-11)

The relationship between the

temperature of a body expressed on
the Celsius scale tc and its tempera-
ture expressed on the Fahrenheit scale

Fig. 14-3 Diagram of the Na- Fig. 14-4 Graph of Fahrenheit

tional Bureau of Standards temperature versus Celsius temper-
triple-point cell, ED, in use in an ature.
ice bath G, within a thermos
flask H. A, water vapor; C,
thermometer well; E, ice mantle;
F, liquid water.

tf can be found by plotting the Fahrenheit temperature as ordinate and

the Celsius temperature as abscissa on rectangular coordinate paper, as
shown in Figure 14-4. The curve relating these two temperatures is a
straight line of slope tg g, which intercepts the Fahrenheit axis at 32°F.
Applying the slope-intercept form of the equation of a straight line
y = m;r + b,
§14-4 THERMAL EXPANSION OF SOLIDS 273

where m is the slope and b is the y intercept, we find

tf = -tt c + 32. (14-12)

This equation may be solved for tc to yield

t c = ~(tf - 32). (14-13)

Illustrative Example. Find the centigrade temperature at which a centi-

grade thermometer will read the same number as a Fahrenheit thermometer.
The relationship between the reading of a centigrade thermometer and a
Fahrenheit thermometer is always given by Equation (14-12) as
tf = -tt c + 32.
The additional condition imposed by the problem may be expressed analytically
as
tf = t e •

Solving these two equations simultaneously, we find

tc = -40.

14-4 Thermal Expansion of Solids

A change in the temperature of a substance is nearly always accompanied

by a change in its physical dimensions. The expansion which takes place
in a perfect crystal when its temperature is increased depends upon the
direction in which the expansion is measured relative to the crystal axis.
Most crystalline solids are made up of grains within which the crystals are
oriented in one direction. The crystal directions are randomly oriented
from grain to grain, so that we find that the expansion of most solids is
often the same in every direction relative to the crystal axis.
The change in length of a solid rod is a smooth function of temperature
and may be represented by an infinite series of the form
Lt = L o[l + (X (~t) + (x' (~t)2 + ... ], (14-14)

where L t is the length of the rod at temperature t, L o is the length of the

rod at some reference temperature to, and ~t is the temperature difference
t - to. The coefficients (x, (x', and so on, are constants which must be
evaluated by experiment for each different material. It is common practice
to approximate the expansion of a solid by making use of a number of
straight lines tangent to the curve which describes the length of a rod as a
function of temperature, as shown in Figure 14-5. In this case only the
first two terms of the right-hand side of Equation (14-14) are required to
represent anyone of these tangent lines, and the equation becomes
(14-15)
274 TEMPERATURE §14-4

where a is called the coefficient of linear expansion at the temperature to.

Since the slope of each of the tangent lines in Figure 14-5 depends upon the
reference temperature at which it is constructed, the coefficient of linear
expansion depends upon the reference temperature at which it is evaluated.

to t
Fig. 14-5

The slope of the tangent line at temperature to, at which the length of the
rod is L o, is given by Loa. If we write the change in length of the rod as
b.L = L t - L o,
we may express the coefficient of linear expansion as
b.L/Lo
a=---' (14-16)
b.t
In other words, the coefficient of linear expansion is the fractional change
in length of the rod per degree of temperature change. The units of a are
therefore (degrees)-l. Since a Fahrenheit degree represents a smaller
temperature interval than a centigrade degree, the coefficient of linear
expansion per degree Fahrenheit is five ninths of the coefficient of linear
expansion per degree centigrade. The linear-expansion coefficients of
several solids are given in Table 14-l.
If two metals, say brass and steel, are welded or riveted together so
that they form a single straight piece at room temperature, then, when the
temperature is raised, the strip will bend in the form of an arc with the brass
on the outside, as shown in Figure 14-6. This is due to the fact that brass
has a greater coefficient of expansion than steel. A bimetallic strip of this
kind is frequently used in thermostats. The strip is fixed at one end, and
the bending of the free end may be used to actuate a switch at some pre-
determined temperature.
A solid in which the expansion is the same in every direction is said to
he isotropic with regard to this property. When the temperature of an
§14-4 THERMAL EXPANSION OF SOLIDS 275
TABLE 14-1 COEFFICIENTS OF LINEAR EXPANSION AND VOLUME EXPANSION

