You are on page 1of 31

12

AIr IensIfy £quafIon


12.1 Intruductiun
In applying aii buoyancy coiiections to weighing, an equation to be used in calculating the density of
aii is iequiied unless one makes a diiect deteimination.
The aii density,
a
, is typically deteimined fiom an equation of state foi moist aii. °In the 1970s, .,
it was appieciated that the equation-of-state itself has gieat impoitance and that seveial such equations
weie in wide use. Fuitheimoie, it had not been demonstiated expeiimentally that any of the equations-
of-state in use weie adequate foi actual mass compaiisons."
1
In 1977, Jones made a defnitive deiivation of °a semi-empiiical equation-of-state based on up-to-
date data."
1
The equation was developed quickly and was published in 1978.
2
The equation developed by Jones, °with minoi changes,"
1
was endoised foi use in mass metiology by
CIPM (Comité Inteinational des Poids et Mesuies) in 1981.
3
The equation given in Ref. 3 is now iefeiied
to as the CIPM-81 equation-of-state foi moist aii and is used foi mass metiology by most national
laboiatoiies. Use of CIPM-81 instead of its piedecessoi
2
makes a negligible change in ioutine mass
calibiations.
1
The effcacy of CIPM-81 and its piedecessoi was tested by deteimining the mass diffeience between
two nominally equal weights with and without ieliance on the equation-of-state. The measuiement
without ieliance on the equation-of-state was typically made in vacuum. Results of this type of compai-
ison done at the Physikalisch-Technische Bundesantalt (PTB)
4
agieed to within the expected unceitainty,
1 10
-4
in aii density,
a
.
The following is quoted fiom an excellent papei by Davis:
5
Recently, Balhoin ¦Balhoin, R., Beiucksichtigung de Luftdichte duich Wagung beim Masseveigleich,
Kochsiek, M., ed., Massebestimmung hohei Genauigkeit, PTB (F.R.G.) Ber., Me-60, 65, June 1984]
has compaied iesults obtained by using the CIPM equation and diiect measuiements based on
Aichimedes` piinciple (the lattei measuiements involve vacuum weighing to fnd mass diffeiences
independent of a buoyancy coiiection). His expected expeiimental unceitainty is at the level of that
estimated by Jones foi unceitainties in the equation-of-state. Balhoin fnds no unexpected iesults.
This is an impoitant confimation because Balhoin has measuied aii density by a buoyancy method.
Similai iesults, but at a somewhat incieased unceitainy, have been obtained in othei laboiatoiies as
well.
Davis
6
pointed out in 1992 that in the 10 oi so yeais since the publication of the BIPM/CIPM equation
(iefeiied to as CIPM-81) one of the most impoitant constant paiameteis in the equation, the molai gas
constant, had become bettei known; in addition, updated values foi some constant paiameteis had
become available.
The Comité Consultatif poui las Masse et les giandeuis appaientees (CCM), on the advice of the
Woiking Gioup on Density, consideied it woithwhile in 1991 to amend seveial of the constant paiameteis
in the CIPM-81 equation; the basic foim of the equation and the piinciples by which it was deiived
© 2002 by CRC Press LLC
iemained unchanged. The CCM iecommended that the 1981 equation incoipoiating the amended
paiameteis be designated as the °1981/1991 equation foi the deteimination of the density of moist aii."
The CIPM accepted the amended equation in 1991 (abbieviated by Davis
6
as the °1981/1991 equation").
The 1981/1991 equation is valid ovei the same iange of piessuie, tempeiatuie, ielative humidity (oi dew-
point tempeiatuie), and caibon dioxide mole fiaction as the 1981 equation.
Davis
6
concluded that:
1. Aii densities calculated using the 1981/1991 equation do not diffei signifcantly fiom those cal-
culated fiom the 1981 equation.
2. The oveiall unceitainty of the 1981/1991 equation is not signifcantly impioved ovei that of the
1981 equation.
3. The ieason foi making the changes is to ensuie that the values foi all constant paiameteis used
in the equation weie the best cuiiently available.
In this chaptei, the development of Jones
2
will be piesented in detail as backgiound mateiial. Then,
the 1981 equation and the 1981/1991 equation will be piesented and discussed. The BIPM equation
should be used in piactice.
12.2 Deve!upment ul the Junes Air Density Equatiun
The ideal gas equation,
(12.1)
ielates the total piessuie, P, the total volume, V, and the absolute tempeiatuie, T, of an ideal gas oi a
mixtuie of ideal gases. The numbei of moles of the gas oi the mixtuie of gases is n, and R is the univeisal
gas constant oi molai gas constant.
In teims of density, , iathei than volume, Eq. (12.1) becomes
(12.2)
wheie M is the moleculai weight of the gas oi the appaient moleculai weight of the mixtuie.
Foi a mixtuie of diy aii (indicated by the subsciipt a) and watei vapoi (indicated by the subsciipt w),
and M aie iespectively, the density and appaient moleculai weight of the aii-watei vapoi mixtuie.
Because
(12.3)
wheie m is the mass of the mixtuie and n is the numbei of moles of the mixtuie,
(12.4)
We now intioduce the watei vapoi mixing iatio, r:
(12.5)
and designate the iatio M
w
/M
a
by , wheieby Eq. (12.4) becomes
(12.6)
PV nRT,
P RT M,
M m n m m n n
a w a w
,
M n M n M n n M n M n M n n
a a w w a w a w w a a w a
1 1 .
r n M n M
w w a a
mass of watei vapoi mass of diy aii ,
M M r r
a
1 1 .
© 2002 by CRC Press LLC
We now substitute Eq. (12.6) in Eq. (12.2) and note that the effective watei vapoi piessuie, e , in moist
aii is defned
7
by
(12.7)
Then
(12.8)
Eq. (12.8) is the ideal gas equation foi a mixtuie of diy aii and watei vapoi with watei vapoi piessuie
of e . If the aii-watei vapoi mixtuie behaved as a mixtuie of ideal gases,
(12.9)
wheie Z is the compiessibility factoi.
