NAFL INST OF STANDARDS & TECH R.I.C.

AIllQE MASEtM
,
A11 102485264
QC100 .US56 N0.174 1986 C.I NBS-PUB-C 19
NBS
UBLICATIONS

THERMOMETER
CALIBRATION

NBS MONOGRAPH 174

Jacquelyn A. Wise

Robert J. Soulen, Jr.

U.S. DEPARTMENT OF COMMERCE

NATIONAL BUREAU OF STANDARDS

"QC
100
.U556
No. 174
1985
C 2
Tf
m he National Bureau of Standards' was established by an act of Congress on March 3, 1901. The
Bureau's overall goal is to strengthen and advance the nation's science and technology and facilitate
their effective application for public benefit. To this end, the Bureau conducts research and provides: (1) a
basis for the nation's physical measurement system, (2) scientific and technological services for industry and
government, (3) a technical basis for equity in trade, and (4) technical services to promote public safety.
The Bureau's technical work is performed by the National Measurement Laboratory, the National
Engineering Laboratory, the Institute for Computer Sciences and Technology, and the Institute for Materials
Science and Engineering

The National Measurement Laboratory

Provides the national system of physical and chemical measurement; • Basic Standards^

coordinates the system with measurement systems of other nations and • Radiation Research
furnishes essentiaJ services leading to accurate and uniform physical and • Chemical Physics
chemical measurement throughout the Nation's scientific community, in- • Analytical Chemistry
dustry, and commerce; provides advisory and research services to other
Government agencies; conducts physical and chemical research; develops,
produces, and distributes Standard Reference Materials; and provides
calibration services. The Laboratory consists of the following centers:

The National Engineering Laboratory

Provides technology and technical services to the public and private sectors to Applied Mathematics
address national needs and to solve national problems; conducts research in Electronics and Electrical
engineering and applied science in support of these efforts; builds and main- Engineering^
tains competence in the necessary disciplines required to carry out this Manufacturing Engineering
research and technical service; develops engineering data and measurement Building Technology
capabilities; provides engineering measurement traceability services; develops Fire Research
test methods and proposes engineering standards and code changes; develops Chemical Engineering^
and proposes new engineering practices; and develops and improves
mechanisms to transfer results of its research to the ultimate user. The
Laboratory consists of the following centers:

The Institute for Computer Sciences and Technology

Conducts research and provides scientific and technical services to aid • Programming Science and
Federal agencies in the selection, acquisition, application, and use of com- Technology
puter technology to improve effectiveness and economy in Government • Computer Systems
operations in accordance with Public Law 89-306 (40 U.S.C. 759), relevant Engineering
Executive Orders, and other directives; carries out this mission by managing
the Federal Information Processing Standards Program, developing Federal
ADP standards guidelines, and managing Federal participation in ADP
voluntary standardization activities; provides scientific and technological ad-
visory services and assistance to Federal agencies; and provides the technical
foundation for computer-related policies of the Federal Government. The In-
stitute consists of the following centers:

The Institute for Materials Science and Engineering

Conducts research and provides measurements, data, standards, reference Ceramics
^
materials, quantitative understanding and other technical information funda- Fracture and Deformation
mental to the processing, structure, properties and performance of materials; Polymers
addresses the scientific basis for new advanced materials technologies; plans Metallurgy
research around cross-country scientific themes such as nondestructive Reactor Radiation
evaluation and phase diagram development; oversees Bureau-wide technical
programs in nuclear reactor radiation research and nondestructive evalua-
tion; and broadly disseminates generic technical information resulting from
its programs. The Institute consists of the following Divisions:

'Headquarters and Laboratories at Gaithersburg, MD, unless otherwise noted; mailing address
Gaithersburg, MD
20899.
^Some divisions within the center are located at Boulder, CO 80303.
'Located at Boulder, CO, with some elements at Gaithersburg, MD.
 RESEAECH nTFOEJflATION
CEIITER
Hierniometer Calibration:

A IVIodel for State Calibration Laboratories

/Oo

no.

Jacquelyn A. Wise
Robert J. Soulen, Jr.

Temperature and Pressure Division

National Measurement Laboratory
National Bureau of Standards
Gaithersburg, MD
20899

o
4r

U.S. DEPARTMENT OF COMMERCE, Malcolm Baldrige, Secretary

NATIONAL BUREAU OF STANDARDS, Ernest Ambler, Director

Issued January 1986

 Library of Congress Catalog Card Number: 85-600636
National Bureau of Standards Monograph 174
Natl. Bur. Stand. (U.S.), Monogr. 174, 88 pages (Jan. 1986)

CODEN: NBSMA6

US. Government Printing Offi<ce

Washington: 1986
For sale by the Superintendent of Documents, U.S. Government Printing Office, Washington, DC 20402
 TABLE OF CONTENTS

Page
1. INTRODUCTION 1

2. LABORATORY AND EQUIPMENT 1

2.1 Laboratory Space and Environment 1

2.2 Temperature Baths 6

2.2.1 Ice bath 6

2.2.2 Variable, constant-temperature baths ... 8

2.3 Laboratory Standards 11

2.3.1 Primary standards 11

2.3.2 Check standards 13

2.4 Personnel 13

3. PERFORMING CUSTOMER'S CALIBRATIONS 14

3.1 Examination and Preparation of Thermometers

for Calibration 14

3.2 Proper Method for Calibrating Specific

Thermometers 15

3.2.1 Total-immersion thermometers 15

3.2.2 Partial-immersion thermometers 15

3.2.3 ASTM thermometers 16

3.2.4 Calorimetric thermometers 16

3.2.5 Beckmann thermometers 17

3.2.6 Clinical standard thermometers 17

3.2.7 Kinematic viscosity thermometers 17

3.2.8 Electronic temperature measuring

systems 18

iii
 Page

3.3 Calibration Procedure 18

3.3.1 Number and choice of calibration

points 18

3.3.2 Use of the ice-point bath 18

3.3.3 Steam point (100 °C) as reference

point 19

3.3.4 Role of primary liquid-in-glass

thermometers 20

3.3.5 Routine adjustment to primary

standard thermometer calibration
data 20

3.3.6 Placement of thermometers in

calibration bath 24

3.3.7 Reading sequence and recording data ... 25

3.3.8 Computation of bath temperature from

readings of primary standard
thermometers 27

3.3.9 Determining corrections for

thermometers being calibrated 30

3.3.10 Stem-temperature corrections 31

4. DOCUMENTATION 36

4.1 Report of Calibration 36

4.2 Report of Test 39

5. MAINTAINING LABORATORY STANDARDS AND ESTIMATION OF

LABORATORY UNCERTAINTY 46

5.1 Estimates of Inaccuracy of Calibration of

Liqu id- in-G lass Thermometers 46

5.2 Maintaining the Accuracy of Primary

Laboratory Standards 46

5.3 Using Check Standards for Quality Assurance ... 47

iv
 Page

6. CONCLUSION . . .' 51

7. ACKNOWLEDGMENTS 54

8. REFERENCES 54

9. COMMERCIAL SOURCES 55

9.1 Suppliers of Calibration Baths 55

9.2 Suppliers of Oils for Calibration Baths 55

9.3 Suppliers of Thermometers 56

9.4 Suppliers of Faden Thermometers 56

10. APPENDIX 57

Liquid-In-Glass Thermometry, NBS Monograph 150 57

V
1. INTRODUCTION

This publication is written primarily to assist State

calibration laboratories to establish temperature calibration
capabilities. It can also be used by any industrial organization
or independent testing laboratory that wishes to establish a
similar facility.
The range of the calibrations to be conducted extends
roughly from 0 to 200 ^C, while the target accuracy is
±0.1 ^C. Several types of thermometers could be used to maintain
this stated range and accuracy, but liquid- in-glass thermometers
have been chosen because they are dependable, easy to use and
maintain, and economical.
Presented in this publication are detailed descriptions of
the laboratory environment, equipment, type of training and
personnel, calibration procedure, calculations, data analysis,
and documentation needed for the calibration of thermometers.
These descriptions and procedures are based on experience
acquired at the National Bureau of Standards (NBS) for over 50
years.
If a is established and maintained as described
laboratory
in liquid-in-glass thermometers and electronic
this publication,
temperature measuring systems can be calibrated to an accuracy
somewhat better than 0.1 over a range of 0 to 200 ^C, and they
will be traceable to NBS. A deviation from the procedures or
specifications given may be acceptable, but any consequent errors
should be determined and evaluated for acceptability by the
manager of the laboratory. If laboratories follow these proce-
dures, there will be more uniformity among these laboratories and
this will result in better service to industry and to the public.
A checklist for establishing a calibration laboratory to the
specifications described in this publication is provided in
Section 6 (see Table 10).

2. LABORATORY AND EQUIPMENT

2.1 Laboratory Space and Environment

Liquid-in-glass thermometers submitted for calibration may

be as long as 615 mm. They may be accidentally broken if an
adequate distance is not kept between laboratory furnishings.
Futhermore, any electronic temperature measuring systems
submitted for calibration will require additional room around the
calibration bath to place the digital readout device or other
electronic components. The room chosen for the calibration lab-
oratory should therefore be spacious; dimensions of approximately
7mx7mx3m are adequate. A photograph of one room of the
liquid-in-glass thermometer calibration laboratory at the NBS is
shown in Figure 1 as an example. A floor plan of the entire
facility is shown in Figures 2 and 3.

1
 4J
to

u
O
4J
fO
1-1

o
(0

c
0
•H
+J
fO
i-l

XI
•H
H
<0
u
u
Q)

<D
•
x:

en fO

.H c

c w
•H

0
•H

•H (D
iJ
3
CQ

to

s-< O

•1-4 fO
fa 2;

2
 Counter and Storage Area Counter Sink

Exhaust Hood

on
Bath

Sand Blaster
and
Thermometer Pick-up
Ice
Shaver

Figure 3. Floor plan of higher-temperature

calibration room and room used for ice
preparation and shipping of thermometers.

The laboratory should be maintained at a constant

temperature and humidity. Many thermometers that will be
submitted for calibration will be intended for use at partial
immersion, in which case it is important that the emergent stem
remain at a constant temperature while being tested. For ex-
ample, if there is a one-degree change in the laboratory
temperature, an error of approximately 0.01 can result in the
calibration of a partial-immersion thermometer with 50 degrees
emergent from the bath. The humidity should be maintained near
50% to keep any electronic components that may be in the
laboratory from being adversely affected.
Any calibration bath that uses an oil as the medium should
be operated under a hood. The hood should not be connected to
the ventilation system of the building; instead it should be
exhausted directly outside. Any filters in the system should be
periodically checked and changed if necessary. Figure 4 shows
the exhaust hood over the oil baths at NBS.
There must be a place in the laboratory to safely store the
primary liquid-in-glass thermometer standards, as well as those
submitted for calibration. A map case, having corrugated paper
placed on the bottom of each drawer to keep the thermometers from
rolling into each other when the drawer is opened, can be used
for this purpose (see Figure 5). Another method commonly used is
to store the thermometers vertically in a thermometer rack
especially designed for this purpose. This method is shown in
Figure 6.

4
Figure 4. Exhaust hood over oil baths.
 Figure 6. Thermometers placed in
thermometer rack.

2.2 Temperature Baths

2.2.1 Ice bath

An ice bath (see Figure 7) is perhaps the most valuable and

essential item in a thermometer calibration laboratory. This is
because all liquid-in-glass thermometers suffer changes in bulb
volume (due to thermal expansion and other factors). The method
for correcting for such changes involves recal ibr at ion in most
cases at the ice point (0 ^c) .(See Section 5.2, NBS Monograph
150 [I]-'-, and a discussion later in this Monograph.)
A typical ice bath consists of a Dewar flask approximately
36 cm deep and 8 cm in diameter. The outside of the flask should
be covered with electrical or waterproof adhesive tape and en-
closed in a metal sleeve to protect the operator in the event
that it should break. The flask should be supported by a struc-
ture that will not allow it to tip over easily. An appropriate

-•-Figures in brackets indicate references at the end of this

Monograph.

6
 .

Figure 7. Ice bath at NBS

clip is placed above the top of the flask to hold the thermometer
perpendicular to the telescope used for reading. A siphon is
placed in the bottom of the flask to remove excess water that
will form due to the melting of the ice during the day. A small
soft mallet, which is used to gently tap the thermometers before
reading, should be available.
Shaved ice (of quality described below) and distilled water
are placed in the Dewar flask and allowed to reach a constant
temperature, a process that takes approximately 15 to 30 minutes.
After this time, any excess water should be siphoned off and
additional ice added. When the ice bath is ready for use, there
should be no floating ice in the flask and no excess water on the
surface. There should be as much ice as can be forced into the
flask and sufficient distilled water to fill the small crevices
between the ice chips. Failure to follow this procedure can lead
to vertical or radial temperature gradients, which can result in
a calibration error as large as 0.01 at the ice point.
Whenever the ice bath is used, care should be taken not to
contaminate the ice or water. Thermometers should be rinsed with
cool distilled water before being placed in the ice and plastic
gloves should be worn by the operator or hands washed often while
taking ice points. It is important that impurities, such as salt
from the hands, not be allowed to come in contact with the ice.
This will cause a depression of the freezing point and result in
an error in subsequent ice-point determinations.
From what has been stated in this section, it is clear that
a calibration laboratory will need an adequate supply of
distilled water and pure ice (see below). A still and storage
tank can be purchased for this purpose. The distilled water that
is produced can be frozen to form blocks of distilled water ice.
The freezer and containers used for this purpose should not be
used for anything else and kept as clean and free from impurities
as possible. The containers should be washed thoroughly with
soap and water and rinsed two or three times with distilled water
before being used initially.
Ice purchased commercially can also be used if it is crystal
clear. Some companies produce large blocks of ice by freezing
the outside first. When ice is made in this manner, the
impurities are forced to the center where the ice will be cloudy
and white. This unclear portion is removed and only the crystal-
clear portion is used.
An ice shaver, which can be purchased from restaurant supply
companies, is used to shave the ice to the proper consistency.
The ice chips should be approximately 2 to 5 mm in diameter.
These particles are smaller than those produced by crushing or
chipping the ice.

2.2.2 Variable, constant-temperature baths

In addition to the ice bath, which provides a calibration at

a single temperature, it will be necessary for the laboratory to
provide calibrations at several other temperatures from +1 to
+200 oc. This requires the use of temperature baths (see
Section 9.1 for sources) in which the temperature can be varied
and then set to constant values for moderately long periods of
time to permit calibrations. Spanning this temperature range
requires, in general, two baths using different working media
(water and oil). The water bath used at NBS is shown in
Figure 8,
For calibration in the range of approximately +1 to +95 ^C,
water can be used as the bath medium. Distilled water is
preferable, since no chemical deposits will form at the top of
the bath. Water is plentiful, clean, and does not produce
harmful vapors. In the range of approximately +95 to +200 ^C, an
oil should be used as the medium. Manufacturers of the calibra-
tion baths will generally recommend the type of oil that should
be used in their baths and list suppliers (see Section 9.2 for
sources). It is important that an exhaust hood be placed over
the oil bath to remove vapors from the laboratory and that the
oil is not used near its flash point. The division of the
temperature range into two regions using two different media is
not always required. In particular situations a laboratory might

8
 Figure 8. Variable, constant-temperature
water bath.

find it more convenient to use a single oil bath to perform

calibrations from +1 °C to +200 °C.
A cover should be placed over any openings of the bath to
prevent foreign material from entering the medium that might
cause contamination or obstruct stirring.
Several factors must be taken into account in the decision
for the purchase or construction or such baths. The most
critical ones are: volume, temperature homogeneity, and
temperature stability.
Constant-temperature baths may be purchased from any of
several companies or they can be made at the calibration labora-
tory. Before the calibration bath is purchased or designed, one
must determine what types of thermometers will be calibrated. If
calibrations will include long, total-immersion liquid-in-glass
thermometers, the depth of the bath must be sufficient to accom-
modate them. If several thermometers are to be tested at one
time, the diameter at the top of the bath must be large enough to
accept a holder containing several thermometers. The bath should
have a provision for overflow or removal of the expanding fluid

9
that will result as the temperature of the medium is increased.
It should also be easy to add fluid so that the liquid level can
be maintained within 6 mm to 12 mm of the bottom of the
thermometer holder.
A constant-temperature bath should have the heating and
cooling coils isolated from the volume where the thermometers
will be placed (see Figure 5 in NBS Monograph 150 [1]). There
should be an unrestricted path of flow for the bath medium, which
is stirred or pumped at a sufficient rate to maintain a uniform
temperature throughout the medium. There should be adequate
insulation around the bath. The holder for the thermometers
should be designed so as to prevent heat loss.
Every bath should be checked to determine the magnitude of
the horizontal and vertical temperature gradients. One way of
doing this is to regulate the bath sequentially at several
temperatures throughout the range. At each temperature a
thermometer is placed in one position and constantly read to
determine if the temperature of the bath is changing. A second
thermometer, either another liquid-in-glass type or preferably a
faster responding one, such as an electronic digital readout
using a thermistor or small resistance device as the sensor, is
placed next to the first one. When it is observed that the
thermometer readings are not changing or are increasing at a rate
of no more than 0.01 to 0.05 in 5 minutes, stability can be
considered to have been reached. At this time both thermometers
are read and the readings are recorded. While the stationary
thermometer is monitored, the second thermometer is moved to
other locations and depths in the bath and readings on both
thermometers are recorded. If a liquid-in-glass thermometer is
used to probe the bath, adequate time must be given for it to
stabilize, since the response time may be slower than that of an
electronic thermometer. After adjusting the reading of the
second thermometer for any change that may be noted by the
stationary thermometer, the change in the second thermometer
reading at various locations will indicate the uniformity of the
bath temperature. The accuracy of either thermometer is not
important for this test, only its temperature sensitivity.
A common cause of bath temperature nonun i f or m i ty is
inadequate circulation of the fluid. If this is suspected to be
the case, it is necessary to increase the circulation rate and
recheck the temperature gradients. On the other hand,
circulation rates cannot be increased beyond the point where
overheating and cavitation of the bath medium will occur
(observable by the generation of temperature fluctuations). It
is best to determine optimum rates experimentally.
A second problem can occur when the bath is used at elevated
temperatures. In this situation, heat escapes from the top of
the fluid, causing a vertical gradient in the upper 25-50 mm
region of the fluid. Clearly, for accurate results, no part of
the bulb or contraction chamber of a liquid-in-glass thermometer
or the sensing portion of an electronic thermometer should be in

10
 .

this region. When the bath is first installed, this gradient

should be determined as a function of temperature.
Attached to all calibration baths, including the ice bath,
should be a telescope mounted perpendicular to the axis of the
thermometers that will be placed in the holder. If the telescope
cannot be mounted on the bath, it can be placed on a stand near
the bath, but care must be taken to see that it is always set
perpendicular to the thermometer axis to avoid reading errors due
to parallax. The telescope should be approximately 10 power and
it is desirable to have a cross-hair to aid in reading the
meniscus within the scale divisions.

2.3 Laboratory Standards

2.3.1 Primary standards

The laboratory should be equipped with two sets of "primary"

standards and one set of "check" standards. The primary stan-
dards should be calibrated versus the International Practical
Temperature Scale of 1968 (IPTS-68). This temperature scale is
the latest version of a series of international temperature
scales that have been developed by national standards laborato-
ries like the NBS throughout the world. A complete description
of this scale is given in the Metrologia article entitled, "The
International Practical Temperature Scale of 1968 Amended Edition
of 1975" [2]
The inaccuracy of IPTS-68 (that is, the best estimate of the
difference between the scale and the correct temperature) in the
region 0 to 200 is estimated to be ±0.04 It is possible
to register temperature in this region with an imprecision much
smaller than the inaccuracy of the temperature scale. For
example, platinum resistance thermometers can be used to
repeatedly register the freezing point of Gallium (near 29 ^C)
with a standard deviation of only 0.0001 (the imprecision) yet
the temperature assigned to the Gallium point on the IPTS-68 will
still remain inaccurate to ±0.04 until a better temperature
scale is developed. When such a temperature scale is developed
(approximately every 10 or 20 years; last done in 1968) and com-
pared to IPTS-68, a new difference table will be constructed. At
that time at each temperature, a different systematic correction
(or offset) must be applied.
By comparison to platinum resistance thermometers, the im-
precision of 1 iqu id - in-g la ss thermometers under the best
conditions has been found to be ±0.03 °C (see Section 5.3).
Thus, a State calibration laboratory can maintain a temperature
scale based on the IPTS-68 using liquid-in-glass thermometers and
the procedures given in this Monograph in which the inaccuracy
and imprecision are comparable. Using liquid-in-glass
thermometers, the laboratory will be able to provide calibrations
with errors of about one-third the tolerance permitted in
customers' thermometers.

11
 .

Since total-immersion thermometers (which have virtually no

mercury-containing stem exposed to ambient temperature) are more
accurate than partial-immersion thermometers, the former should
be chosen as standards. Given in Table 1 are ranges, graduation
intervals, total length, and imprecision of thermometers that are
recommended for use as primary standards. Commercial suppliers
of these thermometers are given in Section 9.3. The reproduci-
bility and accuracy of the thermometers chosen for standards
should be better than or equal to that of the thermometers being
calibrated.

TABLE 1

PRIMARY LIQUID- IN-GLASS THERMOMETERS

ASIM^ Thenrcmeters[3]
Inprecision at
TheriTDneter Graduation Calibration It>ints
Asm No. Range Interval Tbtal Length After NBS Calibration

63C -8 to +32 °C 0.1 °C 374-384 irni 0.03

64C +25 to +55 °C 0.1 °C 374-384 mm 0.03 ^

with auxiliary scale at 0*^

65C +50 to +80 °C 0.1 °C 374-384 mm 0.03 °C

with auxiliary scale at 0°C

66C +75 to +105 °C 0.1 °C 374-384 mn 0.03 °C

with auxiliary scale at 0°C

67C +95 to +155 °C 0.2 °C 374-384 mn 0.05 °C

with auxiliary scale at 0 °C

68C +145 to +205 °C 0.2 °C 374-384 mm 0.1 °C

with auxiliary scale at 0 *-t

ANSI^ Thermometers [4]

Imprecision at
Thermometer Graduat icn Calibration Itoints
SftMA'^ No. Range Interval Tbtal Length After NBS Calibration

CT40 -1 to +51 °C 0.1 oc 460 tim 0.03

CT45 -1 to +101 °C 0.1 °C 610 nm 0.03 °C

CT50 -1 to +201 °C 0.2 "^t 610 mm 0.05 °C \jp to 100 °C

0.1 °C above 100 °C

^ASM acronym for the American Society for Ttesting and Materials.
is an
°MSI an acronym for the American National Standards Institute.
is
^SAMA is an acronym for the Scientific /^aratus Makers Association.
"For imprecision between the calibration points see Section 5.4 of NBS Monograph 150 [1]

These thermometers must be free of foreign material in the

capillary (such as glass chips or oxides of mercury), be numbered
correctly (in proper numerical sequence), and be uniquely

12
identified by a serial number. The scale must be placed on the
thermometer according to NBS specifications (see Section 7.3, NBS
Monograph 150 [1]), which do not permit it to be graduated close
to an enlargement in the capillary. The thermometers must meet
all other requirements given in NBS Monograph 150 [1], such as
being made of the proper glass and giving temperature indications
within the specified tolerances. When ordering liquid-in-glass
thermometers that will be used as standards, it is advisable to
request the company to guarantee in addition that they will pass
the NBS visual inspection. This service may increase the initial
cost of the thermometer, but it will be replaced at no additional
cost in the unlikely event that it is found ineligible for test
by NBS.
To equip the laboratory with adequate primary thermometers,
two liquid-in-glass thermometers from each range should be
purchased from the choices given in Table 1. Then these thermom-
eters should be sent for calibration to the NBS or another
competent laboratory, accompanied by a letter or purchase order
requesting calibration. In the former case, the thermometers
will be calibrated by NBS personnel and returned to the sending
laboratory along with Reports of Calibration, which list the
corrections to be made at each calibration point. By virtue of
their calibration at the NBS, such standard thermometers will
automatically bear the IPTS-68 scale.

2.3.2 Check standards

The third set of liquid-in-glass thermometers, also chosen

from the list given in Table 1, should be purchased for use as
check standards. This set differs from the primary standards in
that it is not sent to NBS for calibration. Rather, the appro-
priate thermometer from the set is incorporated as one of the
thermometers in the calibration sequence (to be described in
Section 3.3.7). By treating the check standard as one of the
thermometers being calibrated in terms of the primary standards,
it is used to check the calibration procedure, helping to insure
that the calibration was done properly. A check standard may also
be used as temporary backup for a primary standard in the event
that one of a primary thermometers is broken. More than one
check standard in each temperature range may be needed due to
bulb volume expansion (see Section 3.3.5).

2.4 Personnel

The staffing level as well as the experience and education

needed by the calibration staff will depend on several factors
that are specific to the laboratory: budgeting, calibration in-
come, calibration volume, and turnaround time. Thus a general
prescription here would not be especially useful. For the pur-
pose of reference, however, we mention the staffing situation at
the NBS. The Liquid-in-Glass Thermometer Calibration Laboratory
at the NBS is staffed by one individual who has had 3 years of

13
college education and 20 years experience in the laboratory.
Approximately 800 liquid-in-glass thermometers and a small number
of other types of thermometers are calibrated per year in this
laboratory. An individual chosen to be employed in a thermometer
calibration laboratory should be adept in mathematics, a stickler
for details, and always striving for perfection. Depending on
the initiative of the individual, we estimate that it would take
12 months of practice and experience for a high school graduate
to operate a calibration laboratory as described in this
publication. We strongly recommend that an individual chosen for
this responsibility attend a session of the NBS Precision
Thermometry Seminar which is given twice yearly. In the Seminar,
participants are given lectures on all common types of
calibration including liquid-in-glass thermometers, and the
attendees also perform simulated calibrations at the NBS
facility.
To summarize: when the laboratory has been properly
equipped and staffed, and when NBS-cal ibr a t ed primary
thermometers have been incorporated into a careful measurement
program, then the State calibration laboratory can be considered
to be ready to perform calibrations, based on the IPTS-68, for
its customers. We now turn to a discussion of the procedures to
be followed in performing those calibrations.

