3.

Instrumentation for temperature measurements

Temperature is one of the major physical parameter, which characterises the condition of
substances involved in processes. In order to measure this parameter we need to choose an
appropriate temperature scale and the unit of temperature.

3.1. Temperature scales, temperature units

Anders Celsius, the Swedish astronomer, devised a scale for measuring temperature, which later
was named after his name. This scale has the symbol C. The Celsius scale was based on two
fixed and easily reproducible points:

• the ice point, ie the temperature of a mixture of ice and water in equilibrium with
saturated air at a pressure 101325 Pa. This temperature was numbered 0 C;

• the steam point, ie the temperature of the water and steam in equilibrium at a pressure
101325 Pa. This temperature was numbered 100 C.

Later in 1954 this scale was redefined and was based on:

• a single fixed point - the triple point of water. This is the temperature at which solid,
liquid and vapour phases of water exist together in equilibrium. The temperature of the
triple point of water has the value of 0.01 C;

• the ideal-gas temperature scale. On this scale the steam point was experimentally found
to be equal to 100.00 C.

The thermodynamic scale of temperature (or the absolute scale) was derived from the second law
of thermodynamics. This scale is independent of any thermometric substance. The relation
between the absolute scale and the Celsius scale is as follows:

T =  + 273.15 , (3.1)

where: T - temperature in the absolute scale, K ;

 
- temperature in the Celsius scale, C .

The unit for the absolute scale is K - Kelvin, named after Lord Kelvin (William Thomson).
1
1, K = of the temperature at the triple point of water.
273.16
However, for practical purposes an International Practical Temperature Scale, IPTS-68 (adopted
in 1968) has been used for 22 years. This scale was based on a number of fixed and easily
reproducible points. Definite numerical values of temperature were assigned to these points. For
interpolation purposes between the fixed points, specified formulas relating temperature to the
readings on certain temperature-measuring instruments were used. Table 3.1 presents the primary
fixed-point temperatures for IPTS-68. (Van Wylen G.J., Sonntag R.E. Fundamentals of Classical
Thermodynamics, 1980).

Table 3.1. Fixed points and corresponding temperatures for IPTS-68.

NN Fixed points Temperature, C

1. Triple point of equilibrium-hydrogen (s+l+v) -259.34
2. Boiling point of equilibrium hydrogen (l+v) at 33.33 kPa -256.108
3. Normal boiling point of equilibrium hydrogen at 101325 Pa -252.87
4. Normal boiling point of neon -246.048
5. Triple point of oxygen -218.789
6. Normal boiling point of oxygen -182.962
7. Triple point of water 0.01
8. Normal boiling point of water 100
9. Normal freezing point of zinc (s+l) at 101325 Pa 419.58
10. Normal freezing point of silver (s+l) at 101325 Pa 961.93
11. Normal freezing point of gold (s+l) at 101325 Pa 1064.43

In this table: s - solid; l - liquid; v - vapour.

The International Committee of Weights and Measures (CIPM) adopted a new International
Temperature Scale (ITS-90) at its meeting in September 1989. It became the official international
temperature scale on January 1, 1990. Table 3.2 shows fixed point for ITS-90. (Burns G.W.,
Scroger M.G., Strouse G.F., Croarkin M.C., Guthrie W.F. Temperature-Electromotive Force
Reference Functions and Tables for the Letter-Designated Thermocouple Types Based on the
ITS-90).

Temperatures in ITS-90 are in closer agreement with thermodynamic values when compared with
IPTS-68. The increased numbers of temperature subranges makes ITS-90 more flexible. There
are certain differences between ITS-90 and IPTS-68 (similar fixed points are shown in italics in
these tables).

Table 3.2. Fixed points and corresponding temperatures for ITS-90.