Since a mixtuie of aii and watei vapoi is not ideal, the magnitude of the nonideality is ießected in
the depaituie of Z fiom 1. Eq. (12.9) then becomes
(12.10)
Eq. (12.10) is the real gas equation foi a mixtuie of diy aii and watei vapoi. By ieaiiangement of
Eq. (12.10), the expiession foi the aii density is
(12.11)
12.2.1 Parameters in the Junes Air Density Equatiun
12.2.1.1 Lniversa! Gas Cunstant, R
The value of the molai gas constant, R, listed in a compilation by Cohen and Tayloi,
8
is 8.31441 ±
0.00026 J/K/mol. Quinn et al.
9
made a new deteimination of R by measuiing the speed of sound in aigon
by means of an acoustic inteifeiometei. Theii value was 8315.73 ± 0.17 J/K/kmol. Gammon
10
deduced
a value of R fiom measuiements of the speed of sound in helium; his latei iepoited value
11
is 8315.31 ±
0.35 J/K/kmol, which, consideiing the unceitainties, was in close agieement with the Quinn et al. value.
Rowlinson and Tildesley
12
inteipieted the expeiimental measuiements of Quinn et al. and aiiived at a
value of 8314.8 ± 0.3 J/K/kmol, which is in close agieement with the Cohen and Tayloi value within the
unceitainties attached to the values. All of the unceitainties listed aie estimates of standaid deviation.
Jones
2
chose to use the Cohen and Tayloi value °with the iealization that in the futuie it might be
ieplaced by a new value." Such a new value has appeaied. Moldovei et al.
13,14
iepoited a new expeiimental
value of the gas constant of 8.314471 ± 0.000014 J/mol/K, wheie the unceitainty quoted is the estimate
of the standaid deviation. This unceitainty is smallei by a factoi of 5 than that of eailiei values. The
Moldovei et al. value is the piefeiied value and will be used thioughout this chaptei.
12.2.1.2 Apparent Mu!ecu!ar Weight ul Air, M
u
The appaient moleculai weight of diy aii, M
a
, is calculated as a summation, , using the ielationship
(12.12)
wheie each M
i
is the moleculai weight of an individual constituent and x
i
is the coiiesponding mole
fiaction, the iatio of the numbei of moles of a constituent to the total numbei of moles in the mixtuie.
e rP r .
P RT M e P
a
1 1 1 .
P RT M e P Z
a
1 1 1 1,
P RTZ M e P
a
1 1 1 .
PM RTZ e P
a
1 1 .
M M x
a i
k
i i
,
© 2002 by CRC Press LLC
The moleculai weights and typical mole fiactions of the constituents of diy aii aie tabulated in
Table 12.1. Othei constituents aie piesent in abundances that aie negligible foi the piesent application.
The values of the atomic weights of the elements
15
aie based on the caibon-12 scale. The moleculai
weights aie taken to be the sums of the atomic weights of the appiopiiate elements.
The value foi the abundance of oxygen is taken fiom Ref. 16.
The value foi the abundance of caibon dioxide is taken fiom an unpublished compilation of data on
atmcspheric concentiation of caibon dioxide at seven locations thioughout the woild. It must be empha-
sized that 0.00033 was the mole fiaction of CO
2
in the atmospheie and should be consideied a backgrcund
value. The mole fiaction of CO
2
in laboiatoiies, which is, of couise, the value of inteiest heie, is in geneial
gieatei than 0.00033 and is variable. Foi example, thiee samples of aii taken fiom a glove box in the
Mass Laboiatoiy at the National Buieau of Standaids (now NIST) had a mean value of 0.00043, and
foui samples of laboiatoiy aii taken at the National Centei foi Atmospheiic Reseaich in Bouldei, CO
had a mean value of 0.00080. Cleaily, then, the optimum utilization of the aii density calculation would
necessitate a measuiement of CO
2
abundance of an aii sample taken at the time of the laboiatoiy
measuiements of inteiest.
One of the options one has in dealing with the vaiiability of CO
2
abundance is to select a iefeience
level, foi example, 0.00033 oi 0.00043, and to piovide an adjustment to M
a
to account foi known
depaituies fiom the iefeience level.
Gluekauf,
17
in discussing the vaiiation of the abundance of oxygen in the atmospheie, stated that °all
majoi vaiiations of the O
2
content must iesult fiom the combustion of fuel, fiom the iespiiatoiy exchange
of oiganisms, oi fiom the assimilation of CO
2
in plants. The fist piocess does not iesult in moie than
local changes of O
2
content, while the lattei two piocesses, though locally alteiing the CO
2
/O
2
iatio, leave
theii sum unchanged."
The constancy of the sum is expiessed by the equation (foi convenience, the subsciipt i has been
ieplaced by the chemical symbol):
(12.13)
The contiibution of O
2
and CO
2
to the appaient moleculai weight of diy aii is
(12.14)
Fiom Eq. (12.13),
(12.15)
and
(12.16)
Theiefoie,
(12.17)
That is, the vaiiation in M
a
due to a vaiiation in CO
2
abundance is equal to 12.011 (the atomic weight
of caibon) multiplied by the vaiiation in CO
2
abundance.
x x
CO O
constant
2 2
0 20979 . .
M x M x x x
O O CO CO O CO 2 2 2 2 2 2
31 9988 44 0098 . . .
x x
O CO 2 2
0 20979 . ,
M x M x x
O O CO CO CO 2 2 2 2 2
12 011 6 7130 . . .
M M x M x
x
a O O CO CO
CO
2 2 2 2
2
12 011 . .
© 2002 by CRC Press LLC
The vaiiation in M
a
due to the diffeience between the iefeience levels of CO
2
abundance 0.00033 and
0.00043 is thus 0.0012 g/mol, which coiiesponds to a ielative vaiiation of 41 ppm in M
a
and a coiie-
sponding vaiiation of 41 ppm in aii density.
The adjusted M
a
accounting foi the depaituie of the CO
2
abundance fiom the iefeience level of 0.00033
becomes
(12.18)
wheie M
a033
is the appaient moleculai weight of diy aii with a CO
2
mole fiaction of 0.00033.
The value of the abundance of aigon in diy aii, 0.00916, is that calculated fiom the mass spectiometiic
deteimination of the iatio of aigon to aigon and nitiogen by Hughes.
18
The value foi the abundance of nitiogen was aiiived at by the usual piactice of infeiiing nitiogen
abundance to be the diffeience between unity and the sum of the mole fiactions of the othei constituents.
The abundances of the constituents neon thiough nitious oxide in Table 12.1 weie taken to be equal
to the paits pei volume concentiation in the U.S. Standaid Atmospheie, 1976.
19
Fiom the data of Table 12.1, the appaient moleculai weight of diy aii with a CO
2
mole fiaction of
0.00033 is calculated by Eq. (12.12) to be 28.963. Foi diy aii with a CO
2
mole fiaction of 0.00043, the
appaient moleculai weight is calculated to be 28.964.