3. PERFORMING CUSTOMER'S CALIBRATIONS

3.1 Examination and Preparation of Thermometers for Calibration
Every thermometer submitted for calibration must be examined
under a microscope (preferably 15 or 20 power) for defects. One
reason for this visual inspection is to detect a separated mer-
cury column or small droplets of mercury that may be at the top
of the thermometer. This condition can occur because of rough
handling encountered during shipping. All of the separated mer-
cury must be reunited with the main column before the thermometer
can be calibrated.
Rough handling during shipping can also cause gas bubbles
to form in the bulb, a defect that also can be detected during
the visual examination. These gas bubbles will expand at a
greater rate than the mercury that should be in these spaces;
therefore, they must be removed. To reunite separated mercury
and to remove gas bubbles in the bulb, see Section 8.4 of NBS
Monograph 150 [1]
The third reason for examining each thermometer under a
microscope is to look for foreign material in the capillary and
defects in the thermometer. The foreign materials most often
found are glass chips or oxides of mercury. These objects can
hold gas or mercury and cause the reading of the thermometer to
vary when the thermometer is repeatedly inserted into a regulated
constant-temperature bath. The capillary of the thermometer must
be clean and free of any foreign material, and it must also

14
appear to have a uniform diameter. If any variation in the
diameter of the capillary can be seen under the microscope, the
thermometer is not acceptable for calibration. The distance
between scale graduations must be uniform, the scale must be
placed on the thermometer according to NBS specifications (see
Section 7.3 in NBS Monograph 150 [1]), and the thermometer must
be numbered correctly and be free of cracks in the glass. If any
of the above conditions are not met in a thermometer that has
been submitted for calibration, it should be ineligible for test
and should be replaced by the customer.

3.2 Proper Method for Calibrating Specific Thermometers

Thermometers usually are calibrated according to their

types. A particular thermometer should be calibrated differently
only if the customer specifically requests it. In this section
is presented a discussion on the calibration of several types of
thermometers.

3.2.1 Total-immersion thermometers

Total-immersion thermometers are calibrated with the bulb
and all but approximately 6 mm to 12 mm of the mercury column
immersed in the calibration bath. The upper 6 mm to 12 mm of the
mercury column extends out of the bath surface to permit reading.
The meniscus (the top of the mercury column) should never be
immersed in the bath fluid, since distillation of the mercury can
occur at high temperatures and cause mercury droplets to condense
at the top of the thermometer.
After remaining at room temperature (23 ^C) for 3 days,
these thermometers are calibrated starting at the lowest tempera-
ture requested and proceeding in sequence to the highest point.
At each calibration point the thermometer must be pushed deeper
into the calibration bath to maintain the proper immersion.
If the thermometer has a contraction chamber (see Section 3
of NBS Monograph 150 [1]), it must be immersed well below the
surface of the bath medium. The contraction chamber contains a
great deal of mercury and this mercury must be at the same tem-
perature as the mercury in the bulb for the thermometer to give a
correct reading. If there is a temperature gradient at the
surface of the bath, as may be the case in high temperature
calibration baths, the contraction chamber must be placed below
this gradient, even if this means changing the calibration point.
If the thermometer being calibrated is too long and cannot
be immersed at the proper depth, a stem-temperature correction
will have to be made as described in Section 5.3 of NBS
Monograph 150 [1] and in Section 3.3.10 of this Monograph.
3.2.2 Partial-immersion thermometers
Partial-immersion thermometers must have either an immersion
line or the proper depth of immersion engraved or printed on the

15
back (most thermometers have both). These thermometers are
immersed only to the immersion line or specified depth. The
portion of the stem that contains mercury above the immersion
line is known as the "emergent stem". It tends to reach ambient
temperature or the temperature prevailing above the calibration
bath. If the emergent stem is to have a specified stem
temperature, as is the case with all ASTM partial-immersion
thermometers, a stem temperature correction must be made as
described in Section 3.3.10,
Like total-immersion thermometers, calibration points are
taken in the same sequence; the thermometer must remain at room
temperature (23 ^C) for 3 days before calibrating; and a
contraction chamber must be immersed properly.

3.2.3 ASTM thermometers

Over 120 different thermometer specifications are listed in

ASTM Standard E-1, Standard Specification for ASTM Ther-
mometers [3]. The thermometers were designed to be used in
specific equipment and test methods and are identified by having
the acronym "ASTM" and a number written on the back. Specifi-
cation E-1 lists the calibration points for all ASTM thermom-
eters, stem temperatures for ASTM partial-immersion thermometers,
and any special procedure that must be followed. An ASTM ther-
mometer should be tested in accordance with the procedure out-
lined in E-1 and E-77, Standard Method for Verification and
Calibration of Liquid-in-Glass Thermometers [5].

3.2.4 Ca lor ime trie thermometers

In some cases it is only necessary to accurately measure a

temperature difference rather than the absolute temperature.
Calorimetric thermometers were designed for this purpose. They
usually are 600 mm long, cover a short range of temperature
(typically 6 to 16 degrees), and are graduated in intervals of
0.01 to 0.05 degree. The thermometers are usually calibrated
every 50 to 100 scale divisions (in intervals of 0.5 or
1 degree). Because of the very small capillary diameter, they
must be tapped lightly at each reading to eliminate a possible
error due to a sticking mercury column, and they must be heated
to the maximum calibration point before testing. Initially
heating to the top point causes the bulb to expand to the maximum
volume and eliminates the error that will occur because of the
gradual expansion of the bulb as the thermometer is heated. In
order to register accurate temperature differences, the bulb must
be expanded before calibration and before use. The thermometer
should be calibrated within 2 to 3 hours after it has been
heated.
These thermometers are capable of measuring temperature
differences with uncertainties of 0.01 ^C, which will be beyond
the capabilities of a State calibration laboratory following the
procedures outlined in this publication. If a customer requests

16
this accuracy, the thermometer must be sent to another
calibration laboratory, such as NBS.

3.2.5 Beckmann thermometers

These thermometers are instruments in which a narrow

temperature interval (typically 6 oc) may be adjusted to fall
within 0 to 100 oc. The graduations are 0.01 oc and the
thermometer is of the enc losed-scale type, which consists of a
capillary tube attached to a scale and enclosed in a glass tube.
Since the narrow interval is adjustable, these thermometers are
to be used only to register temperature differences.
The interval may be moved within the range 0 to 100 oc by
adding or removing mercury to or from the bulb. Consider the
following example for adjusting a Beckmann thermometer to
indicate temperature differences from 20 to 26 o^. Before
calibration, the amount of mercury in the bulb is adjusted to
place the meniscus within ±5 scale divisions of the "zero"
indication when the thermometer is immersed in a medium having a
temperature of 20 oc. This is referred to as a "setting of
20 oc". The thermometer is then calibrated in the range of 20 to
26 °C. For other settings, the temperature difference must be
multiplied by the appropriate setting factor as shown in
Section 6.4 of NBS Monograph 150 [1].
The Beckmann thermometer is a calorimetric thermometer and
must be heated to the highest test point, tapped gently before
reading, and calibrated every 0.5 ^C. They are capable of
measuring temperature differences with uncertainties of 0.005
and, like the calorimetric thermometers, will have to be sent to
another calibration laboratory, such as NBS, if this accuracy is
desired.

3.2.6 Clinical standard thermometers

Clinical standard thermometers are used by the manufacturers

of fever thermometers, which are used for measuring body tempera-
ture. In the United States they are marked in degrees Fahrenheit
with a range of approximately +90 to +112 of. They are provided
with an auxiliary scale at 32 ^F and graduated in intervals of
0.1 Of. The calibration points are usually 32, 98, 102, and
106 °F and they are heated to 108 ^F before calibrating to
initially expand the bulbs. Since they are heated to 108 °F
before calibrating, the ice points are taken after test. They
are generally tested at total immersion and at partial immersion,
immersed to approximately the 92 °F mark.
3.2.7 Kinematic viscosity thermometers

When measuring the viscosity of liquids, it is imperative to

accurately measure one specific temperature. Kinematic viscosity
thermometers are designed for this purpose and have a short 3 to
5 degree range with an auxiliary scale for an ice point. These

17
thermometers are calibrated at two points on the main scale. Ice
points are measured afterwards.

3.2.8 Electronic temperature measuring systems

A thermometry calibration laboratory may occasionally be

requested to calibrate an electronic temperature measuring
system. It usually consists of a digital readout connected to a
probe, which can be a thermocouple, thermistor, or resistor (most
commonly an industrial platinum resistance thermometer). It must
first be determined that the probe can physically be placed in
the calibration bath with enough lead wire to enable the readout
unit to be located near the bath. The probe must be long enough
so that the sensing element is below the region of large tempera-
ture gradient that may occur at the top of the bath. If the
probe cannot be placed directly in the fluid, it must be put in a
tight fitting, protective tube to provide good thermal con-
ductivity. The electronic system is calibrated in the same
manner and readings taken in the same sequence as for liquid-in-
glass thermometers.

3.3 Calibration Procedure

3.3.1 Number and choice of calibration points

A thermometer submitted for test should be calibrated at

several points over at least 80% of the main scale.
The thermometer should be calibrated at the temperatures
specified by the customer or at intervals of approximately every
100 scale divisions. Calibration points taken at no more than
every 100 scale divisions will give acceptable interpolated cor-
rections for temperatures between the calibration points. (See
Section 5.4 of NBS Monograph 150 [1].)
The thermometer should be calibrated at the reference point,
usually 0 or 100 ^C, if there is one on the thermometer. The
reference point is used to recalibrate the thermometer as
described in Section 5.2 of NBS Monograph 150 [1] and in Sections
3.3.5 and 3.3.8 of this Monograph.

3.3.2 Use of ice-point bath

Make a fresh ice bath on the day that a calibration is to be

performed (discarding any previous day's bath). As ice melts
throughout the day, siphon off the water and add new ice. The
bath is ready for use when no ice is floating and no excess water
is on the surface of the ice.
For proper calibration at 0 ^C, each liquid-in-glass
thermometer should be inserted to a specified depth (the
immersion depth) into the ice-point bath. Total-immersion
thermometers should be inserted into the ice-point bath until the
0 °C mark on the thermometer scale is within one scale division
of the top of the bath. For partial-immersion thermometers, the

18
 .

depth is indicated either by an additional line scribed on the

glass (immersion line) or by a number given on the back of the
thermome ter
In order to prepare the bath to accept the thermometer,
loosen an area of ice at the center of the bath with an object
such as a clean glass rod to a depth approximately equal to the
immersion depth. Clean the thermometer to be calibrated at the
ice point by rinsing it with cool distilled water. (Never rinse
with warm water, since this may expand the bulb.) Then gently
place the thermometer through the holder and into the region of
loosened ice. If the thermometer touches a firm surface before
it is immersed to the immersion line or 0 oc mark, then remove
the thermometer and loosen the ice further down into the bath to
permit the thermometer to be immersed to the proper depth. If
the immersion line or 0 °C mark on the thermometer passes below
the surface of the bath before resting on a firm foundation,
repack the ice and loosen the ice to the correct depth.
After the thermometer is properly immersed and perpendicular
to the observer's telescope, firmly pack the ice around the
thermometer. It is important to keep the ice clean while using
the ice bath. The operator's hands must frequently be washed
(with soap and water) or covered with plastic gloves. When the
thermometer is resting on a solid section of ice in the bath and
cannot be immersed further, when it is perpendicular to the
telescope, when ice is firmly packed to the immersion line or one
scale division below the 0 °C mark, and when any excess water is
siphoned from the ice bath, the thermometer is ready to be read
as shown in Figure 9.
The thermometer should remain in the ice bath for approxi-
mately one to two minutes. (Thermometers using an organic fluid
instead of mercury as the liquid may require approximately 15
minutes because the organic fluid tends to cling to the wall of
the capillary.) When stability is reached (the meniscus stops
moving), gently tap the thermometer to free the mercury meniscus
(which can sometimes stick) and record the ice-point reading.

3.3.3 Steam point (100 ^C) as reference point

Although the ice point at 0 °C is most often the reference

point, the steam point at 100 °C is sometimes used. A steam-
point bath or hypsometer, as described in Section 5.1b of NBS
Monograph 150 [1], is the fixed-point apparatus sometimes used at
NBS to realize this temperature. Since the temperature is
dependent on the atmospheric pressure, a barometer accurate to
0.1 mm of mercury must also be available.
It is not necessary to purchase a steam bath or barometer if
the calibration laboratory has standard thermometers and an oil
bath capable of being controlled at 100 °C. A thermometer having
a reference point at 100 can be calibrated at this temperature
in the oil bath. The change in bulb volume can be determined by
subsequent recalibra tion at 100 ^C.

19
 Figure 9. Thermometer properly placed in ice
bath.

3.3.4 Role of primary liquid-in-glass thermometers

Precision liqu id-in-glass thermometers that have been cali-

brated by NBS or another calibration laboratory are used to
realize the current International Practical Temperature
Scale [2] .

When liquid-in-glass thermometers are used as primary

standards, two of them should be used at each calibration point.
Each thermometer is corrected to obtain the accurate temperature
of the bath. The corrected readings of the two primary standards
must agree within a specified tolerance before the calibration
point is acceptable. NBS has found that 0.2 of a scale division
is an achievable tolerance. Using two primary standards usually
will reveal an error due to an incorrect reading of one of the
primary standards. Averaging the carefully taken readings of two
primary standard thermometer temperatures will give a more
reliable value of the temperature of the calibration bath medium,
especially at temperatures that differ from those where the
primary standard thermometers were calibrated.

3.3.5 Routine adjustment to primary standard thermometer

calibration data

There are three conditions that will prevent each primary

1 iqu id- in-g la ss thermometer standard used by the State

20
calibration laboratory personnel from indicating temperatures
correctly. The first is permanent irregularities in the
thermometer, such as nonuniform bore and inaccurate scale
markings. This correction need only be obtained once when the
thermometer is sent to a national standards laboratory such as
NBS. Here the readings of the thermometer are compared with the
temperatures determined from a precision standard platinum
resistance thermometer that has been calibrated on the current
IPTS. On the basis of these measurements, a Report of
Calibration is issued, which gives a table of corrections for
these permanent errors at the calibration points resident on this
thermometer.
The other two conditions, caused by long-term or short-term
changes in the bulb glass, must be corrected by personnel in the
State calibration laboratory on a routine basis according to a
procedure to be described later in this section. Briefly, glass,
as a supercooled liquid, will slowly flow, and thus on a long-
term basis (over the life of the thermometer) the volume of the
bulb will change, causing the readings of the thermometer to
change. This is easily detected and can be corrected by
determining the correction at the ice point and shifting
uniformly the corrections given on the Report of Calibration (see
Section 5.2, NBS Monograph 150 [1]).
There is also a short-term (three days or less) change in
the bulb glass. To understand this, we appreciate the well known
fact that when a liquid-in-glass thermometer is heated to higher
temperatures, the glass as well as the mercury expands. This
expansion of the bulb, which has a greater effect on the
thermometer readings than changes in the stem glass, can
correspond to a change in temperature reading as great as
0.01 oc for every 10 degrees that it is heated. As the
thermometer is cooled (e.g., to the ice point), the mercury will
return to its original volume, but the bulb has been found to
take approximately 3 days to return to its original volume.
During this three-day period the thermometer will indicate that
the temperature is lower than it actually is, because the
expanded glass bulb volume causes the mercury column to fall
lower in the stem. In order to obtain a correct temperature
value, one must compensate for this temporary, short-term change
in the bulb volume known as an "ice-point depression".
To illustrate this procedure, we provide the following
example. Consider a primary thermometer that, for simplicity, we
assume has no error in the graduations, i.e., all corrections on
the Report of Calibration, which were obtained after the
thermometer had been at room temperature (23 OC) for three days
and calibrated from the lowest calibration point to the highest,
are zero. (See column 2 of Table 2.) We conduct the following
measurements: First, a measurement at 0 ^C, taken after the
thermometer had been at room temperature for not less then three
days, shows no correction. The thermometer then is heated to a
given temperature, (e.g., 50 ^C) and then returned within a few
minutes to the ice bath. The volume of the thermometer bulb has

21
been increased by virtue of the heating to 50 so the mercury
will not return to the original height in the glass tube and the
reading at 0 will be lower (e.g., -0.05 OC) We now cycle the
.

thermometer to 100 and note the new reading at the ice point
(e.g., -0.10 OC) and then cycle to 150 and back to 0 for a
fourth ice-point reading (e.g., -0.15 ^C) The data thus obtained
.

for this thermometer may be represented by Column 3 of Table 2.

TABLE 2

EXAMPLE OF ICE- POINT DEPRESSION

Or ig inal
Correction
from
Calibration Report of Ice-Point Ice -Point
Point Calibrat ion Reading^ Correct ion^

(OC) (OC) (OC) (OC)

- 0.00 0.00 0.00 0.00

50.00 0.00 -0.05 +0 .05
100.00 0.00 -0.10 +0.10
150.00 0.00 -0.15 +0.15

^Determined after heating thermometer to calibration

point and then returning to ice bath.

Now suppose the thermometer that has been heated to 150 ^C

is immediately returned to 50 *^C. The reading at 50 ^C will be
influenced by the thermometer's having been heated to a higher
temperature. That is, the reading of the thermometer will depend
on its previous thermal history. The ice-point correction taken
after heating the thermometer to 150 ^C was +0.15 *^C and, in this
case, it will remain that value after the thermometer has been
returned to 50 ^C. The ice-point correction taken after the
thermometer had been heated to only 50 ^C was +0.05; therefore,
when the thermometer is used at 50 ^C after being heated to
150 oc, it will indicate that the temperature of the bath is
49.90 oc. The volume of the bulb had increased by an amount
equal to 0.1 [+0.15-(+0.05) , thus causing the thermometer to
]

read lower at 50 OC by 0.1 ^C.

Having understood the long-term and short-term changes in
the bulb glass, we are now ready to describe the procedure to be
followed routinely by the State calibration laboratory.
The first step is to let the thermometer rest at room
temperature (23 ^C) for 3 days to enable the bulb to contract

22
back to its equilibrium volume. After this period of time an ice
point is taken. If the ice point is the same as the one on the
Report of Calibration, then it can be assumed that the correc-
tions given on the report are current and valid. Since long-term
changes can occur in the volume of the bulb with time and use, it
would not be unusual for the ice point to have a different value.
If the ice point has changed, all of the corrections should be
changed by the same amount and in the same direction (see
Section 5.2 in NBS Monograph 150 [1]). This technique gives the
operator a current set of corrections and is a r ecal ibrat ion
procedure followed by NBS for many years. If the primary stan-
dard thermometers are to be used only once or twice a week, with
a 3-day waiting period between their use, and they will be used
only from some lower point to some higher point, the corrections
determined after changing the corrections for the change in ice
point can be used. If the primary standard thermometers will be
used several times within a week, as most standards are, then one
must determine and compensate for the ice-point depression caused
by the short-term, temporary change in the bulb volume.
The data in Table 2 provides the basis for establishing
"adjusted scale corrections" in order to compensate for the ice-
point depression and long-term changes that will occur in the
volume of the bulb. This is done using the following equation.

Ci = C2 - C3 (1)

Ci = Adjusted scale correction.

C2 = Correction from table supplied in the NBS Report
of Calibration or current set of corrections
obtained from change in ice-point reading.
C3 = Ice-point correction determined by the State
laboratory after heating the thermometer to the
calibration point and returning to the ice bath.

The value C3 can be found after the thermometer is returned

from NBS by taking the ice point after heating the thermometer to
each of the calibration points and cycling back to the ice point
as was done in the example given above (Table 2, Column 3).
Adjusted scale corrections for the primary standard thermometer
used in the example are given in Table 3, Column 4.
Altering the original corrections in this manner enables the
user to know the influence of the ice-point depression on the
thermometer reading as the thermometer is heated. When the
adjusted scale correction is applied to the thermometer reading,
the resulting temperature value will be correct only if the ice-
point reading at that time is 0.00. Since the ice point will not
always read exactly 0 ^C after use, either because of the
temporary ice-point depression or long-term changes in the bulb
volume, it is necessary to take an ice point periodically (see

23
 .

TABLE 3

EXAMPLE OF ADJUSTED SCALE CORRECTIONS

Report of
Calibration Calibration Ice-Point Adjusted Scale
Point Correct ion Correct ion^ Correct ion
(C2) (C3) (Ci)

(OC) (°C) (°C) (°C)

0 .00 0 .00 0 .00 0 .00

50 .00 0 .00 +0.05 -0.05
100.00 0 .00 +0.10 -0.10
150 .00 0.00 +0.15 -0.15

^De termined after heating thermometer to calibration

point and then returning to ice bath.

Section 3.3.7) while the thermometer is in use and to record the

ice-point reading. This procedure will be discussed further in
Section 3.3.8.
Given in Table 4 are adjusted scale corrections computed for
two hypothetical primary standard thermometers, 1 and 2, used in
the sample calibration that will be discussed in Section 3.3.8.

3.3.6 Placement of thermometers in calibration bath

The thermometers to be calibrated should be placed in the

calibration bath in a row with one primary standard liquid-in-
glass thermometer at one end and the other primary standard at
the other end. This placement is shown in Figure 10. Total-
immersion thermometers should be immersed with the meniscus of
the mercury column approximately 6 mm to 12 mm above the surface
of the bath medium and partial-immersion thermometers should be
immersed to the immersion line or to the specified depth. Any
additional thermometers (see Section 3.3.7) used to measure stem
temperatures should be placed on the corresponding test
thermometers. One method of doing this is by joining two
electrical pee-wee clamps together and clamping the additional
thermometer to one clamp. The second clamp is attached to the
thermometer
Ideally, for liquid-in-glass thermometer calibrations or
when 1 iqu id- in-g lass thermometers are used as standards, the
temperature of the calibration bath should be rising slowly (no
faster than one scale division of the primary standard
thermometer in three minutes). It is also acceptable to control
the bath temperature (by observing a more sensitive thermometer)
within ±0.005 ^C. The temperature of the bath should not be de-
creasing during calibration because the mercury column at the

24
i
 TABLE 4

ADJUSTED SCALE CORRECTIONS FOR HYPOTHETICAL

PRIMARY STANDARD THERMOMETERS 1 AND 2

Primary Standard 1

Report of Ice-Point Correction Adjusted

Calibration Determined after Heating Scale
Calibration Correction to Calibration Point Correction
Point {C2) (C3) (Ci)

(°C) (°C) (°C) (°C)

0.00 +0.08 +0.08 0.00

20.00 + .11 + .10 + .01
40.00 + .02 + .10 - .08
60 .00 + .07 + .11 - .04
80.00 + .10 + .12 - .02
100.00 + .06 + .12 - .06

Primary Standard 2

Report of Ice-Point Correction Adjusted

Calibration Determined after Heating Scale
Calibration Correction to Calibration Point Correction
Point (C2) (C3) (Ci)

(°C) (°C) (°C) (°C)

0.00 +0.01 +0.01 0.00

20.00 + .17 + .01 + .16
40.00 + .04 + .02 + .02
60.00 + .11 + .03 + .08
80.00 + .12 + .04 + .08
100.00 + .06 + .04 + .02

meniscus can cling to the wall of the capillary of thethermome-

ter and remain in an erroneously high position even though the
temperature of the bath may have decreased. There is less chance
for this problem to occur if the temperature of the bath is
r ising.

3.3.7 Reading sequence and recording data

After the thermometers have been placed properly in the

calibration bath and it is regulated at the lowest point re-
quested, the thermometers are read in the following sequence:
Primary Standard 1, Thermometers Tl-TX, Primary Standard 2,
Primary Standard 2, Thermometers TX-Tl, and Primary Standard 1.
One of the thermometers Tl-TX should be the check standard.
Any additional auxiliary thermometers used for stem
temperature determinations should be read after this sequence.
An auxiliary thermometer can be any thermometer that is accurate

25
 Figure 10. Correct placement of
thermometers in calibration bath.

to within 1 or 2 degrees. It is used to measure the average

temperature of the emergent stem, both for total -immer s ion
thermometers used incorrectly at partial immersion, and for
partial-immersion thermometers. NBS uses for this purpose a
thermometer approximately 15 cm long with a range of 0 to 100
and graduated in intervals of 1 ^C.
If two calibration operators are available, one should
record the observed thermometer readings while the other reads
the thermometers. The two operators should switch positions and
take a second set of readings. It is preferable to have two
observers, since this procedure will minimize reading errors.
However, this is not absolutely necessary; at the NBS, all cali-
brations for 12 years have been done by one individual.
At specified intervals (typically, every 200 scale
divisions), the two primary thermometers should be removed from
the constant temperature bath and reinserted into the ice bath.
The ice-point readings are recorded. This procedure is used to
determine the "ice-point correction" (see Section 3.3.5 and
Section 3.3.8) for the two primary standard thermometers. Then
they are returned to the constant-temperature bath and further
readings are taken at higher temperatures.

26
 . .

Table 5(a) gives an example in which three unknown

thermometers (Tl, T2 and T3) and a check standard (T4) were
calibrated versus two primary standard thermometers (Primary
Standard 1 and Primary Standard 2). All the thermometers in this
example were total-immersion thermometers with graduations of
0.2 oc and having a range -2 oc to +102 oc. They were calibrated
at the nominal temperatures 0 oc, 20 oc, 40 oc, 60 oc, 80 oc, and
100 oc. These particular points were chosen because they
correspond to intervals of 100 scale divisions, a common practice
used in calibration of 1 iqu id- in-g lass thermometers. The
calibration proceeded from 0 oc to 100 oc. The primary
thermometers were removed from the bath after the readings were
taken at 20 oc, 60 oc, and 100 oc and remeasured in the ice bath.
This allowed ice-point corrections to their readings to be made
at those temperatures (interpolation was used for the 40 oc and
80Oc points)
After a set of data is obtained at a calibration point, it
should be inspected to see that it looks reasonable. The
calibrator should check for reading errors and to insure that the
temperature during the calibration remained constant to i0.005 °C
or was rising at a rate of no more than 1 scale division per
3 minutes. As mentioned earlier, it is not acceptable for the
temperature of the bath to decrease while data is being recorded.

3.3.8 Computation of bath temperature from readings of primary

standard thermometers
Once the data have been recorded in the format shown in
Table 5(a), and after inspection indicates that there are no
suspicious data points, computation of the corrections is done as
shown in Table 5(b).
The first step is to convert the readings of the two primary
thermometers to the correct temperature. This is done by using
the following equation

T = Tr + Ci + C4 (2)

T = True temperature.
Tr = Average of two readings for the primary standard
thermometer
Ci = Adjusted scale correction.
C4 = Ice-point correction taken after thermometer has
been used.