Substance a
Aluminum (20-300°C) 25.7 X 10- 6 per °C
Brass (0-100°C) 19.3
Copper (25-300°C) 17.8
Pyrex glass (21-470°C) 3.6
(550-570°C) 15.1
Invar (20°C) 0.9
Lead (18-100°C) 29.40
Platinum (40°C) 8.99
Steel (40°C) 13.2
Tungsten (0-100°C) 4.3
(1000-2000°C) 6.1
Substance {3
Mercury (0-100°C) 181.8 pcr °C

isotropic solid is changed, the length of each linear dimension is changed.

The area of each element of area is changed, and the volume of each volume
element is changed. We may represent the expansion of area and the
volume expansion by mathematical series analo-
gous to Equation (14-14) and may approximate
the expansion by use of the first two terms of
the series, as before. To represent the area ex-
pansion, we write Bross Steel
Bross
At = A o[l + (T (~t)], (14-17)

where (T (sigma) is the coefficient of area expan-

sion at the temperature to. Similarly, to represent Fig. 14-6 Bending of a
the volume expansion we write bimetallic strip when
heated.
(14-18)

where {3 (beta) is the coefficient of volume expansion at the temperature to.

The coefficient of area expansion and the coefficient of volume expansion
may be expressed by equations analogous to Equation (14-16) as
~A/Ao
(T = ---, (14-19)
~t

~V/Vo
and {3=--' (14-20)
~t

We may relate the coefficient of linear expansion to the coefficient of

area expansion by considering the expansion of a square plate of dimensions
276 TEMPERATURE §14.4

La X La at temperature to. The area of this plate at temperature to is

L~. When the temperature of the plate is t, the length of a side is L t , given
by Equation 14-15, and the new area of the plate is
At = L; = {L o[l + 0'(~t)Jl2
= L~[l + 20' (~t) + 0'2 (~t)2],
At = A o[l + 20' (~t) + 0'2 (~t)2].
In general, the coefficient of linear expansion is a small number. Over the
limited temperature interval for which the coefficient of linear expansion
represents the true expansion of the body, we may neglect the term in 0'2 as
being small in comparison with the other terms on the right-hand side of
the equation. Thus we have
At = A o[l + 20' (~t)].
Comparing the above result with Equation (14-17), we see that the coeffi-
cient of area expansion (T is numerically equal. to twice the coefficient of
linear expansion 0'. In the form of an equation we have
(T = 20'. (14-21)

Similarly, we may find that the coefficient of volume expansion {3 is numeri-

cally equal to three times the coefficient of linear expansion, or
{3 = 30'. (14-22)

Coefficients of areal expansion and of volume expansion of solids are not

usually tabulated in compilations of technical data, for the reason that
they may be readily obtained from the tabulated coefficients of linear
expansion.
Illustrative Example. A hollow copper sphere has an inner radius of 4 cm
and an outer radius of 5 cm at a temperature of 20°C. Find the change in volume
of the spherical cavity when the temperature is raised to 420°C.
The change in volume of a copper sphere is the same whether it is solid or
hollow. We may think of a solid sphere as consisting of a solid central core and
a hollow spherical shell whose inside diameter is the same as the diameter of the
core. The two parts fit perfectly at all temperatures. Thus the volume of a
spherical cavity at any temperature is the same as the volume of a solid sphere
of copper at that temperature. To find the change in volume of a spherical cavity
in a hollow copper sphere, we find the change in volume of a solid sphere of copper
subjected to the same temperature change.
The coefficient of linear expansion of copper valid in the range 0-625°C has
been measured as 16.07 X 1O-6;oC. The change in volume may be obtained
from Equation (14-20). The numerical values are
{3 = 30' = 48.21 X 1O- 6 CC)-1,
§14-5 'l'HI<;}~MAL EXPANSION OF LIQUIDS 277

Thus .6V = iJV o.6t,

.6V=5.17cm 3 •
Thus the spherical cavity expands from a volume of 268 cm 3 to a volume of
273.2 cm 3 •