12.2.1.3 Cumpressibi!ity Factur, Z
The compiessibility factoi was computed using the viiial equation of state of an aii-watei vapoi mixtuie
expiessed as a powei seiies in iecipiocal molai volume, 1/v,
(12.19)
and expiessed as a powei seiies in piessuie,
(12.20)
wheie B
mix
and B
mix
aie second viiial coeffcients and C
mix
and C
mix
aie thiid viiial coeffcients foi the
mixtuie.
The viiial coeffcients of the piessuie seiies aie ielated to the viiial coeffcients of the volume powei
seiies by
(12.21)
TABLE 12.1 Composition of Diy Aii
Abundance Moleculai
Constituent (mole fiaction) Weight
Nitiogen 0.78102 28.0134
Oxygen 0.20946 31.9988
Caibon dioxide 0.00033 44.0098
Aigon 0.00916 39.948
Neon 0.00001818 20.179
Helium 0.00000524 4.00260
Kiypton 0.00000114 83.80
Xenon 0.000000087 131.30
Hydiogen 0.0000005 2.0158
Methane 0.0000015 16.0426
Nitious oxide 0.0000003 44.0128
M M x
a a033 2
12 011 0 00033 . . ,
CO
Z Pv RT B v C v 1
2
mix mix
,
Z Pv RT B P C P 1
2
mix mix
,
B B RT
mix mix
© 2002 by CRC Press LLC
and
(12.22)
Each mixtuie viiial coeffcient is a function of the mole fiactions of the individual constituents and
the viiial coeffcients foi the constituents. The viiial coeffcients aie functions of tempeiatuie only.
Using the viiial coeffcients piovided by Hyland
20
and Wexlei,
21
a table of compiessibility factoi, Z, foi
CO
2
-fiee aii, Table 12.2, has been geneiated. Table 12.2 is applicable to moist aii containing ieasonable
amounts of CO
2
.
Alteinatively, Z can be calculated using the following equations:
Foi P in pascals and t in ºC,
(12.23)
Foi P in psi (pounds pei squaie inch) and t in ºC,
(12.24)
U is ielative humidity (foi example, U · 50 is equal to 50% ielative humidity).
Foi tempeiatuies and/oi piessuies outside the iange of Table 12.2, the table of compiessibility factoi
of moist aii (also CO
2
-fiee) in the Smithscnian Metecrclcgical Tables
22
can be used, with some loss of
piecision since the listing theie is to the fouith decimal place.
12.2.1.4 Ratiu ul the Mu!ecu!ar Weight ul Water tu the Mu!ecu!ar Weight ul Dry Air,
The moleculai weight of watei is 18.0152.
15
The iatio, , of the moleculai weight of watei to that of diy
aii is, theiefoie, 0.62201 foi diy aii with a CO
2
mole fiaction of 0.00033. Foi diy aii with a CO
2
mole
fiaction of 0.00043, the iatio, , is 0.62199.
12.2.1.5 Ellective Water Yapur Pressure, e
Since e is the effective vapoi piessuie of watei in mcist air, a woid of caution with iegaid to infeiiing e
fiom measuiements of ielative humidity is in oidei. Relative humidity, U, can be defned
23
by
(12.25)
wheie e
s
is the effective satuiation vapoi piessuie of watei in moist aii. The intioduction of a second gas
(aii in this case) ovei the suiface of watei incieases the satuiation concentiation of watei vapoi above
C C B RT
mix mix mix
2
2
.
Z P P
t Pt
P t t
Pt P t
0 99999 5 8057 10 2 6402 10
3 3297 10 1 2420 10
2 0158 10 2 4925 10
6 2873 10 5 4174 10
3 5 10 5 0 10
9 16 2
7 10
18 2 9 2
13 2 21 2 2
7 9
2
. . .
. .
. .
. .
. . . RH RH
Z P P
t Pt
P t t
Pt P t
U U
0 99999 4 0029 10 1 2551 10
3 3297 10 8 5633 10
9 5826 10 2 4925 10
4 3349 10 2 5753 10
3 5 10 5 0 10
5 8 2
7 7
11 2 9 2
9 2 13 2 2
7 9
2
. . .
. .
. .
. .
. . .
U e e
s
100%,
© 2002 by CRC Press LLC
the suiface of the watei; the effective satuiation vapoi piessuie of watei, e
s
, is gieatei than the satuiation
vapoi piessuie of puie phase (i.e., watei vapoi without the admixtuie of aii oi any othei substance), e
s
.
12.2.1.6 Enhancement Factur, f
This °enhancement" of watei vapoi piessuie is expiessed by the enhancement factoi, f, which is defned by
(12.26)
TABLE 12.2 Compiessibility Factoi, Z, Calculated Using Eq. (12.23)
t P
RH(%)
(ºC) (Pa) 0 25 50 75 100
19.0 95000 0.99964 0.99963 0.99961 0.99960 0.99957
100000 0.99962 0.99961 0.99959 0.99958 0.99956
101325 0.99962 0.99960 0.99959 0.99957 0.99955
105000 0.99960 0.99959 0.99958 0.99956 0.99954
110000 0.99958 0.99957 0.99956 0.99954 0.99952
20.0 95000 0.99965 0.99964 0.99962 0.99960 0.99958
100000 0.99963 0.99962 0.99960 0.99958 0.99956
101325 0.99963 0.99962 0.99960 0.99958 0.99956
105000 0.99961 0.99960 0.99958 0.99957 0.99954
110000 0.99959 0.99958 0.99957 0.99955 0.99953
21.0 95000 0.99966 0.99965 0.99963 0.99961 0.99958
100000 0.99964 0.99963 0.99961 0.99959 0.99956
101325 0.99964 0.99962 0.99961 0.99959 0.99956
105000 0.99962 0.99961 0.99959 0.99957 0.99955
110000 0.99960 0.99959 0.99958 0.99956 0.99953
22.0 95000 0.99967 0.99965 0.99963 0.99961 0.99958
100000 0.99965 0.99964 0.99962 0.99960 0.99957
101325 0.99965 0.99963 0.99961 0.99959 0.99956
105000 0.99963 0.99962 0.99960 0.99958 0.99955
110000 0.99962 0.99960 0.99958 0.99956 0.99954
23.0 95000 0.99968 0.99966 0.99964 0.99962 0.00059
100000 0.99966 0.99964 0.99962 0.99960 0.99957
101325 0.99965 0.99964 0.99962 0.99960 0.99957
105000 0.99964 0.99963 0.99961 0.99958 0.99956
110000 0.99963 0.99961 0.99959 0.99957 0.99954
24.0 95000 0.99968 0.99967 0.99965 0.99962 0.99959
100000 0.99967 0.99965 0.99963 0.99961 0.99958
101325 0.99966 0.99965 0.99963 0.99960 0.99957
105000 0.99965 0.99964 0.99962 0.99959 0.99956
110000 0.99964 0.99962 0.99960 0.99957 0.99954
25.0 95000 0.99969 0.99968 0.99965 0.99962 0.99959
100000 0.99968 0.99966 0.99964 0.99961 0.99958
101325 0.99967 0.99966 0.99963 0.99961 0.99957
105000 0.99966 0.99964 0.99962 0.99960 0.99956
110000 0.99965 0.99963 0.99961 0.99958 0.99955
26.0 95000 0.99970 0.99968 0.99966 0.99963 0.99959
100000 0.99969 0.99967 0.99964 0.99961 0.99958
101325 0.99968 0.99966 0.99964 0.99961 0.99957
105000 0.99967 0.99965 0.99963 0.99960 0.99956
110000 0.99966 0.99964 0.99961 0.99959 0.99955
f e e
s s
.