Having established the set of adjusted scale corrections the

only term that is necessary to determine the correct temperature
is the ice-point correction determined by taking the ice point
after the thermometer has been used (C4).
As seen in Tables 5(a) and 5(b), if the thermometer is to be
calibrated at several points over a narrow temperature range, it

27
is not necessary to take an ice point after every calibration
point, but one should be taken at intervals of every 200 scale
divisions. The ice-point correction between these points can be
determined by interpolation. Adjusted scale corrections at tem-
peratures between the calibration points can also be determined
by interpolation.
In the far left and right columns of Table 5(b) are shown
the calculations used to obtain T for Primary Standards 1 and 2.
For example, at the nominal temperature 20 ^C, Primary
Thermometer 1 had the following parameters: T = 19.76;
Ci = +0.01; C4 = +0.20, and thus T was calculated to be 19.97 ^C.
By a similar technique. Primary Thermometer 2 also indicated that
the temperature of the calibration bath medium was 19.97 ^C.
After the temperature of the bath has been determined for
each primary standard and the calculated temperatures are found

TABLE 5 (a)

SAMPLE CALIBRATION SHEET

Readings (one observer)

Nominal Pr imary Pr imary

Temp (°C) Standard 1 Tl T2 T3 T4 Standard 2

-0.10 -0.10 -0 . 03 +0.01

20 19 .75 19.92 19 .94 19.9 3 19.99 19.7 2

19 .76 19.93 19 .95 19 .93 20.00 19.73

0 -0.20 -0 .09

40 39.71 39.90 39.90 39.88 39.83 39.70

39 .72 39.92 39.91 39.88 39 .85 39.71

60 59.70 59.86 59.93 59.89 59.80 59.68

59.71 59.86 59 .94 59 .90 59 .81 59.69

-0.22 -0. 10

80 79.79 80.03 79.96 79.90 80.02 79.81

79 .80 80 .04 79 .96 79.91 80 .03 79.81

100 99.71 99.88 99.86 99.86 99.91 99.73

99.71 99.88 99.88 99 .87 99 .92 99 .74

-0.23 -0.11

28
 .

TABLE 5(b)

SAMPLE CALIBRATION SHEET

CtonputatiDns (one observer)

Primary Primary
Standard 1 Tl T2 T3 T4 Standard 2

-0.10 -0.10 -0.03 +0.01

+ .10 + .10 + .03 - .01^

19.76^ 19.75 19.92 19.94 19.93 19.99 19.72 19.72^

+.01^^ 19.76 19.93 19.95 19.93 20.00 19.73 +.16°
+.20^ +.09^
^ 19.92 19.94 19.93 20.00^ —
19.97®
< ^19.97 g

-0.20 : -0.09
+.05 +.03 +.04 -.035"

39.72 39.71 39.90 39.90 39.88 39.83 39.70 39.70

-.08, 39.72 39.92 39.91 39.88 39.85 39.71 +.02^
+.21^ +.10*^
39.91 39.90 39.88 39.84
39-85 < 39.84 ^>
3^.82

.07 -.06 -.04 .00

59.70 59.70 59.86 59.93 59.89 59.80 59.68 59.68

-.04 59.71 59.86 59.94 59.90 59.81 59.69 +.08
+.22 +.10
59.86 59.94 59.90 59.80
59.88 < > ^9.86
59.87
-0.22 -0.10
+.01 -.07 -.03 +.07

79.80 79.79 80.03 79.96 79.90 80.02 79.81 79.81

-.02 79.80 80.04 79.96 79.91 80.03 79.81 +.08
+.22 +.10
80.04 79.96 79.90 80.02
80-00 < 79.99
80.00 ^>

-.04 +.04 +.10 -.02

99.71 99.71 99.88 99.86 99.86 99.91 99.73 99.74

-.06 99.71 99.88 99.88 99.87 99.92 99.74 +.02
+.23 +.11
99.88 99.87 99.86 99.92
99-88 < 99.87
^99.88 ^>

-0.23 -0.11
.00 +.01 +.02 -.04

^Average of two readings for the primary standard thermometer (whai rounding off
.the ev^ nunber was chosen)
Mjusted scale correction corputed for primary standard at indicated tenperature.
^Ice-point correction - tak&n after calibratiOTi point.
Interpolated ice-point ccarrection.
Tteiqperature of calibration bath medium as determined by the corrected primary
^tandard.
Average of two readings for the thermDmeters being calibrated.
^Iteicperature of calibraticm batli medium as determined by average of two cor-
,rected primary standard thermometers.
Corrections given on Report of Calibration for thermometers under test.

29
to agree within 0.2 of a scale division (in our example,
0.2 X 0.2 = .04 °C) or within other limits established by the
,

operator's laboratory, the two values are averaged to obtain the

temperature of the calibration bath medium. In our example this
average is 19.97 oc.

3.3.9 Determining corrections for thermometers being calibrated

Having calculated the correct temperature of the two primary
thermometers as described in Section 3.3.8, we may now obtain the
corrections for thermometers Tl, T2, T3, and T4.
The two readings of each thermometer being calibrated and
the check standard are averaged and compared to the temperature
of the bath medium. If the thermometer under test reads higher
than the bath temperature, the correction is negative; if the
thermometer under test reads lower than the bath temperature, the
correction is positive.
Using the data taken for thermometers in Table 5(b) and the
computed bath temperature of T = 19.97 °C, we calculate that
their respective corrections are +0.05, +0.03, +0.04 and -0.03
at 20 °C. This procedure is followed for the other temperatures,
thereby generating the data for the Reports of Calibration for
these three thermometers. (The Report of Calibration for ther-
mometer Serial Number Tl is given as Figures 17 and 18.)
Ice points should be taken on the thermometers that were
calibrated not less than 72 hours after test. The second ice
point should agree with the initial ice point within the expected
accuracy of the thermometer. Thermometers with a range of 0 to
100 OC and graduated in 0.2 oc have an expected accuracy of
0.05 Oc at the calibration points (see Section 6 of NBS Monograph
150 [1]). The ice points should agree within this limit. If the
two ice points are not in agreement, indicating that the bulb has
not recovered from expanding after heating (due probably to
improper annealing), then the thermometer has to be completely
recalibrated.
The data should be analyzed to see if the results are rea-
sonable. A very useful technique for analysis is to plot the
calibration data. An example for a total-immersion thermometer
graduated in 0.2 oc divisions is shown in Figure 11. The cor-
rections at 80 oc and 100 oc are suspect for two reasons: 1) the
magnitude of the difference between adjacent corrections is close
to one scale division, and 2) the sign of the correction changes.
A deviation curve shown in Figure 11 can occur for liquid-in-
glass-thermometers. If these data were taken by two observers,
the probability for reading error is small and recalibration in
this case is not necessary. If, on the other hand, the data were
obtained by one observer, we recommend recalibration at 80 oc and
100 oc.
The plot may indicate another effect. If the trend of the
curve is downward as the temperature is increased, the
thermometer probably has gas in the bulb. The gas should be
removed (see Section 8.4 of NBS Monograph 150 [1]) and the
thermometer retested.

30
 .

.20

Uncertainty
I
1 Scale
.10 - Division
c
o
o
o
w
o
o
00 -J-

.10
20 40 60 80 100

Temperature °C

Figure 11. Thermometer correction versus temperature,

3.3.10 Stem-temperature corrections

If the thermometer being calibrated is a total-immersion
thermometer used at partial immersion or an ASTM partial-
immersion thermometer requiring a correction that reflects a
specified stem temperature, it will be necessary to correct for
the variation in the stem temperature.
Use the following formula to determine the stem-temperature
correction.

Stem-temperature correction = kn(tobs ~ ^sp) (3)

k = differential expansion coefficient of mercury (or

other thermometric liquid) in the particular kind
of glass of which the thermometer is made,

n = number of scale degrees equivalent to the length

of emergent stem (including the evaluated area
above the immersion line) ,

tobs ~ observed mean temperature of the emergent stem

(auxiliary thermometer reading),

tsp ~ specified mean temperature of the emergent stem

(for which reported scale correction applies)

All stem-temperature corrections require that four values be

determined. The first value is the differential expansion

31
coefficient of mercury (or other ther mome tr ic liquid) in the
particular kind of glass from which the thermometer is made (k).
These values have been established and used for many years and
are given in NBS Monograph 150 [1], Table 4, and discussed in
Section 5.3 of that publication.
The second value needed is the number of scale degrees
equivalent to the length of emergent stem (n). This is
essentially the number of degrees that is in the incorrect
temperature environment. If it is a total-immersion thermometer
with a portion of the stem above the bath, then the value n is
the number of degrees from the degree mark at the surface of the
bath medium to the meniscus (the top of the mercury column). If
the thermometer is a partial-immersion thermometer the value n
is the number of degrees from the immersion line to the meniscus.
Since partial-immersion thermometers are not graduated to the
immersion line, there is a portion of the stem containing no
scale divisions. It must be determined how many degrees could be
inscribed in this portion of stem and that number included in the
value n. One way of doing this is to measure the distance from
the immersion line to the lowest scale graduation, then move the
ruler on the thermometer scale and determine how many degrees
this distance represents.
The third value needed is the observed average temperature
of the emergent stem (tobs)* This value is found by actually
measuring the temperature of the environment where the emergent
stem is located. The preferred way of doing this involves the
use of faden thermometers as described in Section 5.3a of NBS
Monograph 150 [1]. Because the availability of faden ther-
mometers is limited (see Section 9.4 of this Monograph), the
emergent stem temperature can also be measured by using one or
more auxiliary thermometers. Whether one, two, or even three,
auxiliary thermometers are needed depends on the length of
mercury column emergent from the bath. Generally an auxiliary
thermometer should be attached for every 120-mm to 150-mm length
of emergent mercury column. If one auxiliary thermometer is
used, then the center of its bulb should be placed at the
midpoint of the emergent mercury column. If two auxiliary
thermometers are used, then the center of their bulbs should be
placed at 1/3 and 2/3 the height of the emergent mercury column.
An average of the two auxiliary thermometer readings is used as
the temperature of the emergent stem.
The fourth value is the specified average temperature of the
emergent stem (tgp). This value is the temperature at which the
emergent stem should be. If the thermometer is a total-immersion
thermometer having a portion of the stem out of the medium being
measured, the specified temperature of that portion of the stem
should be the temperature of the medium. If it is an ASTM par-
tial-immersion thermometer, the specified temperatures should be
the values given in ASTM Specification E-1 [3].
The stem-temperature correction is added to or subtracted
from the initial correction. Examples are given in Tables 6(a)

32
and 6(b) and a discussion is given in Section 5.3 of NBS
Monograph 150 [1].
In Tables 6(a) and 6(b), Thermometer T6 is an ASTM 33C
thermometer and at 40 the correction should reflect a stem
temperature of 30 as specified in ASTM E-1 [3]. The length of
emergent mercury column is equal to 314 mm and the centers of the
auxiliary thermometer bulbs are placed adjacent to the stem
105 mm and 209 mm from the immersion line. This procedure is
shown in Figure 12.

TABLE 6(a)

EXAMPLE OF STEM-TEMPERATURE CORRECTIONS

Readings (two observers)

T6 T7
ASTM 3 3C ASTM 3 4C
Nominal Pr imary 50 mm 50 mm Pr imary
Temp (°C) Standard 3 Immers ion Immers ion Standard 4 Auxi Aux-

40 39.909 39.96 — 39.882 31 26

39.923 39.97 39.883 31 26

39.919 39.96 ~ 39.901

39.933 39.96 39.902

0 -.138 -.151
-.139 -.150

45 44.770 — 44.85 44.751 33

44.784 44.86 44.750 33

44.784 — 44.86 44.751

44.793 44.87 44.752

0 -.139 -.154
-.139 -.153

The indication on auxiliary 1 thermometer gives the average

temperature of the lower portion of the emergent stem of T6 and
that on auxiliary 2 thermometer the average temperature of the
upper portion of emergent stem of T6. The average temperature of
the emergent stem is the average temperature of the two auxiliary
thermometers. Since in the example there are two observers, each
auxiliary thermometer should be read twice. An average of the
two values for each auxiliary thermometer is used.

33
 TABLE 6 (b)

EXAMPLE OF STEM- TEMPERATURE CORRECTIONS

Computations (two observers)

T6 T7
ASTM 3 3C ASTM 34C
Pr imary 50 mm 50 mm Pr ima ry Aux^ Aux.
Standard 3 Immersion Immers ion Standard 4

a
39 .916^ 39 .909 39 .96 39.882 39 .882 31 26
-.032° 39.923 39.97 39.883 -.013"^ 31 26
+.138° + .150*^
39 .96*
40 .022*^ 40 .019
-40 .020'
+ .06-

39.926^ 3 9.919 39.96 39 .901 39 .902f

-.032° 39 .933 39.96 39 .902 -.013'^
+ .138^^ +.150°
39 ,96*
40 .032' 40 .039"^
-40 .036'
+ .08^
+ .07"
,138 -.02^ -. 151
.139 -.150
+ .05-

44.777 44.770 44 .85 44 .751 44 .750 33

-.042 44 .784 44 .86 44 .750 -.035 33
+ .139 + .154
44 .86
44.874 -44 .872- 44 .869

+ .01

44.788 44.784 44 .86 44 .751 44 .752

-.042 44.793 44 .87 44 .752 -.035
+ .139 + 154
.

44 .86
44 .885 44 .871
-44 .878-
+ .02
+ .02
-.139 -.02 -.154
-.139 -.153
.00

Average of two readings for the primary standard thermometer (when rounding off, the even number
j^was chosen) .

^Adjusted scale correction computed for primary standard at indicated temperature.

^Ice-point correction - taken after calibration point.
^Temperature of calibration bath medium as determined by the corrected primary standard.
^Average of two readings for the thermometer being calibrated.
Temperature of calibration bath medium as determined by average of two corrected primary standard
thermome ter s.
^Correction for thermometer under test excluding possible stem-temperature correction.
^Average correction of two observers.
.Stem-temperature correction (see text for computation).
-"Algebraic addition of average correction of two observers and stem-temperature correction. This
value is given on the Report of Calibration.

34
 /N

I
A
78 Degrees

314 mm

209 mm
i
20 Degrees

120 mm
105 mm i
60 mm
15 Degrees 15 Degrees

N/ Si/

i i

Figure 12. Measuring stem Figure 13. Measuring stem

ten^erature on an ASTM 33C temperature on an ASTM 34C
thermometer using two thermometer using one
auxiliary thermometers. auxiliary thermometer.

35
 The average observed stem temperature (tobs) 28.5
[(31+26)/2], the specified stem temperature (tgp) is 30 °C, the
coefficient of expansion of mercury in glass for Celsius
thermometers is 0.00016 and the number of degrees emergent from
the bath, including the portion from the immersion line to the
first graduation (n), is 93 degrees.

Stem-temperature correction - 0 .00016 n (tobs~*-sp^

= 0.00016 (93) (28.5-30)

= -0.02 OC

Notice that the values tgp and tpj-,5 are reversed in the above
formula, as compared to the one given in NBS Monograph 150 [1] ,

since we are applying this correction to a correction and not to

the thermometer reading itself.
As a second example, thermometer T7 is an ASTM 34C
thermometer. Since the length of the emergent mercury column is
only 120 mm, placing the center of the bulb of one auxiliary
thermometer adjacent to thermometer T7, 60 mm from the immersion
line, is sufficient to measure the average temperature of the
emergent stem. This example is shown in Figure 13.
In this example, the average observed stem temperature
(tobs) is 33 °C, the specified stem temperature is 37 °C, the co-
efficient of expansion of mercury glass is 0.00016 and the
in ,

number of degrees from the immersion line to the mercury meniscus

is 35 degrees.

Stem-temperature correction = 0.00016 n (tobs~tsp)

= 0.00016 (35) (33-37)

= -0.02 OC

4 . DCXniMENTATION

4.1 Report of Calibration

After corrections are determined (see Section 3.3.9 of this

Monograph), a Report of Calibration is generated for each
thermometer. The Report should include the following:

1) Company, agency, or establishment that submitted the

thermometer,
2) Manufacturer's trade name and serial number of the
thermome ter

36
 ,

3) Range and graduation interval,

4) Calibration temperatures and corrections,
5) Temperature scale used (IPTS-68)
6) Instructions for use of corrections,
7) Estimated uncertainties,
8) Other notes concerning how the thermometer was
calibrated and how it can be recalibrated.

The following list contains examples of notes given on NBS

Reports of Calibration.

ALL TOTAL- IMMERSION THERMOMETERS.

The tabulated corrections apply for the condition of total
immersion of the bulb and liquid column. If the thermometer
is used at partial immersion, apply an emergent stem correc-
tion as explained in the accompanying stem correction sheet.
(See Figures 14 and 15.)

ALL PARTIAL- IMMERSION THERMOMETERS WITH NO ST EM -TEMPERATURE

CORRECTIONS.
The thermometer was tested in a large, closed-top,
electrically heated, liquid bath at an immersion of .

The temperature of the room was about 2 3 °C (7 3 °F). If the

thermometer is used under conditions which would cause the
average temperature of the emergent liquid column to differ
markedly from that prevailing in the test, appreciable
differences in the indications of the thermometer would
result.

PARTIAL- IMMERSION THERMOMETERS WITH THE SAME STEM TEMPERATURE AT

EACH CALIBRATION POINT.
These corrections are applicable for immersion and
for an emergent stem temperature of .

PARTIAL- IMMERSION THERMOMETERS WITH DIFFERENT STEM TEMPERATURES

AT EACH CALIBRATION POINT.
These corrections are applicable for immersion and
for temperatures of the emergent mercury column as given
below :

Reading of thermometer
Temperature of emergent mercury .

RECALIBRATION NOTE USING ICE POINT.

The tabulated corrections apply provided the ice-point
reading is If the ice-point reading is found to
.

be higher (or lower) than stated, all other readings will be

higher (or lower) by the same amount.

37 it)
RECALIBRATION NOTE USING ICE POINT FOR TOTAL- IMMERSION
THERMOMETERS GRADUATED UNDER 15 0 °C AND IN GRADUATION INTERVALS
OF 0.2 OC OR LESS.
The tabulated corrections apply provided the ice-point
reading taken after exposure for not fewer than 3 days to a
temperature of about 23 °C (73 °F) is . If the ice-
point reading is found to be higher (or lower) than stated,
all other readings will be higher (or lower) by the same
amount. If the thermometer is used at a given temperature
shortly after being heated to a higher temperature, an error
of 0.01 ° or less, for each 10° difference between the two
temperatures, may be introduced. The tabulated corrections
apply if the thermometer is used in its upright position; if
used in a horizontal position, the indications may be a few
hundredths of a degree higher.

STEAM POINT USED AS REFERENCE POINT.

The tabulated corrections apply provided the reading when
the thermometer is immersed in steam at 100 °C (212 *^F) is
If the reading is found to be higher (or lower)
.

than stated, all other readings will be higher (or lower) by

the same amount. The temperature of steam is 100 *-*C
(212 Of) only if the pressure is 760 mm (29.9 21 inches). If
the pressure differs from 760 mm (29.921 inches), allowance
must be made for this. If the pressure is higher (or lower)
than 760 mm (29.921 inches) the temperature will be higher
(or lower) than 100 ^C (212 °F) by approximately 0.037 °C
per mm difference (1.68 °F per inch difference).

CALORIMETRIC THERMOMETERS HEATED TO TOP POINT BEFORE TESTING.

The thermometer, before testing, was heated to the
temperature of the highest test point. The application of
the tabular corrections to the readings of the thermometer
will give true temperature differences provided the
thermometer is used in its upright position and is heated
previously (within an hour before using) to the highest
temperature to be measured.
SETTING FOR BECK MANN THERMOMETERS.
The tabulated corrections apply for a "setting" of 20 °C.
Setting factors for use with other settings are given on the
accompanying sheet. (See Figure 16.)

ICE POINT TAKEN AFTER TEST, SUCH AS CLINICAL AND KINEMATIC

VISCOSITY THERMOMETERS.
The tabulated corrections apply for the condition of
immersion indicated provided the ice-point reading, taken
after heating to for not fewer than 3 minutes, is
. If the ice-point reading, which should be taken
within 5 minutes after removal of the thermometer from the
heated bath, is found to be higher (or lower) than stated

38
 all other readings will be higher (or lower) by the same
amount.

If there is any other information concerning the thermometer

that the laboratory supervisor thinks is important for the
user to know, a note should be written and included on the
Report. If the thermometer was calibrated against NBS
standards, it might be advisable to state that the ther-
mometer is traceable to NBS and give the serial numbers of
the standards and the NBS test number that the standards
were tested under. (Figures 17 and 18 show a sample Report
of Calibration that would be issued by NBS for Thermometer
Tl.

4.2 Report of Test

A Report of Test is issued for an electronic temperature
measuring system and should also include the name of the company
that submitted the system, identification of the system (serial
number and model), temperature scale used, calibration tempera-
tures, and output of the sensor, which may be in volts, ohms,
hertz, etc. All conditions of the test, such as the immersion
depth of the probe and settings on the electronic readout device,
should also be included in the Report. (Figure 19 is a sample
Report of Test issued by NBS.)

39
 CORRECTION FOR EMERGENT STEM
This stem correction sheet is designed to explain Stem correction=0.00016x65 (85°—38°) =+0.49° C.
the process of computing the corrections for emergent The true temperature is therefore the observed
stem, but has not been specifically adapted to the reading, 84.76° C, + tabular correction as interpolated
individual thermometer with which it is issued. from the report, -(- emergent stem correction
If a total-immersion thermometer is actually used (+0.49° C).
with a part of the liquid column in the capillary
emergent from the bath, a stem correction must be
Example 2. — Suppose the observed reading was
calculated and applied as explained below, making
780° F and the thermometer was immersed to the
200° mark on the scale, so that 580° of the column
use of the formulas given.
projected into the air, and the mean temperature of
The general formula used in computing the correc-
the emergent column was found to be 170° F, then
tion for emergent stem is:
Stem correction=0.00009x580(780°—170°)=+32° F
Stem correction=A;x«(*i — t),
as a first approximation.
where fc=the differential expansion coefficient of
mercury (or other liquid in the thermom- Since the result shows that the bath temperature
eter such as alcohol, toluene, pentane, was approximately 780° F+32° F, a second approxi-

etc.) in the particular kind of glass of mation should be made, using h=S12° F instead of
which the thermometer is made; (for ii=780° F. This gives-
numerical values see below) Stem correction=0.00009x580(812°— 170°)=+34° F.
Tirrnumber of degrees emergent from the bath; The true temperature is therefore the observed
fi=temperature of the bath; reading, 780° F, + tabular correction as interpolated
t=mean temperature of the emergent stem. from the report, + emergent stem correction (+34° F).
Example 1. — Suppose the observed reading was It will be noted that if the average temperature of
84.76° C and the thermometer was immersed to the the stem is below that of the bulb, the sign of the cor-
20° mark onthe scale, so that 65° of the column pro- rection will be +, while if the temperature of the
jected into the air, and the mean temperature of the stem is above that of the bulb, the sign of the correc-
emergent column was found to be 38° C, then tion will be — .

EXPLANATORY NOTES ON THE EMERGENT STEM CORRECTION

Some thermometers are pointed and graduated by the maker to VALUES OF k FOR MERCURY-IN-GLASS THERMOMETERS
read correct, or approximately correct, temperatures when the
For Celsius thermometers For Fahrenheit thermometers
bulb and the entire liquid index In the stem are exposed to the
temperature to be measured, while other thermometers are so Mean k k Mean k A:

pointed and graduated that they will read correct, or approxi- temp. for for "boro- temp. for for "boro-
mately correct, temperatures when the bulb and only a short
t, +t "normal" sillcate" t, + t
"normal" silicate"
2 glass glass 2 glass glass
length of the stem of the thermometer are immersed in the bath,
the temperature of which is to be measured. Thermometers of the 0° 0.000158 0.000164 0° 0.000088 0.000091
former class are known as "total-immersion thermometers," and 100 158 164 200 88 91
150 158 165 300 88 92
those of the latter class as "partial-immersion thermometers." 200 159 167 400 89 93
250 161 170 500 90 95
300 164 174 600 92 97
Total-immersion thermometers are tested under the condition 350 178 700 .000100
of total immersion and the corrections resulting from such a test 400 183 800 103
450 188
will serve to reduce the observed readings of the thermometer to
true temperatures only if the thermometer is used as a total-
If kind of glass of which the thermometer is made is
the
immersion thermometer. If such a thermometer is actually used known, the value of k to be used in computing the stem correc-
as a partial-immersion thermometer, i.e., with a part of the tion mav be taken from the above table. If the kind of glass is
mercury column emergent into the space above the bath, and not known, use fc = 0.00010 for Celsius or 0.00009 for Fahrenheit
thermometers. High-grade thermometers are now generally made
with the emergent stem therefore either colder (or warmer) than of "normal" or "borosilicate" glasses. If a thermometer is grad-
the bulb, the thermometer will obviously read lower (or higher) uated only to about 450° C (850° F), it may be made of either
than it would under the condition of total Immersion. Hence, if of the above glasses; if it is graduated to 500° C (932° F) and
is actually usable at that temperature, it is made of one of the
a total-immersion thermometer is so used, a so-called stem borosilicate glasses or a similar glass.
correction must be applied to the observed reading in addition The expansions of liquids such as alcohol, toluene, etc., vary
to the correction taken from the accompanying table of correc- quite rapidly with the temperature, so that k varies considerably
for different temperature intervals. An approximate stem cor-
tions. This stem correction is very large if the number of degrees
rection for such thermometers may be calculated by taking k in
emergent and the difference of temperature between the bath the above equation =
0.001 for Celsius thermometers or 0.0006
and the space above it are large. It may amount to more than for Fahrenheit thermometers.
20° C (36° F) for measurements made with a mercury The value of f, the mean temperature of the emergent stem, is
the most difficult of the terms in the above formula to estimate.
thermometer at 400° C (750° F). It may be quite accurately measured by the use of special capil-
lary thermometers. This is. however, very rarely done except in
the testing laboratory, and then only when the stem correction
The coefficient k is different for different kinds of glass and, must be determined with considerable precision (to 10 percent or
even for the same kind of glass, it differs for different tempera- better). In general, the value of t may be determined to a suffi-
ture intervals, i. e., different values of (f, —
t). Values for k cient approximation by judgment or preferably by suspending an
auxiliary thermometer close beside the emergent stem, with the
for two widely used thermometric glasses, for use in calculating bulb of the aux lary thermometer somewhat nearer to the top
stem corrections are tabulated as follows of the bath than to the liquid meniscus.

Figure 14. Stem temperature explanation sheet attached to

reports for total -immersion thermometers.

40
 — —— .