Fixed points Temperature, C

N
1. Normal boiling point of helium -270.15 to --268.15
2. Triple point of equilibrium-hydrogen (s+l+v) 259.3467
3. Boiling point of equilibrium hydrogen (l+v) at 33.33 kPa  -256.15
4. Normal boiling point of equilibrium hydrogen at 101325 Pa  -252.85
5. Triple point of neon -248.5939
 6. Triple point of oxygen -218.7916
7. Triple point of argon -189.3442
8. Triple point of mercury -38.8344
9. Triple point of water 0.01
10. Melting point of gallium 29.7646
11. Normal freezing point of indium (s+l) at 101325 Pa 156.5985
12. Normal freezing point of tin (s+l) at 101325 Pa 231.928
13. Normal freezing point of zinc (s+l) at 101325 Pa 419.527
14. Normal freezing point of aluminium (s+l) at 101325 Pa 660.323
15. Normal freezing point of silver (s+l) at 101325 Pa 961.78
16. Normal freezing point of gold (s+l) at 101325 Pa 1064.18
17. Normal freezing point of copper (s+l) at 101325 Pa 1084.62

3.2. Liquid-in-glass thermometers

These thermometers are used for temperature measurements from -200 to 750 C. They are
contact-type thermometers. Fig. 3.1 shows the principle of their design.

5 1 2 3 4

Figure 3.1. Liquid-in-glass thermometer

This thermometer consists of a glass bulb 1, which is connected with a glass capillary tube 2. A
scale 3 in degrees of Celsius or Fahrenheit is placed behind the capillary tube. The bulb, the
capillary tube and the scale are placed in a glass tube 4 to protect them against the damage. A
thermometric liquid 5 fills the bulb and a part of the capillary tube. The operational principle of
these thermometers is based on the difference between the volume expansion of liquids and glass
with temperature. The relationship that governs the operation of this device is

VT = VT0 * (1 +  * T ) (3.2)

3
where, VT - volume of liquid at temperature T , m ;
3
VT0 - volume of liquid at temperature T0 , m ;
 T = T − T0 - difference of temperatures, K ;
1
 - volumetric thermal expansion coefficient, .
K
The volumetric thermal expansion coefficient of glass is much less than that of liquids. The
variation of temperature (up and down) of the bulb causes liquid in the system to expand or
decrease its volume, respectively. As a result of such changes (the internal volume of the glass
bulb and the glass capillary varies negligible), the length of the liquid column in the capillary
tube goes up or down proportionally to the variation of temperature.

The type of thermometric liquid depends on the lower and upper limits of the measuring
temperature range. Table 3.3 presents the most common types of liquids used in these types of
thermometer.

Table 3.3. Types of thermometric liquids.

Liquid Temperature range, C

From To
Mercury -35 750
Toluene -90 200
Ethanol -80 70
Kerosene -60 300
Petroleum Ether -120 25
Pentane -200 20

Among these liquids mercury is the most widely used, because:

• mercury is easy obtainable with high chemical purity;

• mercury does not wet glass (this increases the accuracy of measurement/ reading);

• mercury remains in liquid state in a wide temperature range.

Among disadvantages inherent to mercury-in-glass thermometers we can mention the following:

• mercury is a poisonous element, which affects the central and peripheral nervous
system, its vapour is the most toxic;

• small volumetric thermal expansion coefficient for mercury, therefore, mercury is used
in thermometers with capillaries of small internal diameter;
The solidifying point of mercury, ie 38 C, limits the lowest temperature that can be measured by
mercury-in-glass thermometers. The upper temperature is determined by the temperature at
which glass still retains its solid properties. This temperature is equal about 600 C for glass, and
about 750 C for silicon glass.

When air above mercury in the capillary is removed, a mercury-in-glass thermometer can be used
at temperatures below 300 C, because the boiling temperature of mercury at atmospheric
pressure is equal 356.9 C. In order to increase this temperature range it is necessary to increase
the boiling temperature of mercury (saturation temperature). This can be achieved by increasing
pressure in the capillary. Usually, the space above mercury in the capillary is filled by inert gas
(such as nitrogen, argon) under pressure.
Liquid-in-glass thermometers with organic thermometric liquids are used for temperature
measurements from -200 to 200 C. One advantage of these thermometers is:
• a higher volume thermal expansion coefficient comparing with that for mercury (six
times higher in average).
Disadvantage of thermometers with organic liquids is:
• these liquids wet glass, therefore, in order to increase the accuracy of
measurement/reading, glass capillaries with bigger internal diameters (up to 1 mm) are
used.

Advantages of liquid-in-glass thermometers are as follows:

• they are simple in design;

• they are relatively highly accurate in temperature measurement.

There are several disadvantages inherent to liquid-in-glass thermometers
• they are fragile;
• it is difficult to perform readings due to low visibility of the scale;
• they are not capable of distance transmission of a measuring signal, therefore, they are
used as locally placed devices;
• impossibility to repair;
• high values of time lag;
• low visibility of mercury in the capillary.