© 2002 by CRC Press LLC
A published, expeiimentally deiived value of f
24
at 20ºC and 100,000 Pa, is 1.00400. Theiefoie, the
common piactice of infeiiing e fiom measuied U and tabulated value of e
s
intioduces a signifcant
eiioi in e if f has been ignoied. The coiiesponding ielative eiioi in at 20ºC, 101,325 Pa, and 50%
ielative humidity is about 1.7 10
-5
.
f is a function of tempeiatuie and piessuie. Jones
2
has ftted Hyland`s values of f
24
to a thiee-paiametei
equation in the piessuie (P, Pa) and tempeiatuie (t, ºC) ianges of inteiest in national standaids laboiatoiies.
The iesulting equation is
(12.27)
Ovei the tempeiatuie iange 19.0 to 26.0ºC and the piessuie iange 69,994 Pa (10.152 psi) to 110,004 Pa
(15.955 psi), f ianges fiom 1.0031 to 1.0045. The maximum vaiiation of f fiom a nominal value of 1.0042
is equal to 0.11% of the nominal value. The coiiesponding ielative vaiiation of aii density is equal to
0.00040%, which is negligible.
12.2.1.7 Saturatiun Yapur Pressure ul Water, e
s
The expiession foi e is found by combining Eqs. (12.25) and (12.26) to be
(12.28)
The systematic ielative unceitainties in due to the unceitainties assigned to f
24
and e
s
25
aie appiox-
imately ±1 10
-6
and ±2 10
-7
, iespectively.
The e
s
data of Besley and Bottomley
26
in the tempeiatuie iange 288.15 K (15ºC) to 298.04 K (24.89ºC)
and calculated values foi the iemaindei of the tempeiatuie iange to 301.15 K (28ºC) have been ftted
2
to a two-paiametei equation. The iesulting equation is
(12.29)
wheie e · 2.7182818. is the base of Napeiian logaiithms.
Values calculated using Eq. (12.29) aie suffciently close to expeiimental and calculated values of e
s
,
within ±0.1%, to be used in the piesent application.
Calculated values of e
s
expiessed in mmHg aie tabulated in Table 12.3.
12.2.1.8 Carbun Diuxide Abundance, x
CO2
In a pievious section it was stated that the CO
2
abundance in laboiatoiy aii is in geneial vaiiable. A
vaiiation of 0.0001 in CO
2
mole fiaction is equivalent to a ielative vaiiation of 4 10
-5
in calculated aii
density. The CO
2
abundance should be known foi optimum utilization of the aii density calculation.
Eq. (12.18) can be used to adjust M
a
foi depaituies of CO
2
abundance fiom the iefeience level, 0.00033.
12.2.2 The Junes Air Density Equatiun
2
By combining Eqs. (12.11) and (12.28) and substituting M
w
/M
a
foi , the aii density equation developed
by Jones
2
becomes
(12.30)
By substituting the Moldovei et al. value of R,
13,14
8.314471 J/mol/K, and the value of 18.0152 foi M
w
,
Eq. (12.30) becomes
f P t 1 00070 3 113 10 5 4 10
8 7 2
. . . .
e U fe
s
100 .
e e
s
T
1 7526 10
11
5315 56
. ,
.
PM RTZ M M U fe P
a w a s
1 1 100 .
© 2002 by CRC Press LLC
(12.31)
wheie
(12.32)
Foi T · 293.15 K (20ºC), P · 101,325 Pa, 50% ielative humidity, and M
a
· 28.963 g/mol, the aii
density calculated using Eq. (12.31) is 1.1992 kg/m
3
· 0.0011992 g/cm
3
· 1.1992 mg/cm
3
.
Foi M
a
of 28.964 g/mol (x
CO2
· 0.00043) and the same values of the othei paiameteis above, ·
1.1993 kg/m
3
· 0.0011993 g/cm
3
· 1.1993 mg/cm
3
.
TABLE 12.3 Satuiation Vapoi Piessuie
of Watei Calculated Using Eq. 12.29
t e
s
t e
s
(ºC) (mmHg) (ºC) (mmHg)
19.0 16.480 22.6 20.565
19.1 16.583 22.7 20.691
19.2 16.686 22.8 20.817
19.3 16.790 22.9 20.943
19.4 16.895
19.5 17.000 23.0 21.071
19.6 17.106 23.1 21.199
19.7 17.212 23.2 21.327
19.8 17.319 23.3 21.457
19.9 17.427 23.4 21.587
23.5 21.718
20.0 17.535 23.6 21.849
20.1 17.644 23.7 21.982
20.2 17.753 23.8 22.114
20.3 17.863 23.9 22.248
20.4 17.974
20.5 18.085 24.0 22.383
20.6 18.197 24.1 22.518
20.7 18.309 24.2 22.653
20.8 18.422 24.3 22.790
20.9 18.536 24.4 22.927
24.5 23.065
21.0 18.650 24.6 23.204
21.1 18.765 24.7 23.344
21.2 18.880 24.8 23.484
21.3 18.996 24.9 23.625
21.4 19.113
21.5 19.231 25.0 23.767
21.6 19.349 25.1 23.909
21.7 19.467 25.2 24.052
21.8 19.587 25.3 24.196
21.9 19.707 25.4 24.341
25.5 24.487
22.0 19.827 25.6 24.633
22.1 19.949 25.7 24.780
22.2 20.071 25.8 24.928
22.3 20.193 25.9 25.077
22.4 20.317
22.5 20.441 26.0 25.226
0 000120272 1 1 18 0152 100 . . , PM TZ M U fe P
a a s
M x
a
28 963 12 011 0 00033
2
. . . .