DIFFERENTIAL CORRECTION FOR EMERGENT STEM

This sheet is designed to explain a convenient process The result, 0.006°, should be added to the difference
of computing tlie differential correction for emergent between T'z and T°i, found after applying to each
stem but has not been specifically adapted to the the corrections tabulated on the Report of Calibration.
individual thermometer with which it is issued.
If a calorinietric or differential or a metastatic
—
ICxAMPLE 2. Suppose a Beckmann thermometer was
(Beckmann) thermometer is actually used with a >
immersed to the 0° mark on its scale; its initial reading,
r°,. was 2.058°; its final reading,'r°2, was 5.127°; the
part of its stem emergent from the bath, a differential
stem correction must be calculated and applied as
mean temperature, *°, of the emergent stem was 24°
explained below, making use of the formulas there
C and the setting, S°, was 25° C; then-
;

given.
Differential stem correction =0.00016 (5-2) (25° +
Example 1. — Suppose the point. 7°. to which the 2° +
5° -24°) = +0.004°.
thermometer was immersed was 16° C its initial :

reading, T",, was 24.127° its final reading, T"-, was

: The result + 0.004°, should be added to the difference
27.876° : the mean temperature. t°. of the emergent l>etween T°: and !r°i, found after applying to each the
stem Avas 26° ; then corrections tabulated on the Report of Callbratioffi and
Differential stem correction=0.00016(28 — 24) (24° + multiplying the difference by the setting factor (1.0014
28° -16° -26°) +0.006°. = for a setting of 25° )

EXPLANATORY NOTES ON THE EMERGENT STEM CORRECTION

Calorimetric and differt'iitial and metastatic (Beckmann) 1° = scale reading to which the thermometer was Immersed ;

thermometers are pointed and Kraduated by the makers to t°=mean temperature of the emergent stem.
read correct or approximately correct temperature intervals
when the bnlb and the entire mercury index in the stem are It should be noted that this differential stem correction, as
exposed to the temperature to be measured, and such thermom- well as the difference d" may be either positive or negative,
,

eters are standardized for the condition of "total immer- and the correction must be applied with due regard to these
sion," and the corrections Riven in reports apply for the con- signs.
dition of total Immersion.
emergent stem coiTection for metastatic
Differential
In practicall.v all cases, however, such thermometers are
actually used with only the bulb and a portion of the stem
(licckmnnn) —
thermometers. The differential emergent stem
correction may be computed from the following formula, pro-
immersed, the remainder of the stem projecting into the air vided the thermometer Is immersed to its 0° mark :

above the bath.

Differential stem correction = fcXrf(S° + 7'°i+7"'2— 1°) ;

In such cases the accuracy attained can be increased by

applying' the necessary corrections for the emergent stem. where S° = the "setting" of the thermometer, and fc, d. T°i.
The emergent stem correction corresponding to any reading T°2 and t° have the meanings given in the preceding para-
may be computed by means of the formula graph.

It should be noted that this differential stem correction,

7'°
Stem correction =fcx "( f°) ; as well as the difference d. may be either positive or nega-
tive, and the correction must be applied with due regard to
where fc=:the differential expansion coefficient of mercury in
these signs.
glass ;

=
0.00016 for Celsius thermometers ; A Beckmann thermometer of the ordinary type should not
=
0.00009 for Fahrenheit tlieruKmieters :
be used with any part of the lower portion of the stem ex-
n=number of degrees emergent from the bath : posed, as this part ma.v contain from 5 to 10 times as much
r° =
temperature of the bath: mercury per centimeter as the graduated portion, and if ex-
J° =
niean temperature of the emergent stem. posed introduces a large and uncertain error.

However, in differential measurements the parts of the In case it Is necessary to use such a thermometer with some
scale on which readings are made are. in general, emergent of the lower portion of the stem emergent from the bath,
from the bath and it is usually permissible to assume that f the necessary correction may be computed from the above
will be constant during the measurement. In this case, in- formula, provided S° in the formula is replaced by S°+m°,
stead of calculating the stem corrections for the initial and where TO is the number of degrees the temperature of the
final readings, and applying them separatel.v. the differential thermometer must be lowered to bring the meniscus from
stem eorrection can be more conveniently computed by using the zero position on its scale to the point of immersion.
the appropriate formula of the two given below.
If the thermometer Is immersed to some point other than Its
Dijfferentinl emerfjent stem correction for ordinary thermom- 0° mark, as would ordinarily be the case with thermometers
FterK.- -The differential stem correction may he computed having the 0° graduation at the top of the scale, the differ-
from the following formula : ential stem correction may also be computed from the above
formula, provided 8° in the formula is replaced by iS°+m°,
Differential stem correction = fcXrf(7"'i + 7'°:— 7° — f°) ;
where m Is the number of degrees the temperature of the
thermometer must be lowered to bring the meniscus from the
J-=0.00016 for Celsius thermometers :
zero position on its scale to the point of immersion. Provided
= 0.00009 for I'ahrenhelt thermometers: the points at which readings are made are above the point
7'°i = the Initial reading; to which the thermometer Is immersed, the preceding state-
7'°- = thp final reading: ment is applicable whether the point to which the thermom-
(7 = 7'°~T°i : eter is immersed is on the scale or below it.

Figure 15. Stem temperature explanation sheet attached to

reports for calorimetric thermometers calibrated at total
immersion.

41
 This sheet gives a table of setting factors and an example to explain the processes
employed in calculations. The example, however, has not been specifically adapted to the
individual thermometer with which this sheet is issued.
The thermometer is said to have a given setting, for instance 20°C, if its scale reading is
0° when the temperature of the bulb is 20°C.
The following table, calculated for thermometers of Jena 16'" glass, gives a series of
factors corresponding to different settings by which each observed temperature difference
must be multiplied after applying the corrections given on the preceding page. This factor is
made unity for a setting of 20°C.

For any scale length up to 15°

Setting Factor Setting Factor

0°C 0.9935 55°C 1 .0092

5 0.9951 60 1.0102
10 0.9967 65 1.0112
15 0.9984 70 1.0121
20 1 .0000 75 1.0129
25 1.0014 80 1.0137
30 1.0029 85 1.0146
35 1.0042 90 1.0154
40 1.0056 95 1.0162
45 1.0070 100 1.0170
50 1.0082

As an illustration, suppose the following observations were made:

Setting =25 C. Lower reading = 2.058"

Stem temperature = 24° Upper reading = 5.1 27*'
Lower Upper
Observed reading = 2.058 5.127
Correction from certificate = +.005 —.008

Corrected upper reading = 5.119

Corrected Tower reading = 2.063

Difference = 3.056
Setting factor 1.0014 corr. +.004
Emergent stem correction* = +.004
Corrected difference = 3.064

See accompanying stem correction slieet.

Figure 16. Table of setting factors attached to reports for

Beckmann thermometers.

42
 U.S. DEPARTMENT OF COMMERCE
NflTIONflL BUREAU OF STftNDftRDS
NflTIONflL MEASUREMENT LABORfiTORY
gaithersburg; md. aeasg

REPORT OF CALIBRATION
LIQUID-IN-GLASS THERMOMETER.

TESTED FOR: NATIONAL BUREAU OF STANDARDS

DIVISION 5££
MARKED: SURETY Tl
RANGE: -£ TO DEGREES C IN 0. £ DEGREE

THERMOMETER CORRECTION
READING (IPTS-68)**
-. lei C . IIZI C
£0. m . 05
40. 00 -. 07
60.00 .01
80. 00 -. 04
100.00 .00

**ALL TEMPERATURES IN THIS REPORT ARE BASED ON THE INTERNATIONAL

PRACTICAL TEMPERATURE SCALE OF 1968, IPTS-6a. THIS TEMPERATURE SCALE
WAS ADOPTED BY THE INTERNATIONAL COMMITTEE OF WEIGHTS AND MEASURES AT
ITS MEETING IN OCTOBER, 1968, AND IS DESCRIBED IN "THE INTERNATIONAL
PRACTICAL TEMPERATURE SCALE OF 1968 AMENDED EDITION OF 1975,"
METROLOGIA 1£, NO. 1, 7-17 (1976).

ESTIMATED UNCERTAINTIES IN THE ABOVE CORRECTIONS DO NOT EXCEED 0.05

DEGREE UP TO 102 DEGREES C .

FOR A DISCUSSION OF ACCURACIES ATTAINABLE WITH SUCH THERMOMETERS SEE

NATIONAL BUREAU OF STANDARDS MONOGRAPH 150, LI QUID- IN-GLASS
THERMOMETRY.

IF NO SIGN IS GIVEN ON THE CORRECTION, THE TRUE TEMPERATURE IS HIGHER

THAN THE INDICATED TEMPERATURE; IF THE SIGN GIVEN IS NEGATIVE, THE TRUE
TEMPERATURE IS LOWER THAN THE INDICATED TEMPERATURE. TO USE THE CORREC-
TIONS PROPERLY, REFERENCE SHOULD BE MADE TO THE NOTES GIVEN BELOW.

THE TABULATED CORRECTIONS APPLY FOR THE CONDITION OF TOTAL IMMERSION OF

THE BULB AND LIQUID COLUMN. IF THE THERMOMETER IS USED AT PARTIAL IMMER-
SION, APPLY AN EMERGENT STEM CORRECTION AS EXPLAINED IN THE ACCOMPANYING
STEM CORRECTION SHEET.

TEST NUMBER 999999

COMPLETED l-££-85

Figure 17. Facsimile of page 1 of an NBS Report of Calibration

for thermometer, Serial Number Tl.
 PfiBE £

THE TflBULfiTED CORRECTIONS APPLY PROVIDED THE ICE-POINT READING, TOKEN

OFTER EXPOSURE FOR NOT FEWER THAN 3 DAYS TO A TEMPERATURE OF ABOUT 23
DEGREES C (73 DEGREES F), IS -0.10 C . IF THE ICE-POINT READING IS FOUND
TO BE HIGHER <0R LOWER) THAN STATED, ALL OTHER READINGS WILL BE HIGHER
(OR LOWER) BY THE SAME AMOUNT. IF THE THERMOMETER IS USED AT A GIVEN
TEMPERATURE SHORTLY AFTER BEING HEATED TO A HIGHER TEMPERATURE, AN ERROR
OF 0.01 DEGREE OR LESS, FOR EACH 10-DEGREE DIFFERENCE BETWEEN THE TWO
TEMPERATURES, MAY BE INTRODUCED. THE TABULATED CORRECTIONS APPLY IF THE
THERMOMETER IS USED IN ITS UPRIGHT POSITION; IF USED IN A HORIZONTAL
POSITION, THE INDICATIONS MAY BE A FEW HUNDREDTHS OF A DEGREE HIGHER.

FOR THE DIRECTOR,

NATIONAL MEASUREMENT LABORATORY

TEST NUMBER 999999

COMPLETED l-c:£-85 ROBERT J. SOULEN, JR.
CHIEF, TEMPERATURE AND PRESSURE DIVISION
CENTER FOR BASIC STANDARDS

Figure 18. Facsimile of page 2 of an NBS Report of Calibration

for thermometer, Serial Number Tl.

44
FORM NBS — 259
IREV. I2-7«l

U.S. Department of Commerce

NATIONAL BUREAU OF STANDARDS
GAITHERSBURG, MD 20899

REPORT OF TEST
Surety Model 1234 RTD Digital Thermometer S/N 4567
Resistance Thermometer S/N 999

Submitted by

National Bureau of Standards

Gaithersburg, Maryland

Resistance thermometer S/N 999, which was attached to a Surety Model 1234 RTD
Digital Thermometer S/N 4567, was calibrated by intercomparison with a
standard platinum resistance thermometer in stirred liquid comparison baths at
four temperatures. The unit was allowed to warm up 20 hours before
measurements were made. An ice point was taken before and after test by
immersing the probe in an ice bath. The probe was immersed to a depth of
18 cm for all measurements and the ambient temperature of the laboratory was
23 °C. The results are given below.

Reading of
Bath Temperature Resistance Thermometer

0.00 °C +0.40 °C
50.00 50.49
100.00 100.48
125.00 125.40
150.00 150.23
0.00 +0.40

The uncertainty of the bath temperatures given above is estimated not to

exceed 0.01 °C to 100 °C and 0.02 °C from 100 to 150 °C.

All temperatures in this report are based on the International Practical

Temperature Scale of 1968, IPTS-68. This temperature scale was adopted by the
International Committee of Weights and Measures at its meeting in October
1968, and is described in "The International Practical Temperature Scale of
1968, Amended Edition of 1975", Metrologia 12, No. 1, 7-17 (1976).

For the Director

National Measurement Laboratory

Robert J. Soulen, Jr., Chief

Temperature and Pressure Division
Center for Basic Standards

Test No. 999999

Date : May 31, 1985

Figure 19. Facsimile of an NBS Report of Test for an electronic

temperature measuring system.

45
5. MAINTAINING LABORATORY STANDARDS AND ESTIMATION OF
LABORATORY UNCERTAINTY

5.1 Estimates of Inaccuracy of Calibration of Liquid-in-

Glass Thermometers

The inaccuracy of a temperature measurement refers to a

consistent offset or systematic difference between readings and
the proper temperature. Several factors affect the accuracy of
liquid-in-glass thermometer readings; the ones described in this
Monograph include

o Inaccuracy in IPTS-68. See Section 2.3.1.

o Stem corrections for partial-immersion thermometers.
See Section 3.3.10.
o Improper preparation of ice bath. See Section 2.2.1.
o Temperature gradients in temperature baths. See
Section 2.2.2.
o Bubbles or gas in mercury. See Section 3.1.
o Expansion of bulb glass. See Section 3.3.5.

The first factor is beyond the control of the laboratory;

the others will have to be carefully studied for the particular
laboratory, and their influence on accuracy will have to be
assessed.
To indicate how the values for inaccuracy were determined at
NBS, we note the following. Ice baths prepared over a number of
years were compared with a platinum resistance thermometer that
has a measurement inaccuracy of 0.001 OC. It was found that the
ice bath temperatures varied by ±0.005 OC, so we assign an
inaccuracy of this value to this factor. As for temperature
gradients in the constant temperature baths, three platinum
resistance thermometers read simultaneously were used to
demonstrate that, for baths used from 0 oc to 100 oc, the maximum
variation radially or horizontally was found to be 0.005 OC. For
baths used from 100 °C to 200 ^C the gradients were found to be
0.01 OC at 100 OC gradually increasing to a value of 0.025 oc at
200 *-*C. Concerning errors introduced by poor thermometer
quality, thermometers used as primary standards at NBS did not
have bubbles or gas, so these systematic errors were negligible.
Finally, when 1 iqu id- in-g la ss thermometers were used as
standards, expansion of the bulb was corrected for by the
techniques discussed in this Monograph (Sections 3.3.5 and 3.3.8)
and thus introduced no significant residual inaccuracy. We
summarize these results in Table 7.

5.2 Maintaining the Accuracy of Primary Laboratory

Standards
The experience at the NBS has been that no factors other
than those summarized in Section 5.1 have a significant effect on
the accuracy of primary liquid-in-glass thermometers. Indeed,

46
 TABLE 7

MAJOR SYSTEMATIC ERRORS (INACCURACY) IN THE NBS

LIQUID- IN-GLASS THERMOMETER CALIBRATION SERVICE

Effect Inaccuracy
Range 1 Range 2

0 - 100 oc 100 - 200 oc

Ice bath 0.005 Oc 0.005 Oc

Bath gradients 0.0 05 Oc 0.01 - 0.025 Oc

TOTAL 0.010 Oc 0.015 - 0.030 Oc

many of them refer to laboratory techniques (ice bath, gradients

in constant temperature bath) that, if properly established at
the outset, will not introduce inaccuracies at a later time.
Of the remaining effects, the single largest source of
calibration offset is the bulb volume change of the thermometers.
If adjusted scale corrections are calculated for each primary
standard thermometer, this effect will be accounted for each time
the primary thermometers are used. By using two thermometers,
the likelihood that the corrections are incorrectly applied will
be small.
Use of these aforementioned techniques will preserve the
original calibration accuracy of the primary thermometers
provided by NBS. Thus the thermometers need not be returned to
NBS for recalibration. In the event that the IPTS-68 is revised,
NBS will provide the data so that the adjusted scale corrections
may be changed accordingly.

5.3 Using Check Standards for Quality Assurance

the previous section we reviewed the role of the two

In
primary thermometers used in each calibration and the accuracy of
the temperatures defined by them. In this section we describe
the role of the check standard in evaluating the quality (that
is, the imprecision) of the calibration process. Remember that
the check standard is included in each calibration run. It is
easy therefore to obtain a large number of data points on the
check standard in a short period of time. We recommend that the
laboratory construct a control chart (to be defined below) from
these data. With a sufficient number of points, the data will

47
have statistical significance that, when analyzed, can indicate
clearly the performance of the laboratory.
As an example, we discuss some data obtained at NBS for
check standard Serial Number 48425. This thermometer is a total
immersion type, with a range of -2 to +102 ^C, and with scale
divisions of 0.2 ^c. Using the measurement techniques described
in this Monograph, the measurement resolution was found to be
0.01 OC (the cross hair of the telescope is placed at the center
of the scale division thus dividing the scale division into two
equal parts; the meniscus was read to 1/10 of these divisions).
This thermometer was incorporated as a check standard in several
calibrations conducted from 1976 to 1984. Measurements were
repeated at many temperatures over the full range of the thermo-
meter. In these experiments the bath temperature was not defined
by two primary liquid-in-glass thermometers, but by a calibrated
standard platinum resistance thermometer. The imprecision of
this resistance thermometer is only 0.001 °C, so that no signifi-
cant source of uncertainty was introduced in the temperature
scale. The thermometer was allowed to rest at room temperature
(23 °C) for 3 days before being calibrated, allowing the volume
of the bulb to recover from a previous thermal cycling. In
Figures 20 and 21 we have plotted control charts for this thermo-
meter. Such charts consist of the correction for the thermometer
(before and after the ice point shift was accounted for) plotted
versus time. We chose to show the data at four temperatures.
Such plots immediately reveal the consistency of the calibration
process and can indicate calculation errors or degradation of the
process.
In addition to plotting the data in a control chart format
shown in these figures, we recommend that the same data be
submitted to elementary statistical analysis. At each
temperature the average correction and the standard deviation may
be calculated with the pressing of a few keys on almost any
scientific hand-held calculator. In this way, the quality of the
calibration process may be quantified. When this was done for
the data for check standard Serial Number 48425, it was found
that the standard deviation at all temperatures was 0,01 oc.
This is perhaps understandable, given the fact that the measure-
ment resolution is also 0.01 oc. At any rate, as more data are
accumulated, these quantifiers may be used to watch for any
changes.
For this liquid-in-glass thermometer, the imprecision (i.e.,
standard deviation) was found to be 0.01 *^C. Since the impreci-
sion of the primary thermometer was so small (0.001 °C) in this
case, no additional uncertainty was introduced by the primary
thermometer. This will not be the case in the State calibration
laboratory, where the primary thermometers will be liquid-in-
glass thermometers of performance comparable with check standard
Serial Number 48425. Thus, if two primary thermometers with
performance similar to thermometer Serial Number 48425 are used
to measure the bath temperature, and the measurements are assumed
to be independent, then the standard deviation of the temperature

48
 0.20

o 0.10

u
0)

o
o 0.00

*0>
0)
E
o
E
0)
.c 0.10

-0.20
76 77 78 79 80 81 82 83 84

Year

Figure 20. Control chart for thermometer, Serial Number 48425,

at four temperatures (* at 0 ©C, + at 40 ^C, X at 80 ^c, and
# at 100 °C) The readings were not corrected for ice-point
.

shifts.

0.20

P 0.10

u
0)

o
u
w
0)

E
o
E
0)

Figure 21. Control chart for thermometer, Serial Number 48425,

at four temperatures (* at 0 oc, + at 40 oc, X at 80 OC, and
# at 100 oc) The readings were corrected for ice-point
.

shifts.
49
will be ^
X (0.01 OC) = 0.014 oc. If we now measure a check
standard (again with a performance comparable with the other two)
repeatedly versus these two primary thermometers, and construct a
control chart, the standard deviation will be the square root of
the sum of the squares of the standard deviation of the check
standard and of the thermometers used to define temperature.
Thus we would expect the standard deviation for a State calibra-
tion laboratory to be /3 x (.01) = 0.017 ^C. That is an increase
by a factor of ^ is introduced into the standard deviation
because two liquid -in-glass thermometers are used as the primary
standards instead of a platinum resistance thermometer.
Check standards were studied at NBS versus a standard
platinum resistance thermometer for four ranges in order to
determine the standard deviation. We present these results in
Table 8. Remember: The standard deviation expected at a State
calibration laboratory using two other primary thermometers
should be roughly a factor of larger than that obtained by
NBS. Note too that the standard deviation increases at higher
temperatures, and that this is most likely due to the decreasing
resolution of the liquid-in-glass thermometer and larger time-
varying gradients in the bath.

TABLE 8

STANDARD DEVIATION (IMPRECISION) OF TEMPERATURE

OBTAINED WITH LIQUID- IN-GLASS THERMOMETERS COMPARED
WITH A STANDARD PLATINUM RESISTANCE THERMOMETER AT NBS

Temperature Thermometer
Range Gr aduations Resolution Std. Dev.

(OC) (OC) i°C) (°C)

0 - 50 0.1 0.005 0.006

50 - 100 0.1 0.005 0.008
0 - 100 0.2 0.01 0.009
100 - 200 0.2 0.01 0.028

With this table and the discussion above, a State calibra-

tion laboratory can estimate its best achievable imprecision.
This table should be compared with the value calculated from the
actual data taken at the laboratory. The extent to which the two
agree or disagree and the course of action to be taken depend on
the goal for precision set by the laboratory.
To construct a final total laboratory uncertainty, the
laboratory should assess the systematic errors described in
Section 5.1 and add them to the estimated imprecision. As an
example, we present in Table 9 the total uncertainty estimate for

50
 .

the NBS liquid-in-glass calibration service. (We define the

uncertainty as the sum of the inaccuracy and three times the
imprecision)

TABLE 9

UNCERTAINTY ESTIMATE FOR NBS CALIBRATION SERVICE FOR

LIQUID-IN-GLASS THERM0^4ETERS

(Inaccuracy)
Temperature ( Imprec ision) Systemat ic Total
Range 3 Std. Dev. Error Uncertainty
(OC) (°C) (OC) (OC)

0-50 0 .018 0.010 0.028

50 - 100 0.024 0.010 0.034
0 - 100 0.027 0.010 0.037
100 - 200 0.084 0.015 - 0.030 0.099 - 0.114

This table merits several comments. The uncertainty of the

IPTS-68 is not included; the systematic error originates chiefly
from estimates of the operator's ability to reproduce the ice
point (0.005 ^C) temperature gradients in the baths (0.005 to
,

0.025), plus the thermometer resolution. Note too, that we have

tripled the standard deviation to insure that 99.7% of the
measurements will be included. Assessment of systematic errors
is a difficult task, and each laboratory will have to determine
systematic effects according to the prescription given in this
Monograph. This table is intended to provide an indication of
the performance at NBS; practice (and thus these quantities) may
differ at other laboratories.

6. CONCLDSION
In this monograph we have indicated the procedures by which
a laboratory may be equipped and staffed so that temperature
calibrations in the range 0 to 200 ^C may be performed accurately
within ±0.1 ^C or less. The temperature scale thus established
is shown to be traceable to the IPTS-68. Detailed procedures are
given for using liquid -in -glass thermometers as primary standards
and for deriving calibration corrections for a variety of
laboratory thermometers. A checklist for establishing a
calibration laboratory to the specifications described is
provided in Table 10.

51
 TABLE 10

CHECKLIST FOR ESTABLISHING A

STATE CALIBRATION LABORATORY

A. Range of Calibration: 0 to +200 ^C

B. Accuracy: ±0.1 ^C; traceable to NBS

C. Types of devices that can be calibrated:

1. Liquid-in-glass thermometers

a) total-immersion
b) partial-immersion
c) ASTM
d) calo rime trie
e) Beckmann
f) clinical standards
g) kinematic viscosity

2. Electronic temperature measuring systems

D. Laboratory Specifications:

1. Dimensions: 7m X 7m X 3m

2. Other requirements

a) Constant temperature: 23 ^C ± 2 degrees

b) Constant humidity: approximately 50%

E. Equipment:

1. Thermometer rack or map case to store thermometers

2. Microscope (15- or 2 0 -power)
3. Ice bath

a) tape-wrapped Dewar flask

b) thermometer clip
c) telescope (10-power with crosshair)
d) s iphon
e) soft mallet
f) shaved ice
g) freezer or chest for ice storage
h) distilled water storage tank
i) water distillation apparatus or distilled
water supply
j) plastic gloves
k) ice shaver

52
 Equipment: (continued)

4. Constant temperature baths (one oil bath may be

chosen to span the entire calibration range of
0 Oc to +200 OC)

a) Water bath (range: +1 to +95 ^C)

1) Sufficient to accommodate total-immersion

liquid-in-glass thermometers
2) Overflow provision
3) Heating/cooling coils isolated from
thermome ter s
4) Unrestricted flow for bath medium
5) Insulated
6) Thermometer holder designed to minimize heat
loss
7) Capable of maintaining bath at a constant
temperature 10.005 °C or a rise rate
of no more than 0.1 ^C in 3 minutes
8) Telescope (10-power with crosshair)

b) Oil bath (range: +95 ^C to +200 °C)

1) Same specifications as water bath

2) Oil
3) Hood vented to outside

F. Standards:

1. "Primary"

a) Two sets of total-immersion liquid-in-glass

thermometers (See Table 1) guaranteed to pass
NBS visual inspection
b) Calibrated by NBS or other traceable laboratory

2. "Check"

One or more sets of total-immersion liquid-in-glass

thermometers

G. Personnel Requirements:

1. At least a high-school graduate

2. Detail oriented
3. Adept in mathematics
4. Should attend NBS Precision Thermometry Seminar
5. Twelve months of practice and experience
6. Capable of calibrating up to 800 thermometers per year

53
7. ACKNOWLEDGMENTS

We would like to thank the following individuals for

assisting us in this effort.

Dr. Carroll S. Brickenkamp

Office of Weights and Measures, NBS

Mrs. M. Carroll Croarkin

Statistical Engineering Division, NBS

Mr. L. F. Eason
North Carolina Department of Agriculture
Consumer Standards Division

Mrs. Sylvia C. Ramboz

Temperature and Pressure Division, NBS

Dr. James F. Schooley

Temperature and Pressure Division, NBS

Mr. Charles R. Stockman

Maryland Department of Agriculture
Weights and Measures Section

Mrs. Ruth N. Varner

Statistical Engineering Division, NBS

Mr. Sidney Ween

Ever Ready Thermometer Co., Inc.

8. REFERENCES

1. Jacquelyn A. Wise, Liquid-in-Glass Thermometry. Nat.

Bur. Stand. (U.S.) Monogr. 150; January 1976. 30 pages.

2. "The International Practical Temperature Scale of 1968,

Amended Edition of 1975", Metrologia 12, No. 1, 7-17
(1976)

3. ASTM Designation E-1, Standard Specification for ASTM

Thermometers, Annual Book of ASTM Standards, Vol. 14.01.
4. ANSI Z236.1, Specifications for General Purpose
Laboratory Glass Thermometers.
5. ASTM Designation E-77, Standard Method for Verification
and Calibration of Liquid-in-Glass Thermometers, Annual
Book of ASTM Standards, Vol. 14.01.

54
9. COMMERCIAL SOURCES

9.1 Suppliers of Calibration Baths*

1) Hart Scientific
P. 0. Box 9 34
Provo, UT 84601

2) Willard L. Pearce Associates

Manufacturers Representative
Box 200
Allison Park, PA 15101

3) Rosemount Engineering Company

4900 West 78th Street
Minneapolis, MN 55435

4) Techne Incorporated
3700 Brunswick Pike
Princeton, NJ 08540

9.2 Suppliers of Oils for Calibration Baths*

1) Dow Corning
Chemical Products Division
Midland, MI 48640

2) E. I. Dupont DeNemours and Co.