14-5 Thermal Expansion of Liquids

As a general rule, a liquid expands when its temperature is raised. The

notable exception to this is water, which contracts when its temperature

1.044

1.040

1.036

1.032
1.028
§
.S 1.024

§'" 1.020
~ 1.016

1.012
I

1.008

1.004

1.000
-20 o 20 40 60 80 100
Temperature in °C
Fig. 14-7 Curve showing the volume of water as its temperature is raised from -20°C
to 100°C with its minimum volume or maximum density at 4°C.

is raised in the limited region from ooe to 4°C. Above 4°e water expands
with an increase in temperature. The behavior of water at atmospheric
pressure in the range from -20°C to lOOoe is shown in Figure 14-7.
(NOTE: Water is a supercooled liquid from -20°C to ooe.)
One method for determining the coefficient of volume expansion of a
liquid is to put the liquid into a container of known volume provided with a
narrow tube at one end, as shown in Figure 14-8. A glass container is
usually employed. The level of the liquid in the tube is observed at the
initial temperature to and at the final temperature t. The surface of the
liquid is exposed to the atmosphere so that the pressure of the liquid remains
278 TEMPERATURE §14-5

constant. Since liquids generally have greater coefficients of expansion

than glass, the level of the liquid will rise as the temperature is raised. Only
the relative expansion of the liquid with respect to the container can be
directly determined by this method. If the coefficient of volume expansion
of the container is known, the coefficient of volume expansion of the liquid

t V

1 hz

Fig. 14-8 Expansion of a Fig. 14-9 Dulong and Petit ap-

liquid in a container when paratus.
heated.

can be determined. The apparent change in volume of the liquid is the

difference between the change in volume of the liquid and the change in
volume of the container.
The expansion of a liquid may be determined by a method, introduced
by Dulong and Petit in 1817, in such a way that the measurement does not
depend upon the expansion of the container. In its most elementary form
the apparatus consists of two vertical tubes maintained at different tem-
peratures. The two vertical tubes are connected by a horizontal tube, as.
shown in Figure 14-9. At the level of the horizontal tube, the pressure
must be the same in both columns. The observed difference in the height of
the liquid in the two columns is due to the difference in density of the liquid
at the two temperatures.

14-6 Expansion of Gases

In studies of the expansion of solids or liquids, the substance studied is

generally maintained in an open system, exposed to the atmosphere, so that
the expansion coefficient is determined at constant pressure-the pressure
§14·6 EXPANSION OF GASES 279

of the atmosphere. Since a gas must be studied in a closed container, a

change in temperature may involve a change in both the volume and the
pressure of the gas. It is customary to study the behavior of the gas at
constant pressure, where the volume
changes with temperature, or at con-
B
stant volume, where the pressure
changes with tempeature.
An apparatus suitable for study-
ing the change in pressure of a gas at
constant volume is shown in Figure
14-10. The gas under investigation is
kept in the bulb A and in the narrow
tube leading to the manometer. The
mercury in tube C of the manometer
is always kept at the same level by
being raised to the index point I by
raising or lowering the tube B. The
U-shaped section of the tube R is fre-
R
quently a flexible rubber hose.
Let us define the coefficient of pres-
sure change at constant volume {3' as
, ~P/P
{3 = --. (14-23)
~t

When the coefficient of pressure change Fig. 14-10

at constant volume of hydrogen is
measured at O°C, this coefficient has the value of approximately 0.00366
per degree centigrade. A value very close to this is obtained with other
gases, provided that the pressure is not too high and the temperature is
significantly above the temperature at which the gas liquefies. The inter-
pretation of this numerical value is that the pressure of hydrogen at con-
stant volume will change by 0.00366 or 1/273.2 of its pressure at O°C for
each centigrade degree change of temperature. The use of a constant-
volume gas thermometer has already been discussed in Sections 14-2 and
14-3.
By a slight modification of the apparatus pictured in Figure 14-10,
the pressure of the gas may he kept constant, and its volume may be
measured as a function of the temperature. The coefficient of volume change
at constant pressure {3 may be defined as
{3=~V/V. (14-24)
~t

When the coefficient {3 is measured for hydrogen at O°C, its value is again
280 TEMPERATURE §14-7

found to be 0.00366 per degree centigrade. The coefficient of volume

change at constant pressure of other gases is close to this value, provided
that the pressure is not too high and the gas is far from the temperature at
which it becomes liquid.