CO
© 2002 by CRC Press LLC
12.2.3 Lncertainties in Air Density Ca!cu!atiuns
We now ietuin to the aii density equation as expiessed by Eq. (12.30).
The unceitainty piopagation equation
27
as it applies to aii density is
(12.33)
wheie SD is the estimate of standaid deviation of , the Y
i
aie quantities on the iight-hand side of
Eq. (12.30), and the SD
i
aie the estimates of standaid deviation of the quantities on the iight-hand
side of Eq. (12.30).
12.2.3.1 Lncertainties in Quantities Other than P, T, L, and x
CO2
We fist concentiate on the unceitainties in the quantities othei than the measuiements of P, T, U, and
x
CO2
in oidei to fnd the minimum unceitainty in if P, T, U, and x
CO2
weie measuied peifectly. This
minimum unceitainty can then be consideied the limitation on the deteimination of using the aii
density equation. We shall iefei to these quantities loosely as nonenviionmental quantities.
12.2.3.1.1 Partíal Derívatíves, ( / Y
í
), for the Nonenvíronmental Quantítíes
The paitial deiivatives of with iespect to the non-enviionmental quantities aie now taken and evaluated
using the following values of the paiameteis:
The value of
a
calculated using the quantities above is 1.1992 kg/m
3
.
The paitial deiivatives foi the nonenviionmental quantities aie
Foi convenience, units have not been given foi the paitial deiivatives.
SD Y
i i i
2 2 2
SD ,
P
M
R
T
U
f
e
x
a
s
101325
28 963
8314 471
293 15
18 0152
50
1 0040
2338 80
0 00033
2
Pa
g mol
J kmol K
K
Z·0.99960
M g mol
Pa
w
CO
.
.
.
.
.
.
.
M RTZ P Ufe
R R TZ M M Ufe PM
Z RTZ M M Ufe PM
f Ue RTZ
a s
a w s a
a w s a
s
1 100 4 1106 10
1 100 1 4423 10
1 100 1 1997
100
2
2 4
2
.
.
.
M M
e Uf RTZ M M
M RTZ Ufe
w a
s w a
w s
5 2546 10
100 2 2557 10
1 100 4 8189 10
3
6
4
.
.
.
© 2002 by CRC Press LLC
12.2.3.1.2 Uncertaíntíes ín the Nonenvíronmental Quantítíes (SD
í
)
The estimates of the standaid deviations foi the nonenviionmental quantities aie given below:
12.2.3.1.3 Products of the Partíal Derívatíves and the Estímates of Standard Devíatíon,
( / Y
í
)· (SD
í
), for the Nonenvíronmental Quantítíes
The pioduct of the paitial deiivative and the estimate of standaid deviation foi each of the nonenviion-
mental quantities is given below:
The above pioducts weie squaied and summed as indicated in Eq. (12.33) and the squaie ioot was
taken. The iesulting estimate of the standaid deviation of the aii density,
a
, is 4 10
-5
kg/m
3
. The
estimate of standaid deviation of
a
is dominated by the unceitainty in the appaient moleculai weight
of diy aii, M
a
.
Theiefoie, the minimum unceitainty in the calculation of aii density using the aii density equation,
that which can be consideied to be intiinsic to the equation, is an estimate of standaid deviation of 4
10
-5
kg/m
3
oi 33 ppm in the piesent example. Note that this does not include unceitainties in the
measuiements of piessuie, tempeiatuie, ielative humidity, and mole fiaction of CO
2
.
12.2.3.2 Lncertainties in the Envirunmenta! Quantities
12.2.3.2.1 Partíal Derívatíves, / Y
í
, for the Envíronmental Quantítíes
The paitial deiivatives foi the enviionmental quantities, P, T, U, and x
CO2
aie
M
R
Z
f
e
M
a
s
w
1 10
1 4 10
5 7 10
7 10
3 9 10
2 10
3
2
6
5
2
4
g mol
J kmol K
Pa
g mol
.
.
.
M
R
Z
f
e
M
a
s
w
4 10
2 0 10
6 8 10
4 10
8 8 10
1 10
5 3
6 3
6 3
7 3
8 3
7 3
kg m
kg m
kg m
kg m
kg m
kg m
.
.
.
P M RTZ
T RT Z PM M M Ufe
U RTZ fe M M
a
a a w s
s w a
1 1888 10
1 100
4 0908 10
1 100 1 0551 10
5
2
3
4
.
.
.
© 2002 by CRC Press LLC
Again, foi convenience, the units foi the paitial deiivatives have not been given.
Foi atmospheies in which the mole fiaction of CO
2
, x
CO2
, is othei than the iefeience value of 0.00033,
(12.32)
and
Then,
12.2.3.2.2 Uncertaíntíes ín the Envíronmental Quantítíes (SD
í
)
The estimates of the standaid deviations of each of the foui enviionmental paiameteis aie
12.2.3.2.3 Products of the Partíal Derívatíves and the Estímates of Standard Devíatíon,
( / Y
í
)· (SD
í
), for the Envíronmental Quantítíes
The pioduct of the paitial deiivative and the estimate of standaid deviation foi each of the enviionmental
paiameteis is
The above pioducts weie squaied and summed, and the squaie ioot was taken. The iesulting estimate
of the standaid deviation of the aii density,
a
, is 3 10
-4
kg/m
3
.
The laigest unceitainty in is that due to piessuie measuiement, and the next laigest is that due to
tempeiatuie measuiement. The estimates of unceitainty aie based on state-of-the-ait measuiements of
the vaiiables.
The oveiall unceitainty contiibuted by all of the quantities in the aii density equation is 3 10
-4
kg/m
3
.
12.2.4 Lse ul Cunstant Ya!ues ul F, Z, and M
u
in the Air Density Equatiun
By consideiing the expected vaiiations in piessuie, tempeiatuie, and ielative humidity in the labo-
iatoiy, it might be possible to use constant values of f, Z, and M
a
.