.
3.3. Filled thermal systems
Another class of thermometers that utilise the principle of expansion of substances with
temperature is called filled thermal systems. Depending on the phase of the substance, which fills
these devices, these systems are sub-categorised into gas-, liquid- and vapour-filled systems.
Gas-filled systems are based on a basic law of gases. If a gas is kept in a metallic bulb (or a
container) at a constant volume, then if the temperature varies, so does the pressure according to
the relationship
P2 − P1 = P1 *  * (T2 − T1 ) , (3.47)
where:

P1 and T1 - absolute pressure ( Pa ) and temperature ( K ) at state 1;

P2 and T2 - absolute pressure ( Pa ) and temperature ( K ) at state 2;
 - thermal coefficient of pressure, equal to the volumetric thermal expansion
−1
coefficient, K .

3
4 3

2 4

1
1

Figure 3.6. Gas- or liquid-filled thermometer. Figure 3.7. Vapour-pressure system.

Figure 3.6 schematically shows the design of a gas-filled thermometer. Gas (nitrogen or helium)
1 fills the thermal bulb 2, capillary tube 3 and Bourdon tube of a manometer 4. The thermal bulb
(usually made of a stainless steel) is immersed in the measuring media. Variation of its
temperature causes change in pressure of the gas in the system. The manometer measures this
variation of pressure. The scale of the manometer is graduated in C, but not in Pa. The length of
the capillary tube (usually made of a stainless steel) varies from 0.6 to 60 m. The accuracy of
measurements for these thermometers is greatly influenced by variation of ambient temperature
(since it can change the pressure of a gas in the system). Two methods are used to reduce this
effect:

• a thermal bimetallic temperature compensator is used in the manometer;

• an internal volume of the thermal bulb should be greater than that of the capillary tube,
Vb
the ratio (where Vb and Vc are volumes of the thermal bulb and of the capillary,
Vc
respectively) may vary from 40 to 60; this can be achieved by reducing the internal
diameter of the capillary tube or increasing the internal volume of the thermal bulb. The
longer the capillary tube, the bigger the thermal bulb should be.

Therefore, gas-filled thermometers are not widely used in practice.

Depending on the measured temperature range, the system may be filled with a gas under
pressure higher than atmospheric. That is why variations in atmospheric pressure have no effect
on the indications of gas-filled thermometers.

Gas-filled thermometers have several advantages:

• they have the widest temperature range of all filled systems;

• as follows from the equation (3.47) these thermometers have uniform scales;

• they have the longest capillary length compared with other filled systems.

These thermometers are usually used for temperature measurement in the range from
-200 to 600 C.

Liquid-filled systems have similar design with gas-filled thermometers (see Fig. 3.6).
Organosilicone liquids, propanol and mercury are used as thermometric liquids, which fill the
entire system. Since the total volume of the thermal system is constant, then variation of
temperature of the media, where the thermal bulb is immersed, causes variation in the pressure of
the thermometric liquid. This variation in pressure is proportional to the variation of temperature.
Therefore, scales of liquid-filled thermometers are uniform.

Several factors influence the accuracy during temperature measurements, namely:

• variation in ambient temperature;

• variation in pressure head;

 • variation in atmospheric pressure.

In order to compensate the influence of variation of an ambient temperature it is necessary to

increase the ratio internal volume of the thermal bulb/internal volume of the capillary tube, and
employ thermal bimetallic compensators (see gas-filled thermometers). The error due to variation
of an ambient temperature is bigger in the case of liquid-filled systems, compared to gas-filled
systems. Therefore, the capillary length for liquid-filled systems can not exceed 10 m.

When the thermal bulb is placed below or above the manometer, results of such temperature
measurements will not be correct. This is because of different pressure head of the liquid column
compared with the case when this thermometer was calibrated (the manometer and the thermal
bulb were placed on the same level). In this case the error can be eliminated by zero correction of
manometer. The ultimate elevation distance between the thermal bulb and the manometer are
given in the calibration certificate supplied with the liquid-filled thermometer.

To reduce influence of variation of atmospheric pressure, the system is filled with liquid under
pressure from 0.5 to 2.0 MPa.