14-7 Absolute Scale of Temperature

The fact that the constants (3' and /3 have the same numerical value for
hydrogen, and that these constants are approximately the same for other
gases, has led to the introduction of a scale of temperature known as the
absolute gas scale of temperature. From Equations (14-23) and (14-24) we
see that, if hydrogen is maintained at constant volume, its pressure should
be equal to zero at a temperature of -273.2°C. Similarly, if hydrogen is
kept at constant pressure, its volume should be zero at a temperature of
-273.2°C, provided that it remains a gas.
Let us rewrite Equation (14-23) in terms of the initial pressure Pi, the
final pressure P [, the initial temperature ti, and the final temperature tf.
We find
(3' = (Pf - Pi) .
Pi(t f - ti)
If we take the initial temperature as
ti = O°C,
and set
/3' = 1/273.2 per degree centigrade,
we find, after some algebraic manipulation,
P f = tf +
273.2
Pi 273.2
Let us define the absolute centigrade gas temperature T as the centigrade
temperature plus 273.2, or
T = t + 273.2.
For the case chosen the initial absolute temperature is T i = 273.2. Thus
we have
P T
-f = -f ,
Pi T,:
or P = KT, (14-25)

where K is some constant for a particular quantity of gas maintained at a

constant volume. From Equation (14-24) we may find, by a similar deriva-
tion, that
v = K'T. (14-26)

Equation (14-26) is called Gau-Lussac's law.

 §14-8 THERMAL STRESSES 281

From the basic conception of the absolute gas scale of temperature, we

see that negative numbers for the absolute temperature would lead to such
absurd results as the existence of negative pressures and negative volumes
for a gas. Hence the temperature of a body expressed on the absolute scale
must always be a positive number. It is not possible at this point to say
whether there is a physically achievable absolute zero of temperature. All
substances become liquids at temperatures above the absolute zero. The
lowest temperature measured with a gas thermometer is about -272°C, or
1° abs, using helium at low pressure.
It is apparent from the preceding discussion that what is needed is a
temperature scale which is independent of the properties of a particular
. substance. There is such a scale, known as the absolute thermodynamic scale,
or Kelvin scale oj temperature, which will be discussed in Chapter 19. There
we shall see that temperatures on the thermodynamic scale are in agreement
with temperatures on the absolute gas scale for a perfect gas, and are very
close to temperatures on the absolute gas scale achieved with a constant-
volume hydrogen thermometer.
Although the temperature 273.2° abs is sufficiently accurate for our
purposes, it should be noted that the average value of the best experimental
determinations of the ice point is
T = 273.165° abs.
The lowest temperature which has been obtained experimentally is
18 X 10-6 ° abs. This has been done by a process involving the magnetiza-
tion and demagnetization of certain magnetic materials which were pre-
viously cooled to a temperature of about 1° abs.
The absolute temperature can also be expressed in terms of the Fahren-
heit degree by means of the equation
1' F = *1', (14.27)

where T F is the absolute temperature in Fahrenheit degrees, and T is the

absolute temperature in centigrade degrees. When absolute temperature
is measured in Fahrenheit degrees, the temperature scale is· called the
Rankine scale of temperature. This scale is used in engineering in the
United States and in Great Britain.

14-8 Thermal Stresses

If a rod is heated while its ends are confined, the rod is thereby put into a
condition of internal compressive stress. If the rod is cooled when its ends
are confined, it assumes a condition of internal tensile stress. To calculate
the stress within the rod, we assume that the rod has acquired its final
configuration in a virtual two-step process, in which the first step is the
282 TEMPERATURE §14·8

change in temperature of the free rod, while the second step is the applica-
tion of forces to the ends of the rod to return it to its initial length.
When the temperature of a rod of length L, made of a material whose
coefficient of linear expansion is a, is altered by an amount t::.t, the change
in the length of the rod is given by Equation (14-16) as
t::.L = aL t::.t.
The fractional change in length of the rod is given by
t::.L
L=at::.t.