M x
a
28 963 12 011 0 00033
2
. . . .
CO
M x
a CO2
12 011 . .
x M M x
a a CO CO 2 2
2
4 1106 10 12 011
0 49372
. .
. .
P
T
U
x
20 5
0 05
0 51
0 00005
2
.
.
.
.
Pa
K
CO
P
T
U
x
2 4 10
2 0 10
5 4 10
2 5 10
4 3
4 3
5 3
2
5 3
.
.
.
.
kg m
kg m
kg m
kg m
CO
© 2002 by CRC Press LLC
Foi example, in the Mass Laboiatoiy of NIST, constant values of f (1.0042), Z (0.99966), and M
a
(28.964) aie consideied to be adequate. With these values of f, Z, and M
a
, the iesulting equation foi
calculating aii density is, foi P in pascals and absolute tempeiatuie T · 273.15 - t (ºC),
(12.34)
Foi P in psi and t in ºC,
(12.35)
Foi P · 101,325 Pa · 14.69595 psi, T · 293.15 (t · 20ºC), U · 50, and e
s
· 2338.80 Pa,
a
· 1.1992 kg/m
3
·
0.0011992 g/cm
3
· 1.1992 mg/cm
3
foi both Eqs. (12.34) and (12.31).
12.3 CIPM-81 Air Density Equatiun
The CIPM-81 equation
3
is
(12.36)
wheie p is piessuie, x
v
is the mole fiaction of watei vapoi in moist aii, M
v
is the molai mass of watei
vapoi in moist aii (M
w
in the Jones development), and M
a
is molai mass of diy aii (M
a
in the Jones
development). The mole fiaction of watei vapoi, x
v
, is equal to (U/100) (fe
s
/P) and is deteimined fiom
the ielative humidity, U, oi the dew-point tempeiatuie.
12.4 CIPM 1981]1991 Equatiun
The CIPM 1981/1991
6
equation is the same as the CIPM-81 equation, Eq. (12.36). Davis
6
has tabulated
the amended constant paiameteis appiopiiate to the CIPM 1981/1991 equation. Davis stated that aii
densities calculated fiom the 1991 paiameteis aie smallei by about 3 paits in 10
5
ielative to calculations
using the 1981 paiameteis, and that the oveiall unceitainty foi aii density calculated using the 1981/1991
equation is essentially the same as if the 1981 equation weie used. Davis
6
noted that ITS-90 should be used.
12.5 Recummendatiun
The diffeience between the aii density calculated using the CIPM 1981/1991 and that calculated using
the CIPM-81 aie well within the piactical unceitainty. If one piefeis, the CIPM 1981/1991 equation can
be used.
12.6 Direct Determinatiun ul Air Density
12.6.1 Intruductiun
The aii density in a balance case was deteimined diiectly by Koch et al.
28
using a simple device that
enabled mass compaiisons in aii without the need to coiiect foi aii buoyancy. The device was an
evacuated canistei that was weighed on a laboiatoiy balance with eithei of two masses inside.
The expeiimental method was used to deteimine the mass diffeience between two stainless steel weights
of widely diffeient densities. By a simple weighing of each of the two objects, the mass diffeience
deteimination and the volume diffeience of the two weights weie used to deteimine diiectly the density
of the aii in the balance case.
a s
T P Ue 0 0034847 0 0037960 . . .
a s
t P Ue 24 026 273 15 0 0037960 . . .
pM x M M ZRT
a v v a
1 1 ,
© 2002 by CRC Press LLC
12.6.2 Experimenta! Prucedure
Two stainless steel weights weie used. One was hollow and of low density with mass of M
H
and exteiioi
volume of V
H
; the othei was solid and of high density with mass of M
S
and exteinal volume of V
S
.
The two weights weie weighed, successively, within an evacuated enclosuie of mass M
E
and exteinal
volume V
E
. Foi aii density,
a
, a foice equation can be wiitten:
(12.37)
wheie g is the local acceleiation due to giavity and M iepiesents the small mass diffeience indicated
on the optical scale of the balance, calibiated using platinum weights.
If the volume of the evacuated enclosuie, V
E
, and g aie constant thioughout the pioceduie, the mass
diffeience between the two weights:
(12.38)
is obtained without an aii buoyancy calculation.
If the mean densities of the two objects aie suffciently diffeient, the weights thus chaiacteiized can
be used to deteimine the density of the aii in the balance case. If m is the diffeience in aii weight (divided
by g) between M
H
and M
S
, and if
(12.39)
then,
(12.40)
oi
(12.41)
The values of aii density so deteimined could be compaied to those calculated using an existing aii
density algoiithm and measuiements of baiometiic piessuie, tempeiatuie, and ielative humidity.
Mass compaiisons using the evacuated canistei weie made on a commeicial single-pan balance with
aii damping and an optical scale. The balance had a capacity of 100 g and a iesolution of 10 g.
The weighings in aii of the two mass aitifacts foi checking the adequacy of the aii density algoiithm
weie all made on single-pan balances with optical scales. These balances had a capacity of 20 g and
iesolution of 1 g.
12.6.3 Resu!ts and Cunc!usiuns
The expeiimental and calculated values of aii density agieed thioughout to within 1.0 g/cm
3
(wheie
the noimal aii density was about 1.2 mg/cm
3
). The unceitainty was 0.08% of the aii density, at 1 standaid
deviation.
The expeiimental and calculated values of day-to-day ßuctuations in aii density agieed to within
0.5 g/cm
3
.
A mass deteimination of the following mateiials will have at most the ielative unceitainty indicated
(due solely to the aii density algoiithm) staiting fiom platinum-iiidium standaids:
Silvei 5 10
-8
Stainless steel 8 10
-8
Watei 9 10
-7
M M V g M M V M g
E H a E E S a E
,
M M M
H S
V V V
H S
,
m M V M V
S a S H a H
,
a
M m V .
© 2002 by CRC Press LLC
The measuiements showed that the assignment of the mass unit fiom one kilogiam to anothei can
be made with an unceitainty attiibutable to the aii density algoiithm of no gieatei than 0.10 mg/100 cm
3
in the volume diffeience of the kilogiams.
12.7 Experimenta! Determinatiun ul Air Density in Weighing
un a 1-kg Ba!ance in Air and in Yacuum
12.7.1 Intruductiun
Glasei et al.