Here are the advantages of liquid-filled thermometers:

• small time lag;

• small dimensions of thermal bulb.

These thermometers are used for temperature measurement in the range from -150 to 300 C.

Vapour/pressure systems (see Fig. 3.7) are filled by 2/3 of the volume of the thermal bulb 1 by
liquid 2 which has a low boiling temperature, for example, freon (refrigerant), propylene,
acetone, ethylbenzene, methyl chloride, etc. Another (upper) part of the thermal bulb and the
capillary tube 3 is occupied by saturated vapour 4 of this liquid. Vapour pressure depends only on
the temperature of saturated liquid in the thermal bulb, and therefore, does not depend on the
variation of the ambient temperature (this is an advantage). Relationship between saturation
pressure and temperature for liquids is non-linear (see Fig. 3.8). Hence, the scales of these
thermometers are non-uniform, with more widely spaced increments at high temperatures. The
length of the capillary tube usually does not exceed 25 m.

Disadvantages of vapour-pressure thermometers are as follows:

• narrow temperature range, from -50 to 300 C;

• slow response time (time lag) of about 20 seconds;

• non-uniformity of the temperature scale.

 3.5
3.0
Methyl chloride
Vapour (saturation)

2.5
pressure, MPa

2.0

1.5
1.0

0.5
0.0
0 20 40 60 80 100
Temperature, C

Figure 3.8. Saturation (vapour pressure) curve for methyl chloride.

3.4. Thermocouples

Seebeck in 1821 discovered that thermal electromotive force (t.e.m.f.) is generated in a closed
circuit of two wires made of dissimilar metals if two junction are at different temperatures. One
junction is inserted into a measuring media, and it is called a hot or measuring junction. Another
one, called a cold or reference junction, is kept either at 0 C or at ambient temperature and is
connected to a measuring instrument (millivoltmeter). The electronic explanation of this
phenomenon is as follows:

the density of conduction electrons in two dissimilar metals is different. So, in the
case when metals are brought into contact (welded together), the free (or
conduction) electrons will flow from the metal with high their density to the metal
with low density of the conduction electrons. As the result of this drift, a potential
difference is produced in the boundary between these two metals. This potential
difference will stop the flow of electrons. Since the metals are different, so they
will differently respond to temperature variations. In other words, the variation of
temperature will change the density and velocities of free electrons in two metals
 differently. This will cause the change in the magnitude of the thermal
electromotive force.

Figure 3.9 schematically shows a thermocouple and a measuring instrument.

6
3
+ +

-
-

1 2
4 5
7 8

Figure 3.9. Thermocouple and measuring instrument.

1 - hot junction; 2 - metal A; 3 - metal B;
4 - connection head; 5 - extension wires;
6, 7 - positive and negative terminals, respectively, of a measuring
instrument;
8 - measuring instrument.
T.e.m.f. is proportional to the difference of temperatures between the two junctions. All tables,
correlated t.e.m.f. of thermocouple (measured in mV) and temperature, are developed when the
temperature of a cold junction is equal to 0 C. T.e.m.f. is the function of temperature difference
between the hot and the cold junctions:

E AB = f ( − 0 ) , (3.48)
where:

E AB - t.e.m.f. developed by a thermocouple, mV ;

 and 0 
- temperatures of the hot and the cold junctions of a thermocouple, C .


If the temperature of the cold junction is kept constant (say at 0, C ), then t.e.m.f. is
proportional to the temperature of the hot junction (the measuring temperature), ie

E AB = f (t ) . (3.49)

In reality, in industrial environment, however, it is not possible (or is not convenient) to keep the

temperature of the cold junction at 0, C . Therefore, to evaluate the actual t.e.m.f. and, finally,
the actual measuring temperature, we should introduce a correction. A final equation has the
following form:

E AB ( ,0 ) = E AB ( ,0' ) + E AB (0' ,0 ) (3.50)

where:

E AB ( ,0 ) - t.e.m.f. developed by a thermocouple when the temperature of the hot

junction is equal to  and the temperature of the cold junction is equal to
0 = 0,  C , mV ;

E AB ( ,0' ) - t.e.m.f. developed by a thermocouple when the temperature of the hot

junction is equal to  and the temperature of the cold junction is equal to
0' (different from 0,  C ) – this t.e.m.f. is measured by a millivoltmeter, mV ;