Freon Products Division
Wilmington, DE 19899

3) Exxon Company, U.S.A.

P. 0. Box 1288
Baltimore, MD 21203

4) 3M Company
Commercial Chemicals Division
223-6S-04 3M Center
St. Paul, MN 55144-1000

55
9.3 Suppliers of Thermometers*

1) Brooklyn Thermometer Company, Inc.

90 Verdi Street
Farmingdale, NY 11735

2) Ever Ready Thermometer Company, Inc.

401 Park Avenue South
New York, NY 10016

3) H-B Instrument Co.

American and Bristol Sts.
Philadelphia, PA 19140

4) Walter H. Kessler Company, Inc.

160 Hicks Street
Westbury, Long Island, NY 11590

5) Miller and Weber, Inc.

1637 George Street
Ridgewood, Queens, NY 11385

6) Princo Instruments, Inc.

1020 Industrial Highway
Southampton, PA 18966

9.4 Supplier of Faden Thermometers*

Karl Schneider and Son

Am Bildacken 14
Postfach 13
Werthaim Main I
West Germany

*This list is not all-inclusive. It is intended to be only

representative of commercial products. Certain commercial
equipment, instruments, or materials are identified in this
paper in order to adequately specify the experimental
procedures. Such identification does not imply recommendation
or endorsement by the National Bureau of Standards, nor does it
imply that the materials or equipment identified are
necessarily the best available for the purpose.

56
10. APPENDIX

APPENDIX A-NBS Mongraph 150

I
 Contents
Page

1. Introduction _ 1

2. Thermometer Calibration Services at the NBS 1

2.1. Kinds of thermometers accepted for calibration 1
2.2. Preliminary examination 2
2.3. Reports of Calibration 2
2.4. General instructions to applicants for tests 2
a. Initial arrangements 2
b. Shipping instructions 5

3. Definitions 5

4. Temperature Scales and Standards 6

5. Calibration 8
5.1. Equipment 8
a. Ice Bath 8
b. Steam Bath 8
c. Comparison Liquid Baths 9
5.2. Determination of scale corrections 11
5.3. Corrections for emergent stem 13
a. Measurement of emergent-stem temperature 13
b. Formula for total-immersion thermometers 14
c. Formula for partial-immersion thermometers 15
d. Formula for calorimetric thermometers 16
e. Formula for Beckmann thermometers 16
5.4. Number and choice of test points 16

6. Common Thermometers and Factors Affecting Their Use 17

6.1. Total-immersion thermometers 17
6.2. Partial-immersion thermometers 17
6.3. Low-temperature thermometers 19
6.4. Beckmann thermometers — 19
6.5. Calorimetric thermometers 20

7. Thermometer Design 21
7.1. Materials of construction 21
7.2. Scale design and workmanship 21
7.3. Scale dimensions 22
7.4. Reference point 22
7.5. Marking of partial-immersion thermometers 22

8. Special Notes 23
8.1. Glass changes 23
a. Temporary changes 23
b. Permanent changes 24
8.2. Pressure effects 24
8.3. Lag 24
8.4. Separated columns 25

9. References 26

iii
 ,

Liquid-In-Glass Thermometry

Jacquelyn A. Wise

This Monograph, which supersedes NBS Monograph 90, contains information of general interest
to manufacturers and users of liquid-in-glass thermometers. Instructions explaining how to submit
a thermometer to the National Bureau of Standards for calibration are provided, and the techniques
£md equipment, such as stirred liquid comparison baths, used in the calibration procedures are
described. A discussion of important principles of acceptable thermometer design and factors affect-
ing their use is included. Listed are tables of tolerances reflecting good manufacturing practices
and reasonably attainable accuracies expected with liquid-in-glass thermometers. The calculation of
corrections for the temperature of the emergent stem is given in detail for various types of
thermometers and conditions of use.

Key words: Calibration; emergent stem; liquid-in-glass thermometer; reference point; stirred liquid
comparison bath; temperature scale.

1. Introduction

It the responsibility of the 'National Bureau of

is This Monograph is intended for those who may
Standards NBS) to accurately reproduce the Inter-
( wish to submit thermometers for calibration or who
national Practical Temperature Scale of 1968 [1]^ as desire information on the use of liquid-in-glass
a basis for the uniform measurement of temperature thermometers for precise temperature measurements.
throughout the scientific and industrial community of Practices employed at the NBS in the calibration of
the United States. Because of this responsibility, NBS these thermometers are discussed. Also included is a
accepts with reference to this scale
for calibration brief discussion of the International Practical Tempera-
selected types of temperature-measuring instruments ture Scale of 1968, a description of the calibration
[2] for use as reference or working standards where equipment, and the accuracy capability of liquid-in-
precise temp:^- p.ture measurements are required. Cer- glass thermometers. Details of good thermometer design
tain liquid-in-glass thermometers are among the and factors affecting their use are presented in a
instruments accepted. copious manner.

2. Thermometer Calibration Services

at the NBS

The thermometer is one of many

liquid-in-glass ship, and capable of being placed in the existing cali-
precision instruments calibrated by the NBS. A com- bration facilities. The type known as laboratory or
plete description of calibration services offered by chemical thermometers, consisting of a solid stem
the NBS appears in NBS Special Publication 250 [2] with graduation lines and numbers permanently
which can be obtained by writing to the Office of etched or placed on the stem, are regularly accepted.
Technical Information and Publications, National Others approved for calibration are Beckmann, cal-
Bureau of Standards, Washington. D. C. 20234. orimetric. and enclosed scale (Einschluss) laboratory
Supplements to NBS Special Publication 250 are issued thermometers. Industrial or mechanical type ther-
periodically indicating any changes in these services. mometers with special mountings can be accepted for
test if they can be inserted in the comparison baths,

2.1. Kinds of Thermometers Accepted for or if the thermometer can be easily detached from the
Calibration mounting.
Any thermometer, such as a household thermometer,
Not all thermometers classified as liquid-in-glass consisting of a pa])er, plastic or metal mounting con-
are acceptable for calibration by the NBS. In general taining the scale graduations and attached to the
they must be of an acceptable design and workman- unmarked glass thermometer by metal clips, is not
acceptable for calibration. Maximum self-registering
mercury-in-glass clinical thermometers are no longer
'Figures in brackets indicate the literature references at the end of this
Monograph.
tested at the NBS.
 It advisable to contact the liquid-in-glass ther-
is on the thermometer, the NBS test number and com-
mometry laboratory at NBS if there is any doubt pletion date, and explanatory notes defining the
concerning the acceptability of a thermometer for conditions under which the results of the test are
Every effort will be made to assist and
calibration. applicable. When necessary, accompanying the Report
satisfy needs of the scientific and industrial
the of Calibration is a sheet explaining how to calculate
community with problems involving temperature the corrections for emergent stem. If the thermometer
measurement. is of the metastatic (Beckmann) type, the report will
be accompanied by a second sheet containing a table
of setting factors (See Sec. 6.4) enabling the user to
2.2. Preliminary Examination apply the calibration results for settings other than
Every thermometer submitted for test must be the 20 °C setting for which the corrections are re-

uniquely identified by a serial number and pass a ported. A facsimile of a Report of Calibration is
preliminary examination before final acceptance. These shown in figures 1 and 2.

thermometers are viewed under a microscope having a Any departure from the conditions under which
magnification of 15 or 20X to assure that the mercury the corrections were obtained, as stated on the Report
and capillary are clean. Any foreign matter found in of Calibration, may significantly change the values of
the capillary or bulb that, in the judgment of the the corrections. Conditions of immersion are particu-
laboratory personnel, might tend to make the ther- larly important (Sec. 6). It should be emphasized
mometer readings erratic, is reason for rejection. that the estimates of error assigned to the scale cor-
Glass chips or oxides of mercury in the capillary rections do not assure the user of this accuracy in a
or bulb are considered foreign matter that can cause temperature measurement.
the thermometer to indicate different readings at
different periods of time. [For example, if a small
chip of glass is located in the capillary at the 30 °C 2.4. General Instructions to Applicants
indication, each time the mercury column advances for Tests
past this point, it could cause the reading of the ther-
Testing will be conducted in accordance with the
mometer to vary for temperatures above 30 °C, because policies of NBS as described in NBS Special Publi-
of different quantities of gas that may be trapped
cation 250. The cost of calibration will depend on
around the chip. The movement of the chip to different the number of calibration points requested, the tem-
locations in the capillary, or the trapping of mercury
perature range, and the quantity of thermometers
around the chip when the mercury column recedes, submitted. For tests not sf)ecifically outlined in the
would also cause erroneous readings.]
above publication, NBS should be consulted. If th?
Other reasons a thermometer may be ineligible for
required measurements appear feasible, and, in the
test are given below:
opinion of NBS, sufficiently important to justify the
(a) Defective design or workmanship. work, such tests will be undertaken for a special fee
(b) Part of graduated scale not usable. determined by the nature of the work and time in-
(c) Errors in scale graduation or numbering. volved. In all requests for test the following procedures
(d) Unsuitable bulb glass or inadequate and information are pertinent,
annealing.
(e) Inadequate gas filling. a. Initial Arrangements
(f ) Cracks in the glass.
A or purchase order requesting the test must
letter
A complete list of all possible causes for rejection be sent to NBS either with the thermometer or sep-
is not feasible. The prime consideration in the pre-
arately. No work can be undertaken unless both the
liminary examination is that the thermometer be thermometer and letter or purchase order have been
capable of precise, reproducible readings. received. Information in the request for test should
include the quantity and serial numbers of the ther-
mometers sent, the calibration points required, (See
2.3. Reports of Calibration
Sec. 5.4) and the suggested method for return ship-
In most instances a Report of Calibration will be ment. If it is desired that the thermometer be insured
issued by NBS for every liquid-in-glass thermometer upon return, a statement to this effect must appear
submitted for test that is found JFree from serious on the request, and a value assigned; otherwise, the
defects as determined by the preliminary examination. thermometer will be returned uninsured.
In order to receive a Report of Calibration, the ther- NBS number to each calibration
will assign a test
mometer must be calibrated at two or more calibration request and acknowledge the receipt of the ther-
points. If only one point is requested, a Report of mometer. On the acknowledgment will be stated the
Test will be issued. estimated coippletion date and estimated cost of
The Report of Calibration contains the corrections calibration. Any foreseeable difficulty that may be
determined for each point requested and an estimate encountered involving the calibration will be men-
of the uncertainties associated with the corrections. tioned on the acknowledgment, as well as the name
Also stated on the document are the agency or firm of someone in the testing laboratory who can be
requesting the test, the trade mark and serial number contacted if questions arise.
 U.S. DEPARTMENT OF COMMERCE
NATIONAL BUREAU OF STANDARDS
INSTITUTE FOR BASIC STANDARDS
WASHINGTON, O.C. 20234

REPORT OF CALIBRATION
L IQUIO-IN-GLASS THERMOMETER

TESTED FOR: NATIONAL BUREAU OF STANDARDS

DIVISION 221, SECTION II
MARKED: SURETY 198692
RANGE: -2 TO +102 DEGREES C IN 0.2 DEGREE

THERMOMETER CORRECTION
READING (IPTS-68)**
-.06 C .06 C
20.00 .13
40.00 .05
60.00 .04
80.00 .06
100.00 .04

**ALL TEMPERATURES IN THIS PFPORT ARE BASED ON THE INTERNATIONAL

PRACTICAL TEMPERATURE SCALE OF 1968, IPTS-68. THIS TEMPERATURE SCALE
WAS ADOPTED BY THE INTERNATIONAL COMMITTEE OF WEIGHTS AND MEASURES AT
ITS MEETING IN OCTOBER, 1968, AND IS DESCRIBED IN "THE INTERNATIONAL
PRACTICAL TEMPERATURE SCALE OF 1968, METROLOGIA, VOL. 5, NO. 2, 35
(APRIL 1969).

ESTIMATED UNCERTAINTIES IN THE ABOVE CORRECTIONS 00 NOT EXCEED 0.05

DEGREE UP TO 102 DEGREES C .

FOR A DISCUSSION OF ACCUFACIES ATTAINABLE WITH SUCH THERMOMETERS SEE

NATIONAL BUREAU OF STANDARDS MONOGRAPH 150, L I QUI D-I N-GLASS
THERMOMETRY .
IF NO SIGN IS GIVEN ON THE CORRECTION, THE TRUE TEMPERATURE IS HIGHER
THAN THE INDICATED TEMPERATURE; IF THE SIGN GIVEN IS NEGATIVE, THE TRUE
TEMPERATURE IS LOWER THAN THE INDICATED TEMPERATURE. TO USE THE CORREC-
TIONS PROPERLY, REFERENCE SHOULD BE MADE TO THE NOTES GIVEN BELOW.

THE TABULATED CORRECTIONS APPLY FOR THE CONDITION OF TOTAL IMMERSION OF

THE BULB AND LIQUID COLUMN. IF THE THERMOMETER IS USED AT PARTIAL IMMER-
SION, APPLY AN EMERGENT STEM CORRECTION AS EXPLAINED IN THE ACCOMPANYING
STEM CORRECTION SHEET.

TEST NUMBER 311-11-74

COMPLETED 5-9-74

Figure 1. Facsimile of page 1 of a Report of Calibration.

3
 PAGE 2

THE TABULATED COPRECTIUNS APPLY PROVIDED THE ICE-POINT HEADING, TAKEN

AFTER EXPOSURE FOR NOT LESS THAN 3 DAYS TO A TEMPERATURE OF ABOUT 23
DEGREES C (73 DEGREES F) , IS -0.06 C . IF THE ICE-POINT READING IS FOUND
TO BE HIGHER (OR LOWER) THAN STATED, ALL OTHER READINGS WILL 8E HIGHER
(OR LOWFR) BY THE SAME AMOUNT. IF THE THERMOMETER IS USED AT A GIVEN
TEMPERATURE SHORTLY AFTER BEING HEATED TO A HIGHER TEMPERATURE, AN ERROR
OF 0.01 DEGREE OR LESS, FOR EACH lO-OEGREE DIFFERENCE BETWEEN THE TWO
,

TEMPERATURES, MAY BF INTRODUCED. THE TABULATED CORRECTIONS APPLY IF THE

THERMOMETER IS USED IN ITS UPRIGHT POSITION; IF USED IN A HORIZONTAL
POSITION, THE INDICATIONS MAY BE A FEW HUNDREDTHS OF A DEGREE HIGHER.

FOR THE DIRECTOR,

INSTITUTE FOR BA/SIC STANDARDS

TEST NUMBER 311-11-74

COMPLETED 5-9-74
CHIEF, TEMPERATUP SECTION
HEAT DIVISION

Figure 2. Facsimile of page 2 of a Report of Calibration.

4
 b. Shipping Instructions mitting them, and, if applicable, the company's
purchase order number.
To assure safe arrival of the thermometers, they
Unless otherwise specified, the thermometers will
should be packed firmly in their individual cases.
be returned via the best method in the judgment of
A rigid material, such as wooden slats or corrugated
paper, can be wrapped around^ the thermometer case
NBS. Shipping charges, both to and from the NBS,
will be paid by the applicant. All possible care will
as an additional aid against breakage. The package
should contain enough packing material to absorb any be taken in handling thermometers at NBS, but the
shock that it may receive. Included with each ship- risk of damage either in shipment or testing must be
ment should be a packing list stating the number of assumed by the agency, firm or individual submitting
thermometers shipped, the name of the company sub- them.

3. Definitions

The principle features of a solid-stem liquid-in-glass provided to prevent the buildup of excessive pressures
thermometer are shown in figure 3. Not all of the in gas-filled thermometers as the liquid filling ad-
features shown would necessarily be incorporated in vances toward the top of the scale (see Sec. 7.1).
any one thermometer. Contraction Chamber: An enlargement of the ca-
pillary bore which serves to reduce a long length of
CONTRACTION EXPANSION
CHAMBER CHAMBER capillary or to prevent contraction of the entire liquid
column into the bulb. This chamber is introduced
below the main scale or between the main scale and
AUXILIARY_
SCALE
IMMERSION
LINE
J MAIN
SCALE
an auxiliary scale.
Reference Point: A reference temperature, such as
Figure 3. Principle features of a solid-stem liquid-in- glass
the ice point or steam point, used periodically to
thermometer. check the thermometer for changes in bulb volume.
(See Sec. 5.2). The reference point can be included
Bulb: The reservoir for the thermometer liquid. in the main scale or on an auxiliary scale.

The bulb of a thermometer will contain a volume Total-Immersion Thermometer : A thermometer de-
equivalent to a specific number of degrees of the scale signed to indicate temperatures correctly when the
depending upon the coefficients of expansion of the bulb and the entire liquid column is exposed to the
thermometric liquid and bulb glass. For mercury in temperature being measured. (See the definition for
a bulb made of "normal" glass, the bulb volume is complete-immersion thermometer.)
equivalent to approximately 6,222 times the volume Partial-Immersion Thermometer: A thermometer
of a 1 degree length of the capillary on the scale for designed to indicate temperatures correctly when the
Celsius thermometers, or 11,200 times the volume of bulb and a specified portion of the stem is exposed
a 1 degree length of the capillary on the scale for to the temperature being measured. The remaining
Fahrenheit thermometers. For organic thermometric portion of the stem, referred to as the emergent stem,
liquids with higher coefficients of expansion than will be at the ambient temperature, usually different
mercury, the bulb volumes are correspondingly less. from the temperature being measured. Such thermom-
Stem: The glass capillary tube through which the eters are generally marked with an immersion line to
thermometric liquid advances or retreats with changes indicate the proper depth of immersion.
in temperature. Complete-Immersion Thermometer: A thermometer
Main Scale: The scale graduated in degrees or designed to indicate temperatures correctly when the
multiples or submultiples of degrees; in many in- whole thermometer, including the expansion chamber,
stances the main scale constitutes the only scale. is subjected to the temperature being measured. In gas-

Auxiliary Scale: A short scale including a reference filled thermometers the reading will be different for
temperature such as the ice point, to provide a means complete, as compared to total immersion, as a result
for checking the therm.ometer for a change in cali- of the effect of temperature on the gas pressure in
bration with time (see Sec. 7.4). This scale is added the thermometer (see Sec. 8.2). Although the differ-
when a suitable reference temperature is not included ence in readings under the two conditions is particu-
in the range of the main scale. larly significant at high temperatures, it is also
Expansion Chamber: An enlargement at the top significant at moderate temperatures if the bulb and
end of the capillary bore having a volume equivalent expansion chamber are both relatively large.
to not less than the volume of a 20 mm
length of un- Calibration Points: The temperatures on the ther-
changed capillary. Smaller chambers are not regarded mometer scale (i.e., 0°C) where calibrations are
as expansion chambers. The expansion chamber is performed and corrections reported.

5
 Accuracy: The accuracy of a measurement refers Instability of the thermometer glass, capillary forces
to the difference between the measured value and the at thesurface of the thermometric liquid, non-
true value of the quantity being measured. In this uniformity of capillary bore, and inaccuracies in scale
Monograph the accuracy of a given thermometer refers graduation are among the important factors. With
to its ability to indicate temperatures correctly on the partial-immersion thermometers, uncertainties in cor-
International Practical Temperature Scale of 1968 [1] rections for the emergent stem may greatly limit the
within the uncertainty stated (See tables 5-12) provid- accuracy. Observer errors are also involved, but with
ed all corrections are applied and the thermometer is care these can usually be made relatively small.
used in the same manner as when it was calibrated. Precision: The precision of measurement refers to
The accuracy attainable is principally limited by the the degree of agreement amongst repeated measure-
characteristics of the thermometer itself (See Sec. 6). ments at a given time of the same quantity.

4. Temperature Scales and Standards

The scale to which measurements of temperature The units of Tea and t^a, like the thermodynamic
should ultimately be referred is the Thermodynamic scales, are the Kelvin, symbol K, and degree Celsius,
Kelvin Temperature Scale (TKTS). Values of tem- symbol °C.
perature expressed on the TKTS are designated by Due more sophisticated equipment and
to the use of
the symbol T. The unit of temperature is the Kelvin, more accuracy the defining
the ability to measure with
symbol K, which is a base unit of the Systeme Inter- fixed points, the new temperature scale more closely
nationale (SI), and is defined as "the fraction agrees with thermodynamic temperatures. This scale,
1/273.16 of the thermodynamic temperature of the IPTS-68, is intended to provide scientific and indus-
triple point of water" [3] trial laboratories throughout the world with a com-
Because of the difficulties that are encountered in mon basis for stating temperatures. Calibrations of
the practical realization of the TKTS, it has been thermometers at NBS, therefore, are made with ref-
necessary to define and utilize practical temperature erence to values of temperature on the IPTS-68.
scales. The International Temperature Scale and de- Shown in Table 1 is the approximate difference in
fining text was first adopted in 1927 and later revised degrees Celsius between the values of temperature
in 1948. The Eleventh General Conference of Weights given by the IPTS-68 and the IPTS-48 [1] The differ- .

and Measures 1960 changed the name of the scale

in ences are listed in two separate sections for convenience
to the International Practical Temperature Scale and clarity. It can be seen, by studying these values,
(IPTS) of 1948 and adopted a revised text of the that the correction is important if quality control in
scale [4] ,
although numerically the scale was not the laboratory is to be maintained to the nearest
changed. In 1968, in accordance with the power given 0.01 °C. If, on the other hand, accuracies no better
to it by Resolution 8 of the Thirteenth General Con- than the nearest 0.1 °C are desired, the correction
ference of Weights and Measures, the International becomes less significant in the liquid-in-glass ther-
Committee of Weights and Measures adopted the Inter- mometer region (—200 to +600 °C) of temperature
national Practical Temperature Scale of 1968 [1], measurements.
henceforth referred to as IPTS-68. The IPTS-68 re- Although values of temperatures on the IPTS-68
placed the International Practical Temperature Scale are expressed in degrees Celsius, thermometers that
of 1948 and may be expressed either as a Kelvin scale are graduated on the Fahrenheit Scale can be cali-
or as a Celsius scale. brated with reference to the IPTS-68 by using the
The present Celsius scale has its zero 0.01 below K conversion formula:
the triple point of water (essentially the ice point).
The value of a temperature expressed on the Thermo- temperature value in °F = 9/5
dynamic Celsius Temperature Scale is designated by (temperature value in °C) + 32
the symbol t and is related to the value on the TKTS by
The National Bureau of Standards has been per-
i = r - 273.15 K. forming calibrations with reference to the IPTS-68
since July, 1969. On all Reports of Calibration and
The unit of the Celsius Scale is the degree Celsius,
Reports of Test issued after this date, it has been
symbol °C, which by definition is equal in magnitude
clearly indicated that the new scale was used.
to the Kelvin.
In the range of temperatures normally covered by
Values of temperature expressed on the Interna-
liquid-in-glass thermometry, the IPTS-68 is defined
tional Practical Kelvin Scale are designated by the
by four fixed points: the equilibrium between the
symbol T^a, and values of temperatures expressed on
liquid and vapor phases of oxygen (normal boiling
the International Practical Celsius Temperature Scale
point of oxygen) at —182.962 °C; the equilibrium
are designated by the symbol ios- The relationship
between solid, liquid, and vapor phases of water
between Tug and ies is
(triple point of water) at +0.01 °C; the equilibrium

tas — Tes —273.15 K. between the liquid and vapor phases of water (nor-

6
 Table 1. — Approximate difference ( tss — tis), in degrees Celsius,

between the values of temperature given by the IPTS—68 and IPTS-^8

la. For the range — 180°CtoO°C.

Q -10 -20 -30 -40 -50 -60 —70 — o\j on inn

-100 0.022 0.013 0.003 -0.006 -0.013 -0.013 -0.005 0.007 0.012
- 0 0.000 0.006 0.012 0.018 0.024 0.029 0.032 0.034 0.033 0.029 0.022

lb. For the range 0 °C to 1070 °C.

tts °C 0 10 20 30 40 50 60 70 80 90 100

0 0.000 -0.004 -0.007 -0.009 -0.010 -0.010 -0.010 -0.008 -0.006 -0.003 0.000
100 0.000 0.004 0.007 0.012 0.016 0,020 0.025 0.029 0.034 0.038 0.043
200 0.043 0.047 0.051 0.054 0.058 0.061 0.064 0.067 0.069 0.071 0.073
300 0.073 0.074 0.075 0.076 0.077 0.077 0.077 0.077 0.077 0.076 0.076
400 0.076 0.075 0.075 0.075 0.074 0.074 0.074 0.075 0.076 0.077 0.079
500 0.079 0.082 0.085 0.089 0.094 0.100 0.108 0.116 0.126 0.137 0.150
600 0.150 0.165 0.182 0.200 0.23 0.25 0.28 0.31 0.34 0.36 0.39
700 0.39 0.42 0.45 0.47 0.50 0.53 0.56 0.58 0.61 0.64 0.67
800 0.67 0.70 0.72 0.75 0.78 0.81 0.84 0.87 0.89 0.92 0.95
900 0.95 0.98 1.01 1.04 1.07 1.10 1.12 1.15 1.18 1.21 1.24
1000 1.24 1.27 1.30 1.33 1.36 1.39 1.42 1.44

Examples showing the use of these tables are given below to help clarify any difficulty that may be encountered.

I Standard thermometer reading ~36.56 °C

Correction fromNBS Report of Calibration + .06
True temperature IPTS-45 "36.50 °C
Correction from the above table at ~ 36.50 °C + -02
True temperature IPTS-68 -36.48 °C

II Standard thermometer reading 98.32 °F

Correction from NBS Report of Calibration ~ .02

True temperature IPTS-48 98.30 °F

Converted correction from the above tables* ~ .02

True temperature IPTS-68 98.28 °F

*The converted correction is obtained by:

A. Convert the Fahrenheit temperature to Celsius 98.30 °F 36.83 °C. =

B. Find the correction from the table at 36.83 °C. It is —0.01 °C.
C. Convert this correction to Fahrenheit by multiplying bv 9/5.
-0.01 °C X 9/5 « -0.02 °F.

mal boiling point of water at 100 °C: and the equi-

) low may be used as standards for total-irmnersion
librium between solid and liquid phases of zinc comparisons.
(freezing point of zinc) at 419.58 ^C. Temperatures
in the range —182.96 to 630.7 are defined in
terms of a standard platinum resistance thermometer Smallest Auxiliary
that has been calibrated at the four fixed points, with Range graduation scale

interpolations and extrapolations between and beyond

these points accomplished by employing specified °C °C "C
equations [1]. 0 to 50 0.1
0 to 100 0.2
When the highest accuracy is required in a calibra-
50 to 100 0.1 atO
tion, the thermometer indications are compared di- 100 to 200 0.2 atO
rectly with temperatures obtained using a standard 200 to 300 0.5 atO
platinum resistance thermometer. At NBS all calibra-
tions are performed in this manner, with the exception
of the ice point See Sec. 5.2
I If a lower accuracy is
I . Partial-immersion thermometers, known as "like
required, one of several mercury -in-glass thermometers standards," can be maintained for the calibration of
with ranges and graduated in the intervals li.?ted be- thermometers manufactured to similar specifications.