Let us suppose, for definiteness, that the temperature of the rod has in-
creased, and that the rod has increased in length. We may return the rod
to its original dimensions by applying a compressive stress sufficient to
produce a decrease in the length of the rod equal to t::.L, or to produce a
strain in the rod equal to t::.L/L. From the definition of Young's modulus
of elasticity, we have
stress stress
y=--=--.
strain a t::.t
Thus the stress produced in a rod whose modulus of elasticity is Y wiII be
given by
Stress = aY t::.t, (14-28)

when the rod is confined so that it cannot expand or contract when it is

subjected to the temperature change t::.t. The state of thermal stress in a
rod is determined by its coefficient of linear expansion, Young's modulus,
and the temperature change to which the rod has been subjected. The force
which must be applied to the ends of the rod to produce this stress is de-
termined from the definition of stress as the force per unit area of the rod,
as given in Equation (13-1).
Thermal stresses are widely encountered in practical engineering
problems. Thus the steel rails of a railroad are commonly provided with
expansion joints, but it is possible to lay continuous steel rail to any con-
venient length, provided that the rails are securely anchored to a roadbed
with spikes and crossties capable of exerting a force sufficient to produce the
stress given by Equation (14-28). It is not adequate to exert the force at
the ends of the rails, for the rails would fail by buckling. The continuous
rail must be spiked to the roadbed at suitable intervals. In laying concrete
road or curbing, it is common practice to provide expansion joints because
of the greater expense and difficulty of securing the road or curb to the
ground. Similarly, it is difficult to provide large, long buildings with suffi-
cient restraint, and expansion joints must be provided to keep cracks from
§14-9 THE MEANING OF TEMPERATURE 283

forming in brick walls. Steel tires are fastened to the cast-iron wheels of a
railroad car by a process of heating the steel rim and allowing it to cool and
contract onto the wheel. The stress in the tire exerts forces on a short,
element of length of the tire which have a normal component, holding the
tire to the wheel by frictional forces between the tire and the wheel, as
shown in Figure 14-11. Such a tight fit between two members is called
a shrink fit.

Fig. 14-11 The radial force exerted by a circular hoop under tension 8 upon a section
of length ill is given by 8M, where ilO is the angle subtended by the section ill at the
center of the circle.

When a solid is heated and subsequently cooled, as in heat treatment

and quenching, or in a welded member, internal thermal stresses are often
generated through the irregular cooling of the member. When hot, a solid
body may easily be deformed without development of internal strains,
but when the solid has cooled sufficiently, any further deformation is
accompanied by the generation of internal stresses. Thus an irregularly
shaped forging being quenched after heat treatment is cooled first where it
is thinnest. As the thicker sections subsequently cool and contract, they
are restrained by the cooler sections of the forging, and a state of internal
stress is produced which may be greater than the rupture strength of the
material of the forging, so that cracks develop. A drinking glass into
which boiling water is poured will often crack from the thermal stresses
developed at the boundary between the hot and cold portions of the glass.

14-9 The Meaning of Temperature

In the discussion of the concept of temperature in Sections 14-1 to 14-3,

we have defined the meaning of temperature through the operation of
284 TEMPERATURE §14·9

measuring temperature with the aid of thermometric properties of macro-

scopic systems. Thus the measurement of temperature requires that matter
be present in bulk. The concept of temperature is a macroscopic one. It is
meaningless to ask the question, "What is the temperature of a proton
moving at a speed of 100 mi/hr?"
The definition of temperature was also based upon the concept of
equilibrium; that is, it is assumed that any changes in the temperature of
the system are taking place sufficiently slowly so that the state of the sys-
tem could be considered constant during the time of measurement. If the
state of the system changes rapidly, so that it does not have sufficient time
to come to equilibrium, the meaning of temperature is somewhat ambigu-
ous. This is the case in a flame, or in the exhaust gases of an engine, where
the processes of combustion take place with great rapidity, and the gases
do not have time to come to equilibrium. In such cases it is often found
that different methods of measuring temperature yield different results,
for the measured temperature is a function of the method of measurement
as well as the condition of the object whose temperature we wish to deter-
mine, and the number assigned as the temperature has meaning only in
comparative terms.