E AB (0' ,0 ) - t.e.m.f. developed by a thermocouple when the temperature of the hot
junction is equal to 0 and the temperature of the cold junction is equal to
'

0 = 0,  C , mV .
There are various types of thermocouples:

• Platinum and Platinum - 10% Rhodium (type S) from -50 to 1765 C;
• Platinum - 6% Rhodium and Platinum - 30% Rhodium (type B) from 0 to 1820 C;
• Nickel - Chromium and Nickel - Aluminium (Chromel-Alumel, type K) from -270 to
1370 C;
• Iron and Copper - Nickel (Iron - Constantan, type J) from -210 to 1200 C;
• Copper and Copper - Nickel (Copper - Constantan, type T) from -270 to 400 C;
• Nickel - Chromium and Copper - Nickel (Chromel - Constantan, type E) from -270 to
1000 C.

Figure 3.10 presents experimental curves thermal electromotive force vs temperature for several
types of thermocouples.
 Thermal electromotive force, mV
60

50

40
Type K
30 Type J
Type T
20
Type S

10

-10
-300 0 300 600 900 1200 1500 1800

Temperature, C

Figure 3.10. Experimental curves thermal electromotive force vs temperature.

Requirements imposed to the properties of metals used as electrodes for thermocouples are as
follows:

• reproducibility of material, ie possibility of obtaining of metal wires with the same

properties;

• resistance of metal electrodes should be small and have a weak relationship vs

temperature;

• stability of a static characteristic E AB = f ( ) , ie recovery of properties after

measurements;

• high sensitivity;

• correlation E AB = f ( ) should be close to linear relationship as much as possible.

mV
The highest sensitivity has thermocouple of Type J (Iron - Constantan): S = 0.055, 
.
C
3.5. Resistance temperature detectors

The principle of resistance temperature detectors (RTD) is based on the variation of electrical
resistance of metals with temperature. For this purpose several metals are used, namely, platinum,
copper, nickel. When temperature increases the resistance of these metals increases. Temperature
function of resistance for metals in a narrow temperature interval can be expressed by a
relationship:

R = R0 (1 +  ) , (3.51)

where:

R and R0 - are the values of electrical resistance of a metal conductor at temperatures  and
0,  C , respectively, Ohm ;
1
 - thermal coefficient of electrical resistance, 
.
C

For metals this coefficient is positive. Fig. 3.11 shows relationship between resistance of
platinum and copper RTD and temperature.

Platinum RTDs are used for temperature measurements from -220 to 850 oC (they are used as
reference RTDs, as well), copper RTD - from -50 to 150 oC, and Nickel RTD - from -215 to 320
o
C.
 400

350
Platinum RTD
Copper RTD
300

250
R, Ohm

200
200
150 150
100
100
50
0
50
0 40 80 120 160 200

0
-300 -100 100 300 500 700 900

t, oC

Figure 3.11. Resistance vs temperature for platinum and copper RTD.

Fig. 3.12 shows the assembly of RTDs. Sensitive elements of RTDs are made of a thin wire 1
with outside diameter equal to 0.025 mm (platinum RTD) and 0.1 mm (copper RTD) double
wounded (non-inductive) on a micaceous or porcelain stem 2. For mechanical strength the
sensitive element is placed in the ceramic insulator tube 3 filled by extremely fine granular
powder; extension wires are placed in the ceramic insulator 4, and entire assembly is covered by
a protective sheath of stainless steel 5. The space between the sheath and ceramic insulator is
filled by ceramic packing powder 6. To avoid contact of sensitive element with environment,
sensitive assembly is protected by high-temperature hermetic seal 7. The contact between the
wire of the sensitive element and the ceramic encapsulation permits a rapid speed of response.
 a
b

2 1 5 3 6 4 7

Figure 3.12. RTD assembly.

3.6. Thermistors

If semiconductors or heat-treated metallic oxides (oxides of cobalt, copper, iron, tin, titanium,
etc.) are used as the materials for producing temperature sensitive elements, then these
temperature transducers are called thermistors (the name is derived from the term of ‘thermally
sensitive resistor’). These oxides are compressed into the desired shape from the specially
formulated powder. After that, the oxides are heat-treated to recrystallise them. As the result of
this treatment the ceramic body becomes dense. The leadwires are then attached to this sensor for
maintaining electrical contact.