7
These standards are calibrated for stem-temperature the bulbs are at the same temperature and the stem
conditions expected to prevail during the use of these temperatures are essentially the same for all of the
thermometers or for specified stem temperatures. This thermometers under comparison.
use of like standards eliminates the need for many of All of these liquid-in-glass standards should be
the precautions necessary when dissimilar thermom- calibrated with reference to the IPTS-68 through
eters are compared and permits the direct comparison comparisons with a standard platinum resistance
of the indications of similar thermometers as long as thermometer.

5. Calibration

In most instances, the determination of the liquid- circulation of steam around the thermometers sus-
in-glassthermometer scale corrections is accomplished pended within this space. A
provision is made both
by comparing its scale indications with a known tem- for relieving any excess pressure in the space sur-
perature from a platinum resistance thermometer or rounding the thermometers and for determining the
mercury-in-glass thermometer standards. This com- excess pressure by means of a small differential
parison is achieved by placing the standard and the manometer.
thermometers under test in a series of stirred liquid
comparison baths (See Sec. 5.1c). A discussion of
the equipment needed, the procedure followed, addi-
tional corrections that may be necessary, and the
choice of calibration points is given in the following
sections.

5,1. Equipment
THERMOMETER
THERMOMETER
a. Ice Bath SPACE —
IN
STEAM JACKET AIRJACKET
An can be easily assembled and consists
ice bath
of a container, a siphon tube, ice and distilled water.
A Dewar flask, approximately 36 cm deep and 8 cm
in diameter, can serve as a container for the ice.
A vessel of this type is preferable, since the melting
of the ice is retarded by the insulating properties of
the Dewar flask. A siphon is placed in the flask to
enable excess water to be removed as the ice melts.
The clear or transparent portion of commercially
purchased ice. or ice made from distilled water, can
be used. The ice is shaved or crushed into small chips CONDENSER
measuring 2 to 5 mm. The flask is one-third filled
with distilled water and the shaved ice is added.
This mixture is compressed to form a tightly packed
slush and any excess water is siphoned off. Before BOILER
the bath is used, adequate time 15 to 30 min.) should
(

be given for the mixture to reach a constant tem-

perature throughout.
Ideally there should be as much ice in the flask as
possible, with the small spaces between the chips filled
with distilled water. It will be necessary periodically
to add ice and to remove the excess water while the TO HEATING COIL
bath is being u^ed to maintain this ideal consistency.
Fk.i hk 4. Srhcmalic draiiing of iteaiii balh.
If care is taken to prevent contamination of the ice
and water, the ice point can be realized to better than
0.01 '^C by this means.
When the steam bath is used as a fixed-point ap-
jjaratus, barometer is a necessary accessory, since
a
the temperature of the steam depends upon the pre-
b. Steam Hath
vailing atmospheric pressure. The usual adjustments
A schematic drawing of a steam bath is shown in are made to the barometer reading, including any
figure 1. Steam produced from the boiler circulates necessary corrections for the value of local accelera-
within a double-walled steam jacket permitting free tionOf gravity, for the difference in height between

o
the steam bath and the barometer, and for any excess c. Comparison Liquid Baths
pressure above atmospheric in the steam jacket. After
At NBS two types of stirred liquid baths are used
the corrected pressure reading is obtained, the tem-
for comparison calibrations in the range —110 to
perature of the steam can be derived from the values
given in Table 2. With a barometer accurate to 0.1 540 °C (-166 to 1004 °F). Each is equipped with a
mm Hg. this procedure is capable of an accuracy of stirring unit to provide a uniform temperature
0.002 to 0.003 (0.004 to 0.005 °F). The Fortin throughout the medium and a controlled current to
type barometer will usually suffice for all but the most the heating coils for proper temperature regulation.
exacting measurements. A 5 to 7 cm thickness of insulation surrounding the
The steam bath can also be used as a comparison
baths and an insulated cover is provided to help mini-
bath, with the temperature of the steam being deter-
mize heat loss. Fitted into the top cover is a holder
mined at the time of test by means of a previously
standardized thermometer. This method, which does containing the thermometers. This holder can be ro-
not require the use of a barometer, may be preferable, tated, enabling each thermometer to appear in the

particularlywhen a platinum resistance thermometer field 01 view of a vertically adjustable telescope at-

can be used as the standard. tached to each bath.

Table 2. — (Thermometric) condensation temperature of steam [5]

[Asterisk (*) indicates change in integer]

Pressure in mm mercury (standard)

p 0 1 2 3 4 5 6 7 8 9

Temperature in degrees of International Practical Temperature Scale of 1968

*
500 88.674 0.726 0.778 0.830 0.882 0.935 0.987 *0.039 * 0.090 0.141
510 89.193 .244 .295 .347 .398 .449 .499 .550 .601 .652
520 89.702 .753 .803 .853 .904 .954 * .004 * .054 * .104 .154
530 90.203 .253 .303 .352 .402 .451 .500 .550 .599 .648
* * .086 *
540 90.697 .746 .795 .843 .892 .941 .989 .038 .135

550 91.183 .231 .279 .327 .375 .423 .471 .518 .566 .614
* *
560 91.661 .709 .756 .803 .851 .898 .945 .992 .039 .086
570 92.133 .179 .226 .274 .320 .367 .413 .460 .506 .552
*
580 92.598 .644 .690 .736 .782 .828 .874 .920 .965 .011
590 93.056 .102 .147 .193 .238 .283 .328 .373 .418 .463

600 93.5079 .5527 .5975 .6422 .6868 .7315 .7760 .8204 .8645 .9092
* * * * * * * *
610 93.9534 .9976 .0419 .0860 .1300 .1740 .2179 .2617 .3055 .3493
620 94.39.30 .4366 .4802 .5237 .5671 .6105 .6538 .6972 .7404 .7835
* * * * *
630 94.8266 .8696 .9126 .9555 .9985 .0413 .0841 .1268 .1694 .2120
640 95.2546 .2972 .3396 .3819 .4242 .4665 .5087 .5508 .5929 .6351

* *
650 95.6771 .7190 .7609 .8027 .8445 .8862 .9280 .9696 .0112 .0526
660 96.0941 .1355 .1769 .2182 .2595 .3007 .3419 .3830 .4240 .4650
670 96.5060 .5469 .5878 .6286 .6693 .7100 .7506 .7912 .8318 .8724
* * .1143 * * .1947 * .2347 *
680 %.9128 .9532 .9936 * .0339 .0741 .1546 .2747
690 97.3147 .3546 .3945 .4343 .4741 .5138 .5535 .5931 .6327 .6722

* *
700 97.7117 .7511 .7906 .8300 .8693 .9085 .9477 .9869 .0260 .0651
710 98.1042 .1431 .1821 .2210 .2598 .2986 .3373 .3761 .4148 .4534
720 98.4920 ..5305 .5690 .6074 .6458 .6842 .7225 .7608 .7990 .8372
* .0275 * * * * *
730 98.8753 .9134 .9515 .9896 .0654 .1033 .1411 .1789 .2107
740 99.2544 .2922 .3298 .3673 .4049 .4424 .4798 .5172 .5546 .5920

750 99.6293 .6666 .7038 .7409 .7780 .8151 .8522 .8893 .9262 .9631
760 100.0000 .0368 .0736 .1104 .1472 .1839 .2205 .2571 .2937 .3302
770 100..3667 .4031 .4395 .4760 .5123 .5486 .5845 .6210 .6572 .6934
* *
780 100.7295 .7656 .8016 .8376 .8736 .9095 .9453 .9811 .0169 .0528
790 101.0885 .1242 .1598 .1954 .2310 .2665 .3020 .3376 .3730 .4084

9
 The type of bath shawn in figure 5 is suitable when
the medium is liquid atroom temperature. It consists
of two cylindrical wells of different diameters with
connecting passages top and bottom. The heating
at the
coil and stirrer are located in the smaller well, leaving
the larger well unobstructed for the insertion of the
thermometers. The fluid is forced past the heating
coils, through the bottom opening into the larger well,
around the thermometers, and back into the smaller
well through the connecting passage at the top.

STIR DRIVE

THERMOMETER INSULATION
HOLDER

TESTING AREA

INSULATION AIR SPACE

HEATING —
COILS HEATING COIL

AIR SPACE
SPACE
BETWEEN
TUBES

BEARING

Figure 6. Stirred high temperature bath using liquid having

freezing point above room temperature.

baths are designed to shield the thermometers from

direct radiation from the heating units.
SUPPORT
For calibrations in the range 1 to 99 °C (34 to
210 "F) water is used as the bath liquid. One grade
Figure 5. Stirred liquid baih. of petroleum oil is used between TOO and 200 °C
(212 to 392 ^F) and a second grade between 200 and
The bath illustrated in figure 6 is designed for use 300 (392 and 572 °F). The oils must not have
with a medium that is solid at room temperature. A flash points below the highest test temperatures, nor
bath of this type is used at NBS for high temperature be too viscous to prevent adequate stirring at the
calibrations. Two coaxial cylinders are arranged in lower test temperatures.
such a way as to permit the medium to circulate be- Calibrations from 1 to -110 - (30 to -166 °F)
tween the walls of the two cylinders and through the are made in a cryostat similar to the one described
inner cylinder by means of openings at the top and by Scott and Brickwedde [6]. The cryostat, shown in
bottom. The stirring propeller is situated near the figure 7. consists of an inner Dewar flask, D, which
bottom of the inner cylinder, leaving the majority of contains the bath liquid. This flask is surrounded by
the space for the reentrant tubes into which the ther- liquid nitrogen contained in the outer Dewar flask, C.
mometers are inserted. The heater coils are wound The rate of heat transfer between the bath liquid and
on the outside of the outer cylinder. Both types of the liquid nitrogen is retarded by evacuating the space

10
 between the walls of the inner Dewar flask through
the side tube. M, which is connected to a vacuum
system. Vigorous stirring of the bath liquid is main-
tained by the propeller, I, which circulates the liquid
around the walls of the stirrer tube, P, similar to the
flow of the medium in the high temperature bath.
The temperature of the bath is thermostatically con-
trolled by heater coils, J. wound on the outside of the
stirrer tube. The thermometers are immersed inside
the stirrer tube, thus shielding them from the heater
coils.

The bath liquid is a five-component mixture con-

taining by weight 14.5 percent of chloroform, 25.3
percent of methylene chloride. 33.4 percent of ethyl
bromide. 10.4 percent of transdichloroelhylene. and
16.4 percent of trichloroethylene. This mixture freezes
at approximately -150 °C (-238 °F), but readily
absorbs moisture and becomes cloudy at somewhat
higher temperatures. For this reason calibrations
are not performed in this bath below 110°C —
(-166 °F).
A comparison measurement can be made at approxi-
mately —
196 °C (boiling point of nitrogen) and
— 183 °C (boiling point of oxygen). A silvered Dewar
flask with a narrow transparent vertical strip is used
as a container for the liquid nitrogen or oxygen. The
liquid is agitated by bubbling nitrogen or oxygen gas
in the corresponding liquid through a glass tube with
an outlet near the bottom of the flask.

5.2. Determination of Scale Corrections

Thermometers submitted to NBS for test are gen-

by starting at the lowest test point
erally calibrated
requested and advancing to the higher test points.
In most cases the ice point is the lowest point. Ice is

packed firmly around the thermometer which has been

placed in an ice bath and immersed to the proper
depth as determined by the type of thermometer. The
thermometer is gently tapped before reading to prevent
J
the sticking of a falling meniscus. Care is taken not
to tap too vigorously, as this may cause the mercury
to rebound to an erroneously high reading.
The thermometer scale corrections quoted in the
Report of Calibration apply as long as the ice-point
reading remains the same as observed during the NBS
calibration. Subsequent changes in the ice-point read-
ing of the thermometer will result from small changes
in the glass of the thermometer bulb which affects its
volume. The volume of the capillary also changes, but
the volume of mercury contained in the stem is so
small, in comparison to the amount of mercury in the
bulb, that changes in the stem volume can usually be
ignored. As a result, changes in the ice-point reading
of the thermometer, taken after an exposure of not less
than 3 days at a temperature of approximately 23 °C
(73 °F), will be reflected by similar changes in read-
ings at each point along the scale. Therefore, when
the correction at the ice point is found to be higher
(or lower) than that observed at the time of calibra-
Figure 7. Vertical section of cryostal. tion, the other reported scale corrections will be higher

11
for lower) by the same amount. An example is given Ice-point readings on the thermometers being cali-
below: brated are not taken after each test point on the
scale. For thermometers graduated below 300 °C or
Original Corrections from NBS Report 600 °F, ice points are taken before the comparison
calibrations and 3 days (with the thermometer at room
Thermometer
Reading Correction temperature) after the last test point to assure that
the bulb has recovered within acceptable limits. With
+ 0.011 °C -0.011 °C
thermometers graduated above 300 °C or 600 °F, an
10.000 - .015
20.000 - .020
ice point is taken and it is immediately tested at the
30.000 + .008 highest test point on the scale. After a rest period of
40.000 - .033 3 days at room temperature a second ice point is
50.000 .000 taken. The agreement between the first and second
ice point must be within the stated accuracy limit of
A later ice-point reading may be +0.019 °C. This the thermometer, or it is deemed unsuitable for cali-
means that the ice-point correction will be —0.019 °C bration and no further testing is done.
and a new set of corrections should be made by sub- The ice point is the only test temperature taken in
tracting 0.008 from all of the original corrections. the manner described above. Comparison measure-
The new table would look as follows: ments are made with the temperature of the bath
slowly increasing at a steady rate, which does not
Thermometer
Reading Correction
exceed one scale division in 3 to 10 minutes [7]. All
thermometers are read using a 10 power telescope,
+ 0.019 °C -0.019 °C which is attached to each calibration bath and per-
10.000 - .023
20.000 - pendicular to the thermometers. The telescope is a
.028
30.000 .000 necessity if errors due to parallax are to be avoided.
40.000 - .041 When a liquid-in-glass thermometer is intended to
50.000 - .008
be used as a standard, an ice point should be taken
after heating it at each test temperature. The adjusted
Once a thermometer has been calibrated, an ice point scale correction is determined by adding the ice-point
check is all that is necessary to obtain a current cali- reading to the current correction at the corresponding
bration, provided the thermometer has not been test temperature. The example shown above is modified
abused. in this way.

Thermometer
Current Ice-Point Reading After Adjusted
Reading
Correction Heating to Test Temperature Scale Correction

+ 0.019 "C -0.019 °C

-
+.019
-
0.000 °C
10.000 .023 +.019 .004
20.000 - .028 +.017 - .011
30.000
-
.000 +.010 +
-
.010
40.000 .041 .000 .041
50.000 - .008 -.012 - .020

When this thermometer is used as a standard, the ice ice-point depression (See Sec. 8.1a) that appears
point should again be taken after each calibration when thermometer is heated. The standard can be
a
point. To obtain the actual temperature of the bath, used at any time and at any temperature with no con-
the ice-point correction is added to the appropriate cern about the change in bulb volume, provided the
adjusted scale correction and thermometer reading. ice point is taken after it is used.
If the calibration temperature is 25 °C, the thermom- It advisable to use two liquid-in-glass standards
is
eter reads 24.983 °C, and the ice-point reading taken when calibrating, in order to detect reading errors
after the calibration is +0.014 °C, the actual tempera- and maintain a cross check of the standards. The
ture is computed as follows: calibration procedure can best be described by follow-
ing a hypothetical calibration of four thermometers,
Thermometer reading 24.983 °C T, through T4.
Interpolated Adjusted Scale Correction — .001
Table 3 shows hypothetical observations taken in
Ice-point correction — .014 obtaining the corrections applicable to the thermome-
Actual Temperature of Bath Medium 24.968 °C ters at 20 °C. For simplification, all of the entries in
the table reflect perfect thermometer performance with
This method eliminates the need to wait 3 days for a temperature rise of 0.001 °c between each observation
the bulb to recover and avoids the error due to the and no observer error.

12
 Table 3. — Comparison of test thermometers obtain a correction to the scale of the thermometers.
with liquid-in-glass standards
The thermometer comparisons are repeated in the same
manner at the next higher test point until the calibra-
Ice-point readings of test thermometers
tion completed.
is

SI Tl T2 T3 T4 S2 When a platinum resistance thermometer is used

Observer A .. + 0.02 -0.02 + 0.02 0.00
as a standard, the sequence of observations is the
Observer B .. + .02 - .02 + .02 .00 same, except that the one resistance thermometer is
Mean ice read in place of the separate observations of two liquid-
points + .02 - .02 + .02 .00
in-glass standards. For more information on platinum
Ice-point
corrections - .02 + .02 - .02 .00
resistance thermometers refer to NBS Monograph 126,
Platinum Resistance Thermometry, available from the
Superintendent of Documents, U. S. Government
Thermometer comparisons
Printing Office, Washington, D. C. 20402.
Observer A reading
left to right 19.87 19.98 19.96 20.02 20.03 19.89
Observer A reading 5.3. Corrections for Emergent Stem
right to left 19.88 19.99 19.97 20.03 20.04 19.89
Observer B reading
When a thermometer is calibrated and used under
left to right 19.88 19.99 19.97 20.04 20.05 19.90 conditions of total immersion, no difficulty is encoun-
Observer B reading tered when the reported scale corrections are to be
right to left 19.89 20.00 19.98 20.04 20.05 19.90 applied. The temperature of thermometer bulb
the
Means 19.88 19.99 19.97 20.03 20.04 19.89.-.
and the portion of the stem containing the mercury
is definitely defined as the temperature of the bath

Ice Points of Standards medium. The corrections apply as given on the report,
Observer A -0.01 -0.08
when this thermometer is used at total immersion.
Observer B - .01 - .08 Occasionally it becomes necessary to use a total-
Mean ice points - .01 - .08 immersion thermometer with a portion of the stem
Ice-point emergent from the bath medium. The temperature of
corrections + .01 + .08
the environment above the bath, or apparatus con-
Adjusted scale
corrections + .12 + .04 taining the thermometer, may differ markedly from
the temperature of the thermometer bulb. It is also
possible to have pronounced temperature gradients
Calculations of corrections
along the length of exposed mercury column. A cor-
Correction to rection can be calculated to account for the difference
standards +0.13 +0.12 in temperature between the bulb and the emergent
Mean temperature,
stem. A reliable estimate of the mean temperature of
each standard 20.01 20.01.
Mean temperature the emergent stem is required and should actually
of all readings 20.01 be measured. The determination of the stem tempera-
Corrections to test ture should be repeated each time the thermometer is
thermometers +0.02 +0.04 -0.02 -0.03 used in this way, or the accuracy of the correction
will depend upon the constancy of the environmental
The first observations are the ice points of the
temperature over a period of time. Significant varia-
thermometers under thermometers are
test. The test
tions in the temperature of the emergent stem may
then mounted in the comparison bath between the two
occur due to air circulation and variations in ambient
standards. It is preferable to have two observers
temperature, even though the location of the ther-
(A and B Observer A reads in the order left to right
) .

mometer does not change.

as the thermometers appear in the table, and immedi-
The same situation occurs in the case of partial-
ately repeats the observations in the order right to
immersion thermometers. The reported scale correc-
left, while observer B records the data. Observer B
tions apply only for the indicated depth of immersion
promptly reads in the same manner while observer A
and a particular stem temperature. If the thermometer
records. The bath temperature is increasing linearly is used under conditions other than specified, the
with time and the observations are spaced uniformly
reported scale corrections are no longer applicable,
in time. For this reason the mean of the observations
and a stem temperature correction is required.
with any one thermometer will correspond to the
The following paragraphs describe methods for
mean temperature value of the comparison bath me- determining stem temperatures and calculating cor-
dium durins; the observations of all of the thermome-
rections. It will be seen how important the stem
ters. Immediately after the comparison observations, temperature correction is in relation to a desired
ice points of the two standards are observed and
accuracy.
recorded. With the ice point data and the adjusted
scale corrections for the standards, the temperatures
a. Measurement of Emergent-Stem Temperature
indicated by the standards are calculated, and an
overall mean temperature for the observations is ob- There are two methods available for measuring the
tained. This mean temperature is compared with the approximate mean temperature of the emergent stem.
mean of the observations for each thermometer to The first method consists of placing one or more small

13
auxiliary thermometers near the emergent stem as In calculating the emergent stem correction, it is
shown in figure 8(c). For a more accurate measure- convenient to express the length of the thermometer
ment faden or thread thermometers [8, 9] can he stem adjacent to the faden bulb in terms of degrees
used. These thermometers have long bulbs measuring on the thermometer scale. If a faden thermometer,
variously 5 to 20 cm, with wall thicknesses and bore having a bulb which is 10 cm long, is used for a stem
sizes nearly the same as the stem of an ordinary temperature measurement, then the number of degrees
thermometer. The bulb length is selected to approxi- corresponding to the 10 cm length must be found by
mate that of the emergent stem whose temperature is measuring a portion of the thermometer scale. This
to be measured. The stem of the faden thermometer measurement should be made over the portion of the
has a finer capillary than the bulb and is usually graduated scale which was adjacent to the faden
graduated in intervals of 2, 5, or 10 degrees Celsius. thermometer bulb. This is particularly important with
Stem temperature measurements taken at NBS are high-temperature thermometers, where the length of
based upon the use of faden thermometers whenever a degree is generally not the same throughout the
possible. entire length of the scale. In some instances the adja-
A
convenient method for measuring the emergent- cent portion of the thermometer stemis not graduated.

stem temperature of a total-immersion thermometer This is especially with

partial-immersion ther-
true
that is used at partial immersion involves the use of mometers in the area above the immersion line. This
a faden thermometer. The top of the faden thermome- ungraduated length between the immersion line and
ter bulb is placed on a horizontal plane with the top the first graduation must be evaluated in terms of
of the mercury column of the thermometer whose stem scale degrees and included as part of the distance
temperature is being measured. The faden thermom- covered.
eter chosen must have a bulb which is long enough to
cover the vertical area of unknown temperature b. Formula for Total-Immersion Thermometers
gradient necessary to be measured. This sometimes
involves placing part of the faden thermometer bulb
When a thermometer which has been graduated and
calibrated for use at total immersion is actually used
in the bath medium, since the top portion of the
at partial immersion, the correction for the emergent
medium can be at a temperature different from the
temperature of the total-immersion thermometer bulb.
stem may be calculated by the general formula,

(This is especially true for measurements above stem correction = kn (t-i - t)

150 °C). The reading of the faden thermometer will
indicate the mean temperature value of the area sur- where
rounding the bulb, which is also the mean temperature
value of the adjacent portion of the total-immersion k = differential expansion coefficient of mercury
thermometer stem. A faden thermometer used in this (or other thermometric liquid) in the particu-
manner is illustrated in figure 8(a). lar kind of glass which the thermometer is

If the stem temperature of a partial-immersion made (see table 4),

thermometer is to be measured, a similar approach n = number of thermometer scale degrees adjacent

is followed. In this case it is necessary to measure the to the faden thermometer bulb,
mean temperature from the immersion line to the top t = average temperature of n degrees of the ther-
of the mercury column of the partial-immersion ther- mometer stem (faden thermometer reading),
mometer. One or more faden thermometers with ti = temperature of the thermometer bulb.
appropriate bulb lengths are chosen to accomplish this
measurement. This procedure is shown in figure 8(b). The varies for different kinds of
coefficient k
glass, or for different temperature intervals, i.e., differ-
ent values of (t, —
t). For purposes of computing the
emergent-stem correction, the value of k may be
considered as depending on the average of <i and t.
FADEN Values of k as the function of ('j +
t) /2 for two wide-
- THERMOMETER
BULB ly used thermometric glasses are given in table 4. If
the kind of glass is not known, it is acceptable to use
COLUMN
. LIQUID
OF THERMOMETER k = 0.00016 for mercury thermometers graduated in
FADEN
THERMOMETER
degrees Celsius and k =
0.00009 for mercury ther-
'BULB IMMERSION LEVEL mometers graduated in degrees Fahrenheit.
7777 '/////
The expansions of liquids such as alcohol, toluene,
etc., vary quite rapidly with the temperature causing
paRTiiL (c) PARTIAL
(b)
IMMERSION IMMERSION k to vary considerably for different temperature in-
tervals. An approximate stem correction for such
thermometers may be calculated by setting the value
of k in the above equation as equal to 0.001 for Celsius

(O) TOTAL
thermometers or 0.0006 for Fahrenheit thermometers.
IMMERSION
Calculation of the stem correction may be illustrated
Figure 8. Schemes for measurement of emergenl-slem
by the following example: A total-immersion ther-
temperature. mometer indicates a reading of 90 °C in a bath when

14
 Table 4. — Values of k for mercury-in-glass thermometers

Mean temperature k for "normal" glass k for "borosilicate" glass

ti +t
2

For Celsius thermometers

0° 0.000158 0.000164
100 .000158 .000164
150 .000158 .000165
200 .000159 .000167
250 .000161 .000170
300 .000164 .000174
350 .000178
400 .000183
450 .000188

For Fahrenheit thermometers

0° 0.000088 0.000091
200 .000088 .000091
300 .000088 .000092
400 .000089 .000093
500 .000090 .000095
600 .000092 .000097
700 .000100
800 .000103

immersed to the 80 °C graduation mark on the scale. c. Formula for Partial-Immersion Thermometers
A 10 cm faden thermometer placed alongside the The scale corrections for partial-immersion ther-
thermometer is adjacent to the scale between the 60 mometers calibrated at NBS are applicable when the
and 90 °C graduations and indicates a stem tempera- thermometer is immersed to the immersion mark and,
ture of 80 C. For this case n = 90-60 or 30 degrees.
"
unless otherwise requested, for the unspecified stem
The stem correction is: temperatures which prevailed over the comparison
0.00016 X 30 (90-80) = +0.048 °C. baths at the time of calibration. Frequently it is
necessary to report scale corrections which are ap-
This correction is added to the corrected thermometer plicable when specified mean temperatures of the
reading to obtain the actual temperature of the bath emergent stem are requested. In such cases the emer-
medium. Note that when the temperature of the emer- gent stem temperatures are measured during calibra-
gent stem is lower than the bath temperature, the sign tion and the observations are corrected as necessary
of the correction is +, since the thermometer would to account forany differences found between the
indicate a higher temperature reading if immersed specified stem temperatures and the stem temperatures
properly. observed during test. The magnitude of the stem cor-
If a faden thermometer was not available in the rection will be proportional to the difference between
above example, the emergent-stem temperature could the specified and observed stem temperatures, and
be estimated by suspending a small auxiliary ther- may be calculated for Celsius mercurial thermometers
mometer above the bath adjacent to the thermometer. by using the following formula:
The bulb of the auxiliary thermometer would be ~ —
stem correction 0.00016 n (Up tots),
placed at the center of the emergent stem or at the
85 °C graduation. The reading of the auxiliary ther- where
mometer will indicate the approximate mean tempera-
ture of the 10 degrees (80° to 90 °C) emergent from tsp = specifiedmean temperature of the emergent
the bath. For this correction the value for n would stem (for which reported scale corrections
be 10. If the auxiliary thermometer indicates a read- apply),
ing of 60 "C, the stem correction would be: tobs = observed mean temperature of the emergent
stem (faden thermometer reading),
0.00016 X 10 (90-60) = +0.048 °C.
n = number of scale degrees equivalent to the
length of emergent stem (including the eval-
This method is not usually as reliable as the method
involving the use of a faden thermometer [9]. uated area above the immersion line).