Problems

14-1. At what centigrade temperature will the reading of a Fahrenheit

thermometer be numerically equal to twice the reading of a centigrade ther-
mometer?
14-2. How long must a steel rod be in order that its length will increase by
0.02 in. as a result of a temperature change of lOOC?
14-3. The distance between two markers is measured with a steel tape at
25°C. The reading of the tape is 80 ft. If the calibration of the tape is correct
at O°C, determine the distance between the markers.
14-4. A copper ring has an inside diameter of 4.98 cm at 20°C. To what
temperature must it be heated so that it will just fit on a shaft 5.00 cm in
diameter?
14-5. What will be the stress in the ring of Problem 14-4 after it has shrunk
in place onto the shaft? Assume that the dimensions of the shaft remain constant.
14-6. A glass flask having a volume of 1 liter at 20°C is filled with mercury
at that temperature. What volume of mercury will run over the lip of the flask
when the temperature is raised to lOO°C?
14-7. The pressure of a volume of gas at 27°C is 546 mm of mercury. What
will be its pressure, expressed in millimeters of mercury, if the temperature is
increased to 28°C, the volume being kept constant?
14-8. A constant-pressure air thermometer contains a mass of air whose
volume is 300 cm 3 at O°C. What will be its volume at 50°C?
14-9. The following data were taken in an experiment with a constant-
volume air thermometer, such as that shown in Figure 14-10. Barometric pres-
 PROBLEMS 285

sure 754.0 mm; height of column C 48.4 cm and height of column D 44.7 cm at
the ice point; height of column C 48.4 cm and height of column D 71.0 cm at the
steam point; height of column C 48.4 cm and height of column D 62.3 cm when
the bulb is surrounded with warm water. Determine (a) the coefficient of pres-
sure change of the air and (b) the temperature of the warm water.
14-10. A clock regulated by a seconds pendulum made of brass has a correct
period of 2 sec when the temperature is 70°F. Determine the gain or loss, in
seconds per day, when the temperature rises to 97°F. Assume it to be a simple
pendulum.
14-11. Derive the equation
{3 = 3a

relating the coefficient of volume expansion (3 of an isotropic solid to its coefficient

of linear expansion a.
14-12. A steel rod 1 m long and 0.5 cm in diameter is clamped between two
fixed supports at its ends. The temperature of the rod is raised 30°C. Determine
(a) the stress in the rod and (b) the force exerted by each support. Young's
modulus = 20 X 1011 dynes/cm 2.
14-13. A steel bomb is filled with water at 10°C. If the system is heated to
100°C, determine the increase in pressure of the water (a) neglecting the thermal
expansion of the steel and (b) taking into account the thermal expansion of the
steel. Neglect the change in dimensions of the steel bomb due to the tension it
experiences. Take the bulk modulus of water as 2 X 104 atm- 1 •
14-14. Derive a formula for correcting the reading of a mercury barometer
calibrated at 20°C when the barometer is read at a temperature of tOC. Neglect
the expansion of the scale attached to the barometer. Express your result in
terms of the correction fih, the reading of the barometer h, the coefficient of
volume expansion {3 of mercury, and the temperature. Does the expansion of the
glass container affect your result?
14-15. A block of aluminum is immersed in water at a temperature of 20°C,
and the buoyant force on the aluminum is observed to be 10 lb. What will be
the buoyant force on the block of aluminum at a temperature of 4°C?
14-16. A steel tire 2 in. in width whose inner diameter is 0.999 ft and whose
outer diameter is 1.05 ft is to be heated so that it may be placed on a cast-iron
wheel 1.000 ft in diameter. (a) Assuming that the cast-iron wheel is perfectly
rigid, find the tensile stress in the tire when it has cooled. (b) If the coefficient
of friction between the tire and the wheel is 0.2, find the force which must be
applied to the tire to pry it from the wheel. Take Young's modulus for steel as
29 X 10 6 1b/in. 2 •
14-17. A steel rod 1 ft long is welded at one end to a copper rod of the same
diameter and length, and the two bars are mounted between rigid supports.
Find the stress in each bar when their temperature is increased by 50°C. Young's
modulus for steel is 30 X 10 6 1b/in. 2 ; for copper: 18 X 10 6 1b/in. 2 •
14-18. A steel cable 0.5 in. 2 in cross-sectional area is tightened to a tension
of 20,000 Ib/in. 2 when the cable is at O°C. What is the tension in the cable when
the temperature is increased to 20°C? Young's modulus is 30 X 10 6 Ib/in. 2 •
14-19. Find the coefficient of volume expansion of water at (a) 40°C and
(h) 70°C.