The following relationship applies to most thermistors:

1 1 
B* − 
Rt = R0 * e  T T0 
(3.86)
where,

RT0 - resistance of thermistor at reference temperature T0 , K , Ohm ;

RT - resistance of thermistor at temperature T , K , Ohm ;
B - constant over temperature range, depends on manufacturing process and
1
construction characteristics, .
K
Fig. 3.15 shows relationship between temperature and resistance for a thermistor.
Thermistors have negative thermal coefficient of electrical resistance. It means that when
temperature increases the electrical resistance of thermistor decreases. They have greater
resistance change (this is an advantage) compared with RTD in a given temperature range. For
example, if we compare what change in resistance will be caused by variation of temperature in 1
C for Platinum and Copper RTD (see Fig. 3.11) and for thermistor (see Fig. 3.15) in the
temperature range from 273.15 to 423.15 K (ie, from 0 to 150 C), we will obtain the following
values:
Ohm
• for platinum RTD - 0.38, 
;
C
Ohm
• for copper RTD - 0.04,  ;
C
Ohm
• for thermistor - 0.65,  .
C
550

Thermistor

520
Resistance, Ohm

490

460

430
150 200 250 300 350 400 450 500 550

Temperature, K

Figure 3.15. Thermistor resistance vs temperature curve.

Wheatstone bridge and resistance measuring constant current circuits, similar to that used in the
case of RTDs, are used for resistance measurement of thermistors (see Fig. 3.14). Despite their
high sensitivity, thermistors have a worse accuracy and repeatability (this is the disadvantage)
comparing with metallic RTDs. Since the resistance vs temperature function for thermistors is
non-linear (although, some modern thermistors have a nearly linear relationship of temperature
vs resistance), it is necessary to use prelinearisation circuits before interacting with related
system instrumentation. In addition, due to the negative thermal coefficient of electrical
resistance an inversion of the signal to positive form is required when interfacing with some
analog or digital instrumentation. Therefore, thermistors are not widely used in process
instrumentation field, at least at present. However, they have been well accepted in the food
transportation industry, because they are small, portable and convenient. Another field of their
growing application are heating and air-conditioning systems, where thermistors are used for
checking the temperature in flow and return pipes.
All the discussed above instrumentation for temperature measurement (see from 3.2 to 3.6) refers
to contact-type devices, because their sensitive elements are immersed in the measuring media.
When dealing with temperatures above 1500 C, contact-type temperature measuring devices are
not applicable, because irreversible changes occur in metals which form their sensitive elements.
It is possible to perform non-contact measurement of temperature by optoelectronic transducers.

3.7. Optical and radiation pyrometers

Pyrometer is a device which uses the relationship between the electromagnetic radiation emitted
by a body and the temperature of this body. In order to better understand the phenomenon which
forms the basis of pyrometry, it is useful to explain the concept of the blackbody, and the
differences between it and real objects.

The term blackbody is ideal, and designates a body which radiates more electromagnetic energy
for all wavelengths intervals than any other body of the same area and at the same temperature,
and absorbs all the radiation it intercepts. Fig. 3.16 presents one of the classical blackbody
model.

Light

Figure 3.16. A classical blackbody model.

The temperature of the blackbody determines the nature and extent of such radiation. Stefan-
Boltzmann’s law says, that
 the blackbody with a finite absolute temperature ( T ) emits radiant (ie, in all directions)
electromagnetic radiation (EMR) per unit area of this blackbody and per second with
4
intensity which is proportional to T , according to an equation:

ET =  *T 4 , (3.87)

where,

2
ET - total EM energy emitted by the blackbody in all directions per unit area ( 1, m )
W
and per unit time ( 1, s ), ;
m2
−8 W
 - Stefan-Boltzmann’s constant, equal to 5.67051*10 , ;
m2 * K 4
T - an absolute temperature of the blackbody, K .

Fig. 3.17 shows the relationship between EMR emitted by a perfect blackbody as a function of
temperature. The area under these curves is equal to the total energy (emitted by a black body)
per second per unit area. This body at low temperatures emits EMR in the region of long
wavelengths. This region spreads from far-infrared to microwave region (5 m <  < 100 m,
where  is the wavelength in m, 10-6 m). With increasing the blackbody temperature, the
emission peaks move into the region of shorter wavelengths. At very high temperatures the
blackbody emits in the near visible wavelengths region. Visible region corresponds to the
wavelengths from 0.7 m (red) through 0.62 m (orange), 0.58 m (yellow), 0.53 m (green),
0.47 m (blue) to 0.42 m (violet).