15
This formula may also be used to correct the indica- A Beckmann thermometer should not be used with
tions of a partial-immersion thermometer, when, dur- any part of the lower portion of the stem exposed to
ing use, the actual stem temperatures differ from the ambient temperature. Since this part may contain
specified ones for which the scale corrections apply. 5 to 10 times more mercury per centimeter than the
graduated portion, a large and uncertain error will
be introduced if this section is not in the bath medium.
d. Formula for Calorimetric Thermometers If it is unavoidable, and such a thermometer must

The stem correction is often important when ther- be used way, the necessary correction may be
in this
mometers are used for differential temperature meas- computed from the above formula, provided S in the
urements, as in calorimetry (Sec. 6.5). The correction formula is replaced by S + m, where m is the number
may be computed from the following formula, which of degrees the temperature of the thermometer must
involves the difference between the initial and final be lowered to bring the meniscus from the zero mark
readings, provided the mean temperature of the stem on the scale to the point of immersion.
remains constant. If the thermometer is immersed to some point other
than the zero mark, as would ordinarily be the case
Stem correction = kd {h +iF — / — t)
with thermometers having the zero graduation at the
where top of the scale, the differential stem correction may
be calculated from the above formula if S is replaced
k = factor for relative expansion of glass and by 5 +m. The formula is applicable whether the
mercury, point of immersion is on the scale or below it, provided
ti = initial reading, the points at which readings are made are above the
tF — final reading, point to which the thermometer is immersed.
/ = scalereading to which the thermometer is
immersed,
t = mean temperature of the emergent stem, 5.4. Number and Choice of Test Points
J = — h.
A thermometer is usually calibrated at points spaced
This correction must be applied (added if positive, uniformly over the entire range of the main scale.
subtracted if negative) to the difference of the read- The number of calibration points chosen depends on
ings to give the actual temperature difference. the range of scale, graduation interval, and accuracy
Example: The thermometer was immersed to the desired. The interval between the calibration points
20° mark: the initial reading, ii, was 25 °C; the final should not be unnecessarily small, nor should it be
reading, U, was 30 and the stem temperature was so large as to destroy confidence in interpolated cor-
20 °C. The correction is: rections for temperature values between the calibration
points.
0.00016 x 5 (25 +30—20—20) = +0.012 °C. For thermometers not graduated above approxi-
mately 200 °C, it is generally accepted that the inter-
Since the difference between U and Zf was 5°, the
val between test points should not exceed 100 scale
actual difference between the initial and final tem-
divisions, if the corrected temperature values between
perature readings was:
the calibration points are to have an expected accu-
— + correction = 5.012 °C. racy of approximately one-half of one scale division.
If accuracies of one or two-tenths of a scale division
are desired, it will be necessary to reduce the calibra-
e. Formula for Beckmann Thermometers tion interval to every 40 or 50 scale divisions. If a

The formula used to compute the stem temperature thermometer is graduated above 200 °C, a 40 to 50
correction for Beckmann thermometers (Sec. 6.4) is scale division calibration interval is required to pro-
similar to the one used for calorimetric thermometers. duce corrected temperature values with expected ac-
The formula given below is applicable provided the curacies of approximately one-half of a scale division,
thermometer is immersed near the zero degree indi- and a 20 to 25 scale division calibration interval is
cation on the scale and the temperature of the stem necessary for expected accuracies to be approximately
remains constant. one or two-tenths of a scale division.
The above results were derived from analysis of
stem correction = kd (5 + + ?f — ;), calibration data taken on more than 50 thermometers
purchased from 1930 through 1956 for use as labora-
where
tory standards. The data indicated that there was con-
k — factor for relative expansion of glass and siderable variation between individual thermometers
mercury, and that scale corrections obtained over a given inter-
t\
~ initial reading, val for a particular thermometer were not sufficient to
tr = final reading, predict whether or not more calibration points were
t = mean temperature of the emergent stem, required. The above studies were made with only a
d — If — (1 few of the many types of thermometers submitted to
S = setting of the thermometer (Sec. 6.4). NBS for calibration, and may not necessarily be

16
applicable to other types. Experience with a particular the purchase order, be tested at a reference
it will
type of thermometer seems to be the most reliable point and at approximately every 100
intervals of
guide in the choice of its calibration points. scale divisions. A calibration should never consist of
fewer than two points on the main scale, and should
If a thermometer is submitted to NBS for calibra- always be tested at a reference point, whether on the
tion and the calibration points are not specified on main scale or on an auxiliary scale.

6. Common Thermometers and Factors Affecting Their Use

A and expected accuracies for

listing of tolerances Other factors such as ice-point changes, unless ex-
common types of liquid-in-glass thermometers, which actly accounted for, and differences in external pres-
are accepted for calibration, are given in tables 5 sure may also account for inaccuracies much greater
through 12. The scale tolerances shown are chosen to than the imprecision with which a scale having 0.1 or
be indicative of good manufacturing practice. When 0.2 degree graduations may be read.
a thermometer is manufactured, small errors in point-
ing (marks placed on a blank thermometer at various 6.1. Total-Immersion Thermometers
temperatures to be used as guides for the placement
Thermometers pointed and graduated by the manu-
of the graduation lines) and graduating are inevitable.
facturer to indicate correct temperatures when the
These graduation marks are also subject to variations
bulb and the portion of the stem containing the ther-
due to the inherent properties of the glass. The toler-
mometric liquid are subjected to the temperature being
ances must be sufficiently restrictive to insure a satis-
measured are known as total-immersion thermometers.
factory high-grade thermometer, and at the same time
While these thermometers are designed for immersion
not cause undue manufacturing difficulties.
of all the mercury, it is not desirable to immerse the
In addition to the scale tolerance limit, the error in
portion of the stem above the meniscus. The heating
any temperature interval should not exceed 5 percent
of this portion to high temperatures could cause ex-
of the nominal value of the interval. The purpose of
cessive gas pressures resulting in erroneous readings
this requirement is to eliminate thermometers having
and possibly permanent damage to the bulb.
large corrections of alternating signs, which lead to
In practice, a short length of the mercury column
uncertainties in the interpolation of scale corrections
often must be left emergent from the bath or medium
between the calibration points.
being measured so that the meniscus is visible. If the
The word "accuracy" used in these tables refers to
temperature difference between the bath medium and
the best values attainable in the use of thermometers
its surroundings is large, an appreciable temperature
when all corrections are applied. The accuracy bounds
gradient may exist around the thermometer stem near
may seem broad in some instances, but the definite
the surface of the bath. This condition becomes more
thermometry become ap-
limitations of liquid-in-glass
serious when a total-immersion thermometer is inten-
parent when all factors are considered. For example,
tionally used at partial immersion. If either situation
if the scale is expanded by reducing the diameter of
exists, an emer,s;ent stem correction, as explained in
the capillary, a practical limit is reached beyond
section 5.3b, will be necessary.The correction may be
which capillary forces, in combination with the elas-
as large as 20 Celsius degrees (36 Fahrenheit degrees)
ticity of the thermometer bulb, will prevent a smooth
if the length of the emergent liquid column and the
advance or retreat of the mercury column. The move-
difference in temperature between the bath and the
ment of the mercury meniscus may be erratic and
space above it are large. Tolerances and accuracies
occur in steps appreciably large in comparison to the
expected of total-immersion thermometers are given
graduation interval. This is particularly true when
in tables 5 and 6.
the temperature of the medium is decreasing. Less
rigid bulbs (relatively large diameters and/or thin
6.2. Partial-Immersion Thermometers
walls), as well as capillaries of small diameters, may
cause large "meniscus jumps." Excessively elliptical In many instances the use of a total-immersion
or flattened bores are not recommended for the same liquid-in-glass thermometer for temperature measure-
reason. Therefore, increasing the length of a degree ments is inconvenient or impossible. For this reason
on the scale, for practical bulb sizes, improves ther- partial-immersion thermometers are designed with
mometric performance only to a certain point. Beyond scales graduated to indicate correct temperatures when
reading may readily be mistaken
this point precision of the thermometers are immersed to specified depths. Un-
for accuracy in temperature measurement. A study of less otherwise stated, each Report of Calibration issued
the effects of bulb and capillary dimensions on ther- by NBS gives corrections which are applicable for
mometer performance, made by Hall and Leaver [10], temperatures prevailing above the comparison baths.
provides valuable guidelines for design purposes. No stem temperature correction is necessary when

17
 Table 5. — Tolerances for Celsius total-immersion mercury emergent stem, t gether with the inherent difficulty
thermometers of estimating or measuring the emergent-stem tem-
perature with sufficient accuracy, can contribute
Temperature range in Graduation Tolerance Accuracy markedly to the uncertainty of a given thermometer
degrees interval in in in degrees indication. For these reasons tolerances and accu-
degrees degrees racies expected of partial-immersion thermometers,
as given in tables 7 and 8, are not as stringent as
Thermometer graduated under 150 °C those for total-immersion thermometers.

°C —
Table 7. Tolerances for Celsius partial-immersion
0 up to 150 1.0 or 0.5 0.5 0.1 to 0.2
mercury thermometers
0 up to 150 .2 .4 .02 to .05
0 up to 100 .1 .3 .01 to .03

Temperature range in Graduation Toler- Accuracy ^

Thermometers graduated under 300 °C degrees interval in ance in in degrees

degrees ^ degrees

0 up to 100 f 0.5 ).l

( to 0.2
I 1.0 or 0.5 Thermometers not graduated above 150 °C
Above 100 up to 300 .... ) 1.0 .2 to .3
0 up to 100 ( 0.4 .02 to .05
Above 100 up to 200 .... 1 1 -5 .05 to .1
0 up to 100 1.0 or 0.5 1.0 0.1 to 0.3
0 up to 150 1.0 or 0.5 1.0 0.1 to 0.5

Thermometers graduated above 300 °C

Thermometers not graduated above 300 °C
0 up to 300 { 2.0 0.2 to 0.5
2.0
Above 300 up to 500 .... 1 ( 4.0 .5 to 1.0
0 up to 100 1.0 1.0 0.1 to 0.3
0 up to 300 ( 2.0 .Ito .5
1
1.0 or 0.5 Above 100 up to 300 .... 1.0 1.5 .5 to 1.0
Above 300 up to 500 ....
I 4.0 .2 to .5

Thermometers graduated above 300 °C

Table 6. — Tolerances for Fahrenheit total-immersion 0 up to 300 f 2.5 0.5 to 1.0

2.0 or 1.0
Above 300 up to 500 .-. 1 5.0 1.0 to 2.0
mercury thermometers I

Temperature range in Graduation Tolerance Accuracy

degrees interval in in in degrees —
Table 8. Tolerances for Fahrenheit partial-immersion
degrees degrees
mercury thermometers

Thermometers graduated under 300 °F

Temperature range in Graduation Toler- Accuracy l>

degrees interval in ance in in degrees

degrees » degrees
32 up to 300 2.0 1.0 0.2 to 0.5
32 up to 300 1.0 or 0.5 1.0 .Ito .2
32 up to 212 .2 or .1 0.5 .02 to .05
Thermometers not graduated above 300 °F

Thermometers graduated under 600 °F

32 up to 212 2.0 or 1.0 2.0 0.2 to 0.5
32 up to 300 2.0 or 1.0 2.0 0.2 to 1.0
32 up to 212 1.0 0.2 to 0.5
2orl I
Above 212 up to 600 .... 1 1 2.0 .5

Thermometers not graduated above 600 °F

Thermomete rs graduated a bove 600 °F
32 up to 212 2.0 or 1.0 2.0 0.2 to 0.5
32 up to 600 ] c ( 4.0 0.5 to 1.0 Above 212 up to 600 .... 2.0 or 1.0 3.0 1.0 to 2.0
Above 600 up to 950 ....
^ 7.0 1.0 to 2.0
{ |

32 up to 600
Above 600 up to 950 ....
(
j
oil
^ or i ^
3.0
g ^
0.2 to 1.0
^ ^ q Thermometers graduated above 600 °F

32 up to 600 5.0 1.0 to 2.0

5.0 or 2.0
Above 600 up to 950 10.0 2.0 to 3.0

these thermometers are used under conditions similar

to those existing during calibration. However, if con- " Partial-immersion themiometers are sometimes graduated in

ditions differ, especially the temperature of the emer- smaller intervals than shown in these tables, but this in no way
improves the performance of the thermometers, and the listed
gent stem, the necessary emergent stem correction
tolerances and accuracies still apply.
must be calculated as shown in section 5.3c. An un- ^ The accuracies shown are attainable only if emergent stem

steady or irreproducible environment surrounding the temperatures are closely known and accounted for.

18
 6.3. Low-Temperature Thermometers Table 9. — Tolerances for low-temperature total-immersion
thermometers
The use of mercury-in-glass thermometers for low
temperature measurements is limited bv the freezing
Temperature Type
point of mercury, which is —
38.9 °C' (— 38.0°F). range in
of
thermometer
Graduation
interval in
Toler-
ance in
Accuracy
in degrees
This limit may be extended to lower temperatures by degrees degrees degrees
forming an eutectic alloy consisting of mercury and 8.5
percent by weight of thallium. The freezing point of Celsius thermometers
this alloy is approximately —
59 °C ( —
74 °F). How-
ever, small differences in the quantity of thallium
-35 to 0 Mercury 1 or 0.5 0.5 0.1 to 0.2
present have the effect of markedly raising the freez- -35 to 0 do .2 .4 .02 to .05
ing point of the alloy. It has also been found that -56 to 0 Mercury- .5 .5 .1 to .2

some thermometers containing this liquid have be- thallium.

haved erratically range of —59 to —57 °C.

in the
-56 to 0 do .2 .4 .02 to .05
-200 to 0 Organic 1.0 2.0 .2 to .5
Consequently, thermometers using a mercury-thallium liquid.
alloy as the liquid should not be used below 56 °C —
(—69 °F).
Fahrenheit thermometers
Temperature measurements from 56 °C ( 69 — —
°F) to approximately —200 °C (—328 °F) can be -35 to 32 Mercury 1 or 0.5 1.0 0.1 to 0.2
made by using thermometers containing organic li- -35 to 32 do .2 0.5 .05
quids. Alcohol, pentane, or other organic
toluene, -69 to 32 Mercury- lor .5 1.0 .Ito .2
thallium.
liquids, alone or in mixtures, have been used as fluids
-69 to 32 do .2 0.5 .05
for low temperature thermometers. All of these fluids -328 to 32 Organic 2 or 1.0 3.0 .3 to .5
have limitations, and thermometers containing organic liquid.
liquids are not considered to be as reliable as ther-
mometers containing mercury or mercury-thallium
alloy. Table 10. — Tolerances for low-temperature partial-immersion
thermometers
Organic liquids have the disadvantage of wetting a
surface and leaving a film on the wall of the capillary
Temperature Type of (Graduation Toler- Accuracy
when the liquid column recedes. This may lead to range in thermometer mterval in ance in in degrees
significant errors in the temperature indications if degrees degrees degrees
sufficient precautions are not taken. The thickness of
the film on the capillary wall will depend, among other Celsius thermometers
things, on the viscosity of the liquid, the interfacial
action between the liquid and glass, and the rate at
-35 to 0 Mercury 1.0 or 0.5 0.5 0.2 to 0.3
which the thermometer is cooled. Whenever possible -56 to 0 Mercury- 1.0 or .5 .5 .2 to .3
the rate of cooling should be slow with the bulb cooled thallium.
first,enabling the viscosity of the organic fluid in -90 to 0 Organic 1.0 3.0 .4 to 1.0
liquid.
the capillary to be kept as low as possible until the
final temperature is reached. This minimizes the
amount of liquid left on the capillary wall. Sufficient Fahrenheit thermometers
time should always be allowed to assure complete
drainage. Under adverse conditions this may take an -35 to 32 Mercury 1.0 or 0.5 1 0.3 to 0.5
hour or more. -69 to 32 Mercury- 1.0 or .5 1 .3 to .5
thallium.
In addition to good drainage characteristics, a satis- - 130 to 32 Organic 2 orl 5 .8 to 2.0
factory low-temperature fluid should not contain water, liquid.

dirt, or other foreign material which will separate

from the liquid. Because low-temperature thermom-
eters are frequently designed for use above room 6.4. Beckmann Thermometers
temperature, the vapor pressure of the liquid must A metastatic, or Beckmann thermometer,
usually of
be low to prevent distillation at these higher tempera- the enclosed-scale type, constructed in a manner
is
tures. Any dye, added to the fluid for improved
that permits mercury to be removed from, or added
visibility, should be chosen for good color fastness
to, the bulb, making it possible to use the same ther-
with respect to light exposure or chemical action with mometer for differential temperature measurements in
the organic liquid. various temperature ranges [11]. The scale usually
Tolerances and accuracies applicable to low-tem- consists of 5 or 6 Celsius degrees, although some micro
perature thermometers are given in tables 9 and 10. types have a scale containing only 3 Celsius degrees.

19
 The "setting" of a Beckmann thermometer refers setting factor is 1.0000 at this temperature. For a set-
to the temperature of the bulb when the reading on ting other than 20 °C, the observed temperature differ-
the scaleis 0°. When the setting is changed, enabling ence must be multiplied by the appropriate setting
the thermometer to be used at a higher or lower tem- factor as shown in the example given below the table.
perature range, the quantity of mercury affected by a For most Beckmann thermometers, the large bulb
temperature change is not the same. Therefore, two is joined to a fine capillary, which is backed by a

equal changes in temperature at different settings milk-glass scale, by a capillary of much larger diam-
cause different indications on the scale, and a "setting eter. This large capillary is a source of some un-
factor" must always be used to convert reading differ- certainty when the thermometer is used at partial
ences into actual temperature differences whenever immersion and this portion is not entirely immersed.
the thermometer is used at any setting other than the For appropriate emergent-stem corrections refer to
one at which the scale was calibrated. These setting section 5.3e.
factors combine corrections for the different quan- Tolerance requirements for Beckmann thermometers
tities of mercury during equal temperature changes, are given in table 12. Under the heading "Accuracy
and the difference between the mercury-in-glass scale of interval in degrees" is given the estimated accuracy
and the International Practical Temperature Scale of attainable in the measurement of any interval within
1968. Table 11 lists setting factors calculated for the limits of the scale. No tolerances for scale error
thermometers of Jena 16"' - glass or, its American are given, although it is desirable that it be no larger
equivalent, Corning normal. The scale calibrations for than 0.02 °C over a 1.0 °C interval.
Beckmann thermometers as reported by NBS are
applicable for a setting of 20 °C. Consequently, the
Table 12. — Tolerances for Beckmann and calorimetric
thermometers
Table 11. — Setting factors for Beckmann thermometers

Gradua- Allowable change Accuracy of

Setting Factor Setting Factor Type of tion in- in correction in interval in
thermometer terval in degrees degrees
degrees
0 °C 0.9935 55 °C 1.0092
5 0.9951 60 1.0102
10 0.9967 65 1.0112 Beckmann 0.01 °C 0.01 over 0.5° 0.002 to 0.005 ..-

15 0.9984 70 1.0121 interval for

20 1.0000 75 1.0129 setting of 20 °C
25 1.0014 80 1.0137 Bomb calorimetric .01 °C 0.02 over 1.5° .005 to .01
30 1.0029 85 1.0146 interval
35 1.0042 90 1.0154 Do .02 °C 0.02 over 1.5° .005 to .01
40 1.0056 95 1.0162 interval
45 1.0070 100 1.0170 Do .05 °F 0.04 over 2.5° .01 to .02

50 1.0082 interval
Gas calorimetric ... .1 -F 0.15 over a 5° .02 to .05
interval

As an suppose the following observations were made:

illustration,
25 °C.
Setting — Lower reading = 2.058°
Stem temperature 24 °C.= Upper reading = 5.127°
Lower Upper 6.5. Calorimetric Thermometers
Observed reading = 2.058 5.127
Correction from certificate — +0.005 "0.008 Calorimetric thermometers include a specialized
Corrected upper reading 5.119 group of solid-stem mercury-in-glass thermometers
Corrected lower reading 2.063
which are used for accurate differential measurements.
Difference = 3.056
Since the accuracy of these thermometers at any one
Difference multiplied by setting factor (1.0014) = 3.060
Emergent stem correction (see accompanying temperature is of less iinportance than the accuracy
stem correction sheet) — + .004 of the temperature intervals, no reference point is
Corrected difference — 3.064 required.
Table 12 gives the scale error required of some
typical calorimetric thermometers. Like the Beckmann
^Certain commercial products and instruments are identified in this paper
in order to specify adequately the experimental procedure. In no case does thermometer, no tolerances for scale error are given,
such identification imply recommendation or endorsement by the National
Bureau of Standards, nor does it imply that the products or equipment
but it is desirable that the scale corrections be no larger
identified are necessarily the best available for the purpose. than approximately five graduation intervals.

20
 7. Thermometer Design

Reproducibility and accuracy of readings are influ- and causing only small changes in the thermometer
enced markedly by the design, choice of materials, readings with continued use. This is especially impor-
and actual construction of the thermometer. The cali- tant for a thermometer graduated above 300 °C or
bration laboratory personnel exercise judgment in 600 "F. The quality of the thermometer glass and the
deciding whether individual thermometers are likely adequacy of the annealing process may be judged by
to maintain a calibration; however, general recom- the stability of the reference-point readings. A method
mendations for acceptable thermometer design and for testing bulb stability is described in The American
construction are given in the following sections. Society for Testing and Materials Method E77.
All high-temperature thermometers should be filled
7.1. Materials of Construction with a dry inert gas such as nitrogen under sufficient
pressure to prevent separation of the mercury at any
While the cleanliness of the thermometer bulb, bore, temperature indicated on the scale. Total-immersion
and liquid filling have a pronounced effect upon the thermometers graduated above 150 °C or 300 °F must
performance of a thermometer, equally important is be gas filled to minimize the distillation of mercury
the proper choice of the glass from which the ther- from the top of the mercury column. An expansion
mometer is manufactured. The thermometer, and chamber at the top of the capillary or a specified
especially the bulb, must be made of glass suitable length of unchanged capillary above the highest grad-
for use in the temperature range indicated on the stem. uation is essential for thermometers containing a gas.
Some types of glasses commonly used
in the manu- The gas must have an adequate area for compression
facture of thermometer bulbs and reasonable upper when the mercury is advanced toward the top of the
temperature limits are estimated by Thompson [12] scale; otherwise, excessive pressure in the capillary
in table 13. Thompson's estimates are based upon the will cause the thermometer to burst.
work of Liberatore and Whitcome [13]. The results For thermometers graduated below 150 °C or
show that significant changes in bulb volume may 300 °F a gas filling is optional, but is strongly rec-
occur if the bulb is heated for long periods of time ommended. The mercury column of a vacuous ther-
at temperatures higher than 130 °C (234 °F) below mometer will tend to separate easily if the thermometer
the strain point of the bulb glass. The strain point of a is inverted or subjected to a sudden shock.
glass is defined as the temperature at which the glass
has a viscosity of 10"^ poises [14]. Thermometers
may be used intermittently, to approximately 70 °C 7.2. Scale Design and Workmanship
(126 °F) of the strain point. It should be noted that Solid-stem thermometers must have the graduation
the use of a glass with a high strain point, such as
marks placed directly on the stem and opposite the
borosilicate glass, results in better thermometer per-
enamel back. If the thermometer is of the enclosed-
formance and stability even in thermometers used at scale type, the graduated scale must be securely
temperatures much lower than the exposure limits
fastened to prevent relative displacement between the
given in table 13.
scale and the capillary, or a datum line should be
conveniently located to indicate at any time whether
Table 13. — Temperature exposure limits for various
the scaleis in its original position. The graduation
thermometer glasses ^
marks should be clear cut, straight, of uniform width,
and in a plane perpendicular to the axis of the
Exposure limits thermometer.
Strain
point
The scale should be graduated in 1.0, 0.5, 0.2, or 0.1
Continuous Intermittent degree intervals, or in decimal multiples. The divisions
should be numbered in such a manner that the iden-
tification of any graduation is not unnecessarily
"C "C °F "C
''430* 1-805* difficult. Thermometers with scales graduated in 0.25
Corning normal 7560 500 370 700
Kimble R 6 490 360 680 420 790 degree intervals, or submultiples, are sometimes diffi-
Jena 16 III 495 365 690 425 795 cult to read and the discontinuance of their production
Coming borosilicate 8800 529 400 750 460 860
is desirable. If the thermometers are graduated in 0.1
Jena borosilicate 2954 548 420 790 480 900
1005 600 1110 or 0.2 degree intervals (or decimal multiples), every
Corning 1720 668 540
Jena Supremax 2955 665 535 995 595 1100 fifth mark should be longer than the intermediate
ones and a number should appear at every tenth mark.
a From reference
Thermometers graduated in 0.5 degree intervals (or
[12].
b *405 °C or 760 T
if Corning Standard Thermometer 0041 decimal multiples) require three lengths of graduation
glass is used for the stem. marks. These consist of alternating short and inter-
mediate marks, with every tenth line distinctly longer
The thermometer must also be adequately annealed than the others. A number should appear at every
bulb volume to remain reasonably stable
to enable the 10th or 20th mark.

21
 The must not be extended to temperatures
scale graduation next above, if the graduation is not above
beyond which the particular thermometer glass is 100 °C (212 ° F) a 30; mm
length if the graduation
suited. A thermometer made of borosilicate glass, for is above 100 °C (212 °F).
example, should never be graduated to 500 °C (d) A 10 mm length of unchanged capillary above
(932 °F). It would be ruined in a short time if used the highest graduation, if there is an expansion cham-
at that temperature. ber (see definitions, Sec. 3); a 30 mm length if there
is no expansion chamber.