29
made expeiimental deteiminations of aii density by compaiing two mass aitifacts of about the
same mass and suiface but diffeient well-known volumes on a 1-kg mass compaiatoi in aii and in vacuum.
The density of aii,
a
, was calculated fiom the diffeiences in °appaient mass" in aii and in vacuum.
Appaient mass, m , was defned as the diffeience between the mass of the aitifact, m, and the pioduct
of the volume, V, of the object and the aii density, (m - V
a
).
The following foimula was used to calculate
a
:
(12.42)
wheie m
v
is the diffeience of appaient mass in vacuum, (m
1
- m
2
), of the two aitifacts, m
a
is the
diffeience of appaient mass in aii, m
a
· (m
1
- m
2
)
a
, and V is the diffeience in volume of two aitifacts
(Nos. 1 and 2).
Suiface desoiption effects, loss of mass fiom suifaces of the aitifacts duiing evacuation, can be
minimized by designing the aitifacts to have as neaily equal suiface aieas as possible.
In geneial, the mass, m, of an aitifact changes due to the desoiption of mass m
s
by
(12.43)
Eq. (12.43) is then modifed to become
(12.44)
wheie m
s
is the diffeience in desoibed mass.
Desoiption inßuences weie estimated by the use of aitifact No. 1 and an aitifact, No. 3, of equal
volume to that of No. 1 but of diffeient suiface aiea.
12.7.2 Resu!ts and Cunc!usiuns
Weighings weie made of the thiee aitifacts in aii and in vacuum on a 1-kg mass compaiatoi. The aii
compaiisons weie coiiected foi buoyancy using the BIPM-iecommended aii density equation.
3
An estimation was made of the unceitainy of the aii density deteimined by use of the iecommended
equation and of the aii density deteimined in these expeiiments.
It was concluded that the ielative unceitainty foi the expeiimental measuiements (at a level of
1 standaid deviation) was 5 10
-5
; this value was calculated to be smallei by a factoi of 2.4 than that
associated with the iecommended equation. The unceitainties used foi the aii density equation, howevei,
can be questioned, paiticulaily the unceitainty foi the appaient moleculai weight of aii.
The suiface adsoiption and desoiption effects in vacuum and in aii weie found to be small compaied
with the expeiimental unceitainty.
Foi an aitifact of mass 1 kg of stainless steel compaied with one of platinum-iiidium, the buoyancy
coiiection is about 94 mg at an aii density of 1.2 kg/m
3
.
A ielative unceitainty of 5 10
-5
coiiesponds to an unceitainty of about 0.1 mg in the compaiison
of a stainless steel kilogiam with a platinum-iiidium kilogiam.
a v a
m m V,
m m m
v s
.
a v a s
m m m V,
© 2002 by CRC Press LLC
12.8 A Practica! Appruach tu Air Density Determinatiun
30
12.8.1 Intruductiun
Mass standaids aie used diiectly oi indiiectly with the electionic balance in many common measuiement
applications such as mass, density, volumetiic capacity, foice, and piessuie. These applications geneially
iequiie some knowledge of aii density. This section offeis the metiologist, bench scientist, and laboiatoiy
technician a piactical guide to the deteimination of aii density and the associated unceitainty estimate.
The measuiement piocess usually begins with a pioblem defnition and then an eiioi analysis. Many
pioblems aie iepetitive and set foith in papei standaids. Often, the metiologist oi expeiimentei is faced
with a measuiement pioblem that will be peifoimed a veiy limited numbei of times and foi which the
metiologist has little expeiience. Metiologists often peifoim tasks by iote as passed along by otheis oi
follow wiitten piesciiptions that aie without detail. These ciicumstances often lead to unnecessaiy woik,
the puichase of inappiopiiate equipment, and inaccuiate measuiement iesults and unceitainty estimates,
all of which give iise to economic ineffciencies. This section piovides a basis foi the deteimination of
aii density wheie coiiection foi the buoyant effect of aii is necessaiy.
Sometimes it is unnecessaiy foi the usei to deteimine aii density while at othei times the usei must
appeal to an aii density equation that iequiies the measuiement of othei paiameteis. This section examines
the use of aveiage aii density foi a given location and the appiopiiate measuiement of tempeiatuie,
baiometiic piessuie, and ielative humidity when one applies an equation. The Oiganization of Legal
Metiology (OIML) document R111
31
sets foith a fiamewoik foi laboiatoiy mass standaid specifcations
that foim a useful basis to examine the pioblems associated with the deteimination of aii density.
The piimaiy puipose of R111 is the cieation of a system foi iefeience weights that in noimal usage
aii density, that is, buoyancy coiiections, can be ignoied. This is accomplished by ielying on a veiy old
concept that in the past was iefeiied to as °appaient mass" and then, in R111, as the °conventional value
of the iesult of weighing in aii" (see Chaptei 15) as defned by OIML R33.
32
Unfoitunately, the small maximum peimissible eiiois (MPE) foi class E
1
and E
2
weights (laboiatoiy
mass standaids) and the depaituie of aii density fiom the standaid aii density of 1.2 mg/cm
3
foice the
usei to measuie mass (sometimes called °tiue mass") and compute the conventional value oi to make
tedious coiiections to obtain the coiiect conventional value.
Waining is given in R33 iegaiding the incoiiectness of conventional values when the weighings aie
peifoimed at aii densities diffeiing fiom 1.2 mg/cm
3
by moie than 10%. This waining applies to weights the
densities of which aie within the domain of R111, and may not be adequate foi objects outside this domain.
12.8.2 Air Density
In the use of foice machines, and in some instances in the use of dead-weight piessuie gauges, the
metiologist need only account foi the gioss effect of aii buoyancy, and theiefoie an aveiage aii density
foi a given location is adequate. In the °conventional value" usage one may not want to depait by moie
than 10% fiom the iefeience aii density, 1.2 mg/cm
3
. These applications aie easily accommodated by
consulting a weathei buieau foi a location aveiage aii density oi by calculating an aii density based on
the station elevation. One need only considei elevation as natuie limits local vaiiations in aii density
ielated to vaiiation in aii piessuie to ±3%. Laboiatoiy tempeiatuie and ielative humidity (RH) is usually
maintained neai 25ºC and 50% RH, iespectively. Noimal vaiiations fiom these values have a negligible
effect when making appioximate buoyancy coiiections.