Real objects emit and absorb less EMR than blackbodies, and this difference is dependent on the
wavelength, so nonblackbodies can not exactly follow relationship shown in Fig. 3.17. For this
purposes corrections should be used, otherwise, the apparent temperature will be lower that the
actual temperature. Also, it is necessary to take into account the loss of emitted radiation when it
passes through the media between the emitting body and a measuring instrument.
 0.8
2000 K
0.7
EM rediation energy density

1600 K
0.6 1200 K
(relative scale)

800 K
0.5
500 K
0.4

0.3

0.2

0.1

0
0 1 2 3 4 5 6

Wavelength, microns

Figure 3.17. EM radiation emitted by the blackbody at various temperatures.

There are two types of pyrometers: optical (monochromatic or narrowband) and radiation (total
radiation or broadband) pyrometers. The last devices originally were called radiation pyrometers,
then radiation thermometers, and more recently infrared thermometers. However, the first their
name (radiation pyrometers) is still widely used at present. These devices have high accuracy of
0.01 C as a standard instruments, and from 0.5 to 1% for industrial purposes.

a). Optical pyrometers, sometimes referred to as brightness thermometers, generally involve

wavelengths only in the visible part of the spectrum. When the temperature of the body increases,
so does the intensity at any particular wavelength. If two bodies have the same temperature, then
intensities of those two objects are equal. In this type of a pyrometer the intensity of a certain
wavelength of a heated body is compared with that of a heated platinum filament of a lamp (see
Fig. 3.18).
 9
4
5 6 7 8
1 2 3

10
11

Figure 3.18. An optical pyrometer.

An object 1 which temperature is to be measured, emits electromagnetic radiation with intensity

proportional to its absolute temperature. This radiation passes through lens 2 and red optical filter
3. Optical filter picks out only the desired wavelength - red. Then radiation focuses on the
platinum filament of a lamp 4, and passes through another filter 5, lens 6, viewing system 7. The
viewer 8 sees the platinum filament superimposed on an image of the object 1. When the
temperature of the filament is low comparing with that of the object, the viewer sees the filament
as a dark line on the bright background image of the object. The lamp 4 is connected in series
with an electrical battery 9, a variable resistor 10 and an ampermeter 11. By reducing the
resistance of the resistor an electrical current passing through the filament increases. So does the
temperature of the filament and its brightness. For a certain value of an electrical current
(corresponded to a certain value of an object temperature), the brightness of the platinum filament
will match the brightness of the object 1. At this setting the viewer cannot distinguish between
the image of the object and the filament. At this time the measurement of temperature is
performed. The scale of the ampermeter is calibrated in the units of temperature.

The lower temperature limit for optical pyrometers is determined by the temperature at which
objects become visible in red (about 225 C). However, there are devices which are able to
measure even lower temperatures down to -50 C. The upper limit varies from 600 to 3000 C,
and is limited by the melting point of the platinum filament. An accuracy is typically varied from
5 to 10 K.
b). Radiation pyrometers, being very simple and cheap, use an exponential relationship between
a total emitted EMR energy and given temperature. In radiation pyrometers (see Fig. 3.19) EMR
energy emitted at infrared (2.5 <  < 20 m) to visible wavelengths (0.42 <  < 0.7 m) from an
object 1 is focused by a spherical reflector 2 on a series of micro-thermocouples attached to a
blackened platinum disc 3. The radiation is absorbed by the disc, which temperature is increased,
so does thermal electromotive force U developed by the series of thermocouples. This thermal
electromotive force is proportional to the temperature of hot junctions of thermocouples, and,
finally, to the temperature of the object 1. The advantage of these pyrometers is that their
operation slightly depends on the wavelength.

1 2

Figure 3.19. A total radiation pyrometer

The lower limits for radiation pyrometers vary from 0 to 600 C, the upper limits vary from 1000
to 1900 C. The accuracy varies from 0.5 to 5 K, depending on cost. They are widely used for
temperature measurements in metal production facilities, glass industries, semiconductor
processes, etc.