7.3. Scale Dimensions

Although optimum line width depends in some 7.4. Reference Point

measure upon the intended use of a particular ther-
Thermometers graduated above 150 °C or 300 °F,
mometer, coarse graduation marks do not represent
or precision thermometers with an expected accuracy
good design. If the thermometer indications are to be
of better than 0.1 °C or °F when calibrated for
0.2
observed precisely (for example to 0.1 of a division),
actual temperature measurement, must have a refer-
the width of the graduation marks should not exceed
ence point. The effects of changes in the bulb volume
0.2 of the interval between the center lines of the
on the thermometer indications may be followed
graduations. In instances where the thermometer must
throughout the life of the thermometer by periodic
be read quickly or in poor light, and less precision is
From the reference point
testing at the reference point.
expected, somewhat wider lines may be acceptable.
observations, the thermometer corrections can be kept
The graduation marks must not be too closely
current as explained in section 5.2. If a suitable ref-
spaced. The closest permissible spacing depends upon
erence point, such as the ice or steam point, is not
the fineness and clearness of the marks. In no case
included in the range of the main scale, a short aux-
should the distance between center lines of adjacent
iliary scale containing a fixed point should be provided.
graduation marks on a solid-stem thermometer be less
To prevent the thermometer from being unduly long,
than 0.4 mm. The minimum permissible interval be-
a contraction chamber may be introduced between the
tween graduation marks for an enclosed-scale ther-
auxiliary scale and the main scale. Any auxiliary scale
mometer is 0.3 mm if the lines are ruled on a milk-
must have graduations identical to those on the main
glass scale; otherwise, the minimum is also 0.4 mm.
scale, both dim.ensionally and in terms of temperature,
The minimum no case represents good design, and
in
and they must extend for a short interval both above
well-designed thermometers will have graduation inter-
and below the reference point. Similarly, when the
vals considerably larger than the specified minimum.
main scale ends near a reference point, the graduations
In order for a thermometer to be usable over the
must be continued for a short interval beyond the
entire range of the scale, the graduation marks must
reference temperature.
not be placed too close to any enlargement in the
Reference points are not needed on thermometers
may result if any of the
capillary. Appreciable errors
intended for differential measurements (such as calori-
following conditions exist. (1) There is insufficient
metric thermometers), nor on thermometers graduated
immersion of the mercury in the main bulb or a capil-
below 150 °C or 300 °F, if they are not to be cali-
lary enlargement; (2) the graduation marks have
brated to an accuracy of better than 0.1 '°C or 0.2 °F.
been placed over parts of the capillary that have been
changed by manufacturing operations; or (3) the
graduations are so close to the top of the thermometer 7.5. Marking of Partial-Immersion
that excessive gas pressure results when the mercury Thermometers
is raised to this level. The distances given below are

considered as minimum limits commensurate with Partial-immersion thermometers must be plainly

good thermometer design marked "partial immersion" or the equivalent (for
(a) A 13 mm length of unchanged capillary example, "76 mm immersion"), or a conspicuous line
between the bulb and the lowest graduation, if the permanently placed on the stem to indicate the proper
graduation is not above 100 °C (212 °F); a 30 mm depth of immersion. This line must not be less than
length if the graduation is above 100 °C (212 °F). 13 mm above the top of the bulb. Special partial-
fb) A 5 mm length of unchanged capillary be- immersion thermometers, made for use in a specific
tween an enlargement and the graduation next below, manner or instrument (for example, viscometers and
except at the top of the thermometer. flash-point testers where the thermometer is held in
c( A 10
) mm
length of unchanged capillary be- a ferrule or other mounting arrangement suited to the
tween an enlargement, other than the bulb, and the instrument), need not be marked.

22
 8. Special Notes

The following brief notes on the characteristic be- responding to that particular high temperature will
havior of mercury-in-glass thermometers are added to appear to be reached. If the thermometer is cooled
aid the user in understanding the behavior of these slowly through critical temperature regions, the glass
thermometers, and to better utilize the information will nearly return to its initial state, and the ice-
contained in the Reports of Calibration. point reading will show no change. If, on the other
hand, the thermometer is cooled rapidly (such as
cooling naturally in still air), the bulb will retain a
8.1. Glass Changes portion of its expanded condition, and the ice-point
reading will be lower than the reading taken before
The changes which occur in the glass of a ther-
mometer bulb after first heating to a high temperature heating. This phenomenon is known as "zero, or ice-
point depression." Thermometers which have been
within the acceptable exposure limits of the glass and
heated to high temperatures recover from this ice-
subsequent cooling to ambient temperatures, are an
point depression in an unpredictable way, and fre-
involved function of time and temperature. They will
quently there will be no significant recovery after a
depend upon the thermal history of the glass (both
period of one year at room temperature. Since the
during manufacture and previous use) the time of .

ice-point depression has a reproducible value, ice-

exposure to the high temperature, and the rate of
point readings may be used reliably to show changes
cooling. Evidence from many investigations [13, 15,
in the volume of the thermometer bulb with time and
16] seems to indicate that when a glass is held indefi-
use, provided the thermometer is allowed to cool in
nitely at some fixed temperature, density (and volume)
still air and the ice point is taken within a reasonable
changes proceed toward a preferred density corres-
period of time, not to exceed one hour, after being
ponding to a quasi equilibrium condition character-
heated.
istic of the particular kind of glass and the temperature.
Since these changes involve molecular rearrangements, Thermometers used below approximately 100 °C will
they proceed more rapidly at high temperatures where usually exhibit a relatively rapid recovery from the
the viscosity of the glass is lower, and the molecular ice-point depression, and the original bulb volume will
mobility consequently higher. For this reason, a close recover within the equivalent of 0.01 or 0.02 degrees
approach to quasi equilibrium may be reached in the C in approximately 3 days. This phenomenon has
order of hours at annealing temperatures, while infinite an important bearing on the precision attainable with
time may be required at much lower temperatures. mercury thermometers, and must be taken into con-
If a glass that has been heated to a high tempera- sideration, especially in the range of 0 to 100 °C.
ture is allowed to cool rapidly, it will be seen that If a thermometer is used to measure a given tempera-
equilibrium is not reached at the lower temperatures ture, it will indicate an erroneously low value if it
during cooling, and an equilibrium density more has, within a short period of time, previously been
nearly corresponding to the high temperature is exposed to a higher temperature. With the better
"frozen" into the glass. This characteristic behavior grades of thermometric glasses the error resulting from
of glass has a lasting effect on the performance of this hysteresis will not exceed (in the interval of
liquid-in-glass thermometers. For the entire lifetime 0 to 100 °C) 0.01 of a degree for each 10-degree
of the thermometer it may retain a "memory" of the difference between the temperature being measured
thermal history at the higher temperatures. and the higher temperature to which the thermometer
The techniques of good manufacture are designed to has recently been exposed. With the best glasses the
produce in the thermometer glass a state which will error may only be a few thousandths of a degree for
result in maximum stability for the range of tempera- each 10-degree difference. The errors due to this
ture indicated on the scale. To achieve perfect stability hysteresis become somewhat erratic at temperatures
for all conditions of use is not possible in thermometer above 100 °C. For these reasons it is customary, in
manufacture; therefore, changes in the ice-point read- precision thermometry, to apply a scale correction
ings are observed periodically. The changes observed based upon an ice-point reading taken immediately
at the ice point are reflected at all points on the scale after the temperature measurement (see Sec. 5.2).
by the same magnitude and sign, since they are the Where the range of temperature is small and the
result ofchanges in the bulb volume (see Sec. 5.2) time between observations is short (as in the use of
(changes in the stem have very little effect). Because calorimetric thermometers), it is more satisfactory,
of this behavior of glass, the changes in the bulb each time the thermometer is used, to first heat to
volume can be either temporary or permanent. the highest temperature to be measured, so that all of
the depression has taken place before the observations
are begun. The condition to be observed is that the
a. Temporary Changes
time required for observations is so short that no
Upon heating to a high temperature, the bulb of a appreciable recovery shall have taken place during
thermometer will expand from its initial state. After this time. As this condition is fairly well satisfied in
a short period of time, an equilibrium condition cor- calorimetric work, and as it is the only one for which

23
consistentmeasurements of the same temperature where and R\ are external and internal radii of
intervalcan be made repeatedly, calorimetric ther- the bulb, and A: is a constant containing elastic prop-
mometers should be used in this way. erties of the glass and a conversion factor for express-
ing the volume change in terms of change of

b. Permanent Changes thermometer reading in degrees. In the above formula,

the external pressure coefficient ^ is defined as the
A "secular change" in thermometer glasses, which change in scale reading in degrees resulting from a
may progress with time, results in a non-recoverable change of 1 mm Hg in external pressure. For Celsius
decrease in the bulb volume as indicated by an in- thermometers, Guillaume found a value of 5.2 X 10~^
crease in the ice-point reading. At room temperature degrees C/mm Hg for k, but Hall and Leaver [10],
there may be a gradual change which will continue by experiment, found a value approximately 25 percent
for years, but at high temperatures the changes will lower for their thermometers. In cases where an ac-
be markedly accelerated. With better grades of ther- curate correction is necessary, the value, /?e, should
mometer glasses the change will not exceed 0.1 °C be determined experimentally. A simple apparatus for
over a period of many years, provided the thermom- this determination is shown in figure 9.
eter has not been heated to temperatures above approxi-
mately 150 °C. Initially, at high temperatures, the
secular change usually progresses more rapidly, but
with continued heating and time it tends toward a
lower rate of change. The rate of secular change will
depend upon the kind of glass used in the thermometer
bulb and the particular heat treatment given the
thermometer in manufacture. Thermometers manu-
factured according to good practices will evidence
only small secular changes. However, thermometers
made of glass unsuitable for the temperature range
indicated on the scale, or improperly annealed, may
show changes as large as 12 °C or 21 °F after heating
for approximately 200 hours at high temperatures
[17] Permanent changes in the bulb volume have
.

also been observed when thermometers have been

repeatedly cycled at low temperatures (between —30
and +25 °C) [18].
When using thermometers for high temperature
measurements, one must use care to avoid overheating. Figure 9. Apparatus for measurement of the external pressure
After only a few minutes of heating the thermometer coefficient.
at a temperature higher than its intended range, the
increased gas pressure above the liquid column may The internal pressure coefficient, /Sj, is more diffi-
cause a permanent distortion of the bulb resulting in cult to determine accurately, but may be calculated
lower thermometer indications. from the value, /3e, by means of the relation,

8.2. Pressure Effects

/?, = iSe + 1.5 X 10-6

for thermometers in Celsius degrees, or

Since glass exhibits elastic properties, the volume
of a thermometer bulb will change when either the j8i = /3e + 2.7 X 10-6
internal or external pressure changes. Therefore, at a
given temperature, the reading of a thermometer in a for thermometers graduated in Fahrenheit degrees.

horizontal position will differ from the reading in a

vertical position. Thermometer readings will also vary 8.3. Lag
with altitude. Changes of approximately 0.1 °C
Practically all theoretical treatments concerning
(0.2 °F) per atmosphere have been observed for
thermometer lag are based on the assumption that
many thermometers with bulb diameters between 5
Newton's law of cooling is applicable (that the rate
and 7 mm. This value can be used with some confi-
of change in the reading of the thermometer is propor-
dence for estimating the probable effect of an external
tional to the difference between the thermometer
pressure change. The effect of change of internal
temperature and the bath temperature). Consequently,
pressure is approximately 10 percent greater.
when a thermometer is immersed in any medium, it
Formulas for both external and internal pressure
does not indicate the temperature immediately, but
coefficients have been derived by Guillaume [19].
He found
approaches it asymptotically. A detailed discussion
the relation.
of this subject has been given by Harper [20].
If the temperature of the medium is varying uni-

I3.
= k formly, the thermometer always indicates what the
temperature was at some previous time. The thermom-

24
 eter readings are said to "lag" behind the temperature 1 1 1
1
r

of the medium by an amount which may or may not

be negligible, depending upon the rapidity of the
temperature variation and the physical characteristics
of the thermometer. In this case, the lag. A, may be
defined as the interval in seconds between the time
when the bath reaches a given temperature and the
time when the thermometer indicates that temperature.
This lag is dependent upon the dimensions and material
of the thermometer bulb, the medium surrounding
the thermometer and the stirring rate of the medium.
If a thermometer is suspended in still air, the lag
may much
be as as 50 times that of the same ther-
mometer when it is immersed in a well-stirred water
bath. Since the value of the lag for mercurial thermom-
eters is not large from 2 to 10 seconds in a well-stirred
I

water bath), it is not generally necessary to correct

for it. For example, if two thermometers, one having
a lag of 3 seconds and another of 8 seconds, are read
simultaneously in a bath with the temperature rising
at a rate of 0.001 degree in 5 seconds, the former will
read 0.001 degree higher than the latter, due to the
lag. In the intercomparison of thermometers, the rate
of temperature rise can usually be kept small, making
the lag correction negligible.
A second interpretation of thermometer lag involves
immersing a thermometer in a bath where the tem-
perature of the medium remains constant. A certain
time must elapse before the thermometer reading
agrees with the temperature of the medium to 0.1 TIME, seconds
of a degree, and still longer for agreement to be
within 0.01 of a degree. In this case the lag. A, is the Figure 10. The approach to temperature in a stirred water bath
for three thermometers uith typical lag constants.
time required for the original difference in tempera-
ture between the thermometer and bath medium to be
of A for a given thermometer in a well stirred oil bath
reduced to 1/e (that is 1/2.7) of itself. In a length
of time 4 A the difference will be approximately 1.5
will be approximately twice its value in a water bath.
percent of the original difference, and in a length of For thermometer which is used to measure tem-
a

time 7 A approximately 0.1 percent. Determinations perature changes (as in calorimetry it has been ) ,

of A for solid-stem laboratory thermometers, represen- shown by White [21] that the lag enters into the
tative of American manufacture, have yielded values
observations in such a way as to be eliminated from
the results in applying the usual radiation corrections.
of approximately 2 to 3 seconds in a well-stirred water
bath. Figure 10 shows the approach of thermometer
Therefore the lag need not be considered, providing
the initial and final readings are made when the tem-
readings to the water bath temperature for three
perature is varying uniformly. This is not strictly
selected thermometers having different values of A.
If the thermometer having A = true, however, in the case of some Beckmann ther-
2.2 seconds is initially
at 25 "C. and is immersed in a constant temperature
mometers, which have no true value of A, as is ex-
bath at 75 °C, the thermometer reading will be within plained in the paper referred to above.
0.05 "C (0.1 percent of 50 °C) of the bath tempera-
ture in 15 seconds, and within 0.01 °C in 19 seconds. 8,4. Separated Columns
The curve showing A =
3.1 was obtained for an
Many inquiries are received concerning separated
American Society for Testing and Materials (ASTM)
mercury columns which occur especially during ship-
specification 56C calorimetric thermometer with an
ment. Since no means of avoiding such occurrences
outside bulb diameter of 7.9 mm
and a bulb length
has yet been found, some directions for joining the
of 44 mm. The value of A = 2.2 was found for an
mercury may be helpful and are described below.
ASTM 7C thermometer having bulb dimensions of 5.4
by 12 mm. The third curve, where A =
1.7, was ob- (a) The bulb of the thermometer may be cooled
tained for a bulb with dimensions of 5.4 by 34 mm. in a solution of common and water (or other
salt, ice,
It is probable that most solid-stem thermometers of cooling agent) to bring the mercury down slowly into
American manufacture will have values of A lying the bulb. If the salt solution does not provide sufficient
within the range of the three curves shown. It will cooling, carbon dioxide snow (dry ice) may be used.
be noted that, according to Harper [20], the value Since the temperature of dry ice is approximately

25
— 78 °C — 108 °F), and mercury
( freezes at approxi- warming into droplets tiny enough to leave space for
mately —40 °C (— 40 °F), the mercury will solidify. the gas to by-pass. The thermometer is heated, and
Cool only the bulb and never the stem or mercury the droplets are collected by the rising mercury
column. Moderate tapping of the bulb on a rubber column.
stopper or similar soft spongy object, or the applica- The procedure for thermometers containing organic
tion of centrifugal force, by swinging the thermometer liquids is similar. Separated liquid in the stem can be
in a short arc (i.e. use of centrifugal force), usually vaporized and permitted to drain down the capillary.
serves to unite the mercury in the bulb. Care must be Another method consists of gently tapping the stem
taken to warm the top of the bulb first, so that pres- above the separation against the palm of the hand,
sures in the bulb due to expanding mercury may be forcing the organic liquid to break away from the
relieved. wall of the capillary and flow down the bore to join
(b) If there is a contraction chamber above the the main column.
bulb or an expansion chamber at the top of the ther- Minute gas bubbles, which are sometimes found
mometer, the mercury can sometimes be united by along the surface of the mercury in the thermometer
warming the bulb until the column reaches the sep- bulb, may be collected by "washing" the bulb with
arated portions in either enlargement. Great care is a large gas bubble. Bring all of the mercury into the
necessary to avoid filling the expansion chamber com- bulb as outlined in section (a). Hold the thermometer
pletely with mercury, which might produce pressures in a horizontal position and gently tap it against the
large enough to burst the bulb. (The expansion cham- hand form a large gas bubble. Force the bubble to
to
ber should never be more than 2/3 full.) Joining the travel around the walls of the bulb by rotating the
mercury is more readily accomplished if the quantity thermometer and tapping it against the palm of the
in either cavity has first been shattered into droplets hand. When the entire surface has been "washed"
by tapping the thermometer laterally against the hand. rotate the bubble to the top of the bulb and reunite
This procedure should not be used if it requires the mercury as described above.
the thermometer to be heated above 260 °C (500 °F) All of these manipulations require patience, and
and the bulb should never be heated in an open flame. experience is helpful, but they will yield results if
(c) As a last resort, especially for thermometers care is used. A convenient method of ascertaining
having no expansion chambers, small separated por- that all of the liquid has been joined is a check of the
tions of the column can sometimes be dispersed by ice point or some other reference point on the scale.

9. References

[1] "The International Practical Temperature Scale of 1968," [11] Ween, Sidney, "The Beckmann Differential Thermometer:
Metrologia, Vol. 5, No. 2, 35 (April, 1969). Its Principles of Construction and Application," Ma-
[2] Calibration and Test Services of The National Bureau of terials Research and Standards, MTRSA, Vol. 12,

Standards, NBS Spec. Publ. 250 (1970 Edition) Part 5. ,

No. 8, August 1972, p. 31.
(Copies available from NBS on request). [12] Thompson, R. D., Recent developments in liquid-in-glass
thermometry. Temperature, Its Measurement and Con-
[3] Comptes Rendus des Seances de la Treizieme Conference
trol in Science and Industry 3. Part 1 (Reinhold Pub-
Generale des' Poids et Mesures (1967-1968), Resolu-
lishing Corp., New York, 1962) p. 201.
tions 3 and 4, p. 104.
[13] Liberatore, L. C, and Whitcomb, H. J., Density changes
[4] Stimson, H. F., The International Practical Temperature in thermometer slasses, J. Am. Ceram. Soc. 35, 67
Scale of 1948, J. Res. Nat. Bur. Stand. (U.S.), 65A (1952).
(Phys. and Chem.), No. 3, 139-145 (May-June 1961). [14] American Society for Testing and Materials Designation
[5] Based on the data of Osborne, N. S., and Meyers, C. H., C162-56, Standard Definitions of Terms Relating to
A formula and tables for the pressure of saturated Glass and Glass Products.
water vapor in the range 0 to 374 °C, J. Res. NBS 13. [15] Lillie, H. R., and Shaver, W. W., Method of tempering
1 (1934) RP691. glass, U.S. Patent No. 2,148,630 (Feb. 28, 1939).

[6] Scott, R. B., and Brickwedde, F. G., A precision cryostat [16] Liberatore, L. C, Method of stabilizing the molecular
with automatic temperature regulation, BS J. Res. 6. arrangement of glass thermometers, U.S. Patent No.
401 (1931) RP284. 2,610,445 (Sept. 16, 1952).

American Society [17] Ruh, E. L., and Conklin, G. E., Thermal stability in
[7] for Testing and Materials Designation
E77-72, Calibration at Temperatures Other Than Fixed
ASTM thermometers, ASTM Bui. No. 2,33. 35 (Oct.
1958).
Points, Section 11.2.
[18] Martin, W. I., and Grossman, S. S., Calibration drift with
[8] Buckingham, E., The
correction for emergent stem of the
thermometers repeatedly cooled to ~30 °C, ASTM
mercurial thermometer, Bui. BS 8. 239" (1912) S170.
Bui. No. 231. 62 (July 1958).
[9] Pemberton, L. H., Further consideration of emergent col- [19] Guillaume, C. E., Traite Practique de la Thermometrie,
umn correction in mercury thermometry, J. Sci. Instr. Gauthier-Villars et Fils, Paris (1889) p. 99.
41, 234 '(1964).
[20] Harper 3d. D. R., Thermometrie lag, Bui. BS 8, 659
[10] Hall, J. A., and Leaver, V. M., The design of mercury (1912) S185.
thernK^meters for calorimetry, J. Sci. Instr. 36. 183 [21] White, W. P., Lag effects and other errors in calorimetry,
(195P). Phys. Rev. 31, 562 (1910).

26
•CrV.S. GOVERNMENT PRINTING OFFICE; 198 6-1+91 070 20331+
NBS-n4A (REV. 2-8CI

U.S. DEPT. OF COMM. 1. PUBLICATION OR 2. Performing Organ. Report No. 3. Publication Date
REPORT NO.
BIBLIOGRAPHIC DATA
SHEET (See instructions) NBS/MONO-174 January 1906
4. TITLE AND SUBTITLE

Thermometer Calibration: A Model for State Calibration Laboratories

5. AUTHOR(S)
J. A. Wise and R. J. Soulen, Jr,
6. PERFORMING ORGANIZATION (If joint or other than NBS, see in struct/on sj 7. Contract/Grant No.

NATIONAL BUREAU OF STANDARDS 8. Type of Report & Period Covered

DEPARTMENT OF COMMERCE
GAITHERSBURG, MD 20899 Final
9. SPONSORING ORGANIZATION NAME AND COMPLETE ADDRESS (Street, City. State, ZIP)

Same as #6.

10, SUPPLEMENTARY NOTES

Library of Congress Catalog Card Number 85-609636

I I
Document describes a computer program; SF-185, FlPS Software Summary, is attached.
11. ABSTRACT (A 200-word or less factual summary of most significant information. If document includes a significant
bi bliography or literature survey, mention it here)

This document describes the means by which a state calibration laboratory

can establish a calibration service based on liquid-in-glass thermometers.
Discussed are: ice-point baths, controlled-temperature baths, thermometer
inspection, calibration techniques, and control chart procedures.

12. KEY WORDS (Six to twelve entries; alphabetical order; capitalize only proper names; and separate key words by semicolons)

Calibration; IPTS-68; liquid-in-glass thermometers; temperature

13. AVAILABILITY 14. NO.OF

PRINTED PAGES
I
>j Unl imited

I I
For Official Distribution. Do Not Release to NTIS 88
f~53 Order From Superintendent of Documents, U.S. Government Printing Office, Washington, D.C.
20402. 15. Price

[ I
Order From National Technical Information Service (NTIS), Springfield, VA. 22161

fr Sir
II
 NBS Technical Publications

Periodical

I
Journal of Research —The Journal of Research of the National Bureau of Standards reports NBS research
{
and development of the physical and engineering sciences in which the Bureau is active.
in those disciplines
These include physics, chemistry, engineering, mathematics, and computer sciences. Papers cover a broad
range of subjeas, with major emphasis on measurement methodology and the basic technology underlying
standardization. Also included from time to time are survey articles on topics closely related to the Bureau's
technical and scientific programs. Issued six times a year.

Nonperiodicals

Monographs — Major contributions to the technical literature on various subjects related to the Bureau's scien-
tific and technical activities.

—
Handbooks Recommended codes of engineering and industrial practice (including safety codes) developed in
cooperation with interested industries, professional organizations, and regulatory bodies.
Special Publications — Include proceedings of conferences sponsored by NBS, NBS annual reports, and other
special publications appropriate to this grouping such as wall charts, pocket cards, and bibliograpiiies.
Applied Mathematics Series — Mathematical manuals, and studies of special interest to physicists,
tables,
engineers, chemists, biologists, mathematicians, computer programmers, and others engaged in scientific and
technical work.
National Standard Reference —
Data Series Provides quantitative data on the physical and chemical properties
: of materials, compiled from the world's literature and critically evaluated. Developed under a worldwide pro-
gram coordinated by NBS under the authority of the National Standard Data Act (Public Law 90-396).
NOTE: The Journal of Physical and Chemical Reference Data (JPCRD) is published quarterly for NBS by
theAmerican Chemical Society (ACS) and the American Institute of Physics (AIP). Subscriptions, reprints,
and supplements are available from ACS, 1155 Sixteenth St., NW, Washington, DC 20056.
Building Science Series —
Disseminates technical information developed at the Bureau on building materials,
components, systems, and whole structures. The series presents research results, test methods, and perfor-
mance criteria related to the structural and environmental functions and the durability and safety
characteristics of building elements and systems.

Technical Notes — Studies or reports which are complete in themselves but restrictive in their treatment of a
subject. Analogous to monographs but not so comprehensive in scope or definitive in treatment of the subject
area. Often serve as a vehicle for final reports of work performed at NBS under the sponsorship of other
•

I
government agencies.
Voluntary Product Standards — Developed under procedures published by the Department of Commerce in
Part 10, Title 15, of the Federal Regulations. The standards establish nationally recognized re-
Code of
quirements for products, and provide all concerned interests with a basis for common understanding of the
characteristics of the products. NBS administers this program as a supplement to the activities of the private
' sector standardizing organizations.

;
—
Consumer Information Series Practical information, based on NBS research and experience, covering areas
f
of interest to the consumer. Easily understandable language and illustrations provide useful background
knowledge for shopping in today's technological marketplace.
Order the above NBS publications from: Superintendent of Documents, Government Printing Office,
Washington. DC 20402.
Order the following NBS publications—FIPS and NBSIR 's—from the National Technical Information Ser-
vice, Springfield, VA 22161.
Federal Information Processing Standards Publications (FIPS PUB) — Publications in this series collectively
I constitute the FederalInformation Processing Standards Register. The Register serves as the official source of
information in the Federal Government regarding standards issued by NBS pursuant to the Federal Property
,
and Administrative Services Act of 1949 as amended. Public Law 89-306 (79 Stat. 1127), and as implemented
by Executive Order 1 1717 (38 FR 12315, dated May 11, 1973) and Part 6 of Tide 15 CFR (Code of Federal
Regulations).

NBS Interagency Reports —

A sp)ecial series of interim or final reports on work {performed by NBS
(NBSIR)
for outside sponsors (both government and non-government). In general, initial distribution is handled by the
sponsor; public distribution is by the National Technical Information Service, Springfield, VA 22161, in paper
icopy or microfiche form.
U.S. Department of Commerce
National Bureau of Standards
Gaithersburg, MD 20899

Official Business
Penalty for Private Use $300