Fiom the laboiatoiy elevation, h, in meteis, one can calculate the standaid piessuie, P, in millimeteis
of meicuiy by application of eithei of the following two foimulas deiived fiom data given in the
Smithscnian Metecrclcgical Tables
33
:
(12.45)
(12.46)
ln . . P h 6 6342205 0 00012157179
P h exp . . . 6 6342205 0 00012157179
© 2002 by CRC Press LLC
The piessuie unceitainty (at k · 1) is 0.44 mmHg. The standaid piessuie at sea level (0 m) is 760 mmHg
at 15ºC. The standaid piessuie is neai the uppei extieme of the natuial piessuie vaiiation of ±3%.
Theiefoie, the aveiage station piessuie, P
A
, calculated fiom P is P
A
P - 3% P. If the station elevation is
unknown it can be deteimined fiom many souices (topogiaphical maps); howevei, a neaiby aiipoit
elevation may be adequate. Otheiwise, an aiiciaft altimetei can be set to iead the aiifeld elevation and
then tianspoited to the laboiatoiy. The indicated elevation is unceitain (at k · 1) by 25 m. Foi the
ieadei`s convenience, a few calculated aii piessuies, P, fiom sea level to 2000 m aie piovided in Table 12.4.
Having obtained an aveiage station piessuie, one can apply it to an equation that expiesses the density
of aii in teims of the aii tempeiatuie, t, ielative humidity, U, and aii piessuie, P. The inteinational
community geneially accepts the equation developed by Jones
2
and now slightly modifed and known as
the CIPM 81/91 aii density equation.
6
A vaiiant of this equation is offeied heie that is simplei to use
and easiei to follow in discussion
35
:
(12.47)
wheie
a
is aii density in kg/m
3
, T is the aii tempeiatuie in kelvin, Z is the compiessibility factoi foi
CO
2
-fiee aii, P is ambient aii piessuie in pascals, U is the ielative humidity in %, and e
s
is the satuiation
vapoi piessuie of watei in pascals. e
s
is obtained by application of the foimula:
(12.48)
wheie is the base of natuial logaiithms. A value foi Z is obtained fiom the following foimula:
(12.49)
wheie P is in pascals, t in ºC, and U in %. Foi the ieadei`s convenience, Table 12.5 piovides values of Z
foi a few selected values of P, t, and U.
Foi many mass and mass-ielated measuiements such as density and volumetiic capacity, gieatei
knowledge of the aii density is iequiied along with an unceitainty estimate. Eq. (12.47) is appiopiiate
foi these applications but iequiies measuiement of aii piessuie, aii tempeiatuie, and ielative humidity.
TABLE 12.4 Elevation and Aii Piessuie
Elevation, Aii Piessuie,
meteis mmHg
0 (sea level) 760
300 733.4
600 707.2
900 (10% limit) 681.8
1200 657.4
1500 633.9
1609.3 625.5 (Denvei, CO, USA)
1800 611.2
2000 596.5
a s
TZ P Ue 0 0034836 0 0037960 . . ,
e
s
T
1 7526 10
11
5315 56
. ,
.
Z P P
t Pt
P t t
Pt P t
U U
0 99999 5 0857 10 2 6402 10
3 3297 10 1 2420 10
2 0158 10 2 4925 10
6 2873 10 5 4174 10
3 5 10 5 0 10
9 16 2
7 10
18 2 9 2
13 2 21 2 2
7 9 2
. . .
. .
. .
. .
. . ,
© 2002 by CRC Press LLC
A backgiound value foi CO
2
concentiation was assumed to be 430 ppm in the deiivation of Eq. (12.47).
Each paiametei sensoi must be calibiated and accompanied with an unceitainty estimate befoie an eiioi
piopagation can be peifoimed foi the calculated aii density. Sensoi calibiation will now be discussed
befoie pioceeding with the eiioi piopagation foi Eq. (12.47).
12.8.2.1 Temperature
Foi the most accuiate woik, theimometeis used to measuie aii tempeiatuie should be placed inside the
weighing chambei and close to the weight on the weighing pan. Of couise, the theimal histoiy 24 h piioi
to the weighing should be monitoied to assuie theimal equilibiium between the object undeigoing test
and the suiiounding aii.
36
If the data show a change gieatei than 0.5ºC/h, the local enviionment may not
be adequate foi applications involving laige buoyancy coiiections and wheie high accuiacy is iequiied.
Regaidless of the type of theimometei used, that is, platinum iesistance, theimistoi, meicuiy-in-glass,
oi theimocouple, it should be peiiodically tested by the usei, as shculd the pressure and humidity senscrs.
A single point check is not always adequate because it fails to check the iesponse slope of the sensoi.
Theiefoie, a moie thoiough theimometei check iequiies two widely sepaiated tempeiatuie iefeience
points that span the noimal laboiatoiy woiking tempeiatuies. Foitunately, natuie has piovided thiee
fxed iefeience tempeiatuie souices that aie easy and economical to access. They aie the melting point
of ice (0ºC), the tiiple point of watei (0.01ºC), and the steam condensation tempeiatuie (appioximately
100ºC). Techniques to implement these fxed points that aie quite simple and ieliable aie now desciibed
in detail.
12.8.2.2 Me!ting Puint ul Ice
The melting point of ice is achieved by packing fnely divided ice (ciushed ice) made fiom distilled watei
into a dewai ßask oi common Theimos' bottle to pievent iapid melting.
37
The theimometei to be
checked is not allowed to come into contact with the walls of the ßask while ice is packed by fim hand
piessuie aiound the theimometei; see Figuie 12.1. The theimometei stem must be immeised at least ten
diameteis (moie is bettei) except foi meicuiy-in-glass types. The lattei usually have an immeision line
oi may iequiie total immeision. Then, distilled watei is added to the ßask and the excess watei is foiced
out of the ßask by adding moie ciushed ice. The constituents should be in theimal equilibiium within
1
/2 h. Peiiodically, ciushed ice must to be added to displace liquid watei cieated by the melting ice. The
amount of liquid watei must be minimized.
Theimal equilibiium is confimed by noting a stable theimometei ieading foi 10 min. The theimom-
etei coiiection is
wheie t
ief
is the iefeience tempeiatuie 0ºC and t is the theimometei ieading in ºC. The iefeience
tempeiatuie unceitainty is 0.003ºC with a coveiage factoi k · 2.
To achieve a ieliable ice point, it is extiemely impoitant that eveiything that comes into contact with
the distilled watei and ice not impait any impuiities. Theiefoie, the hands of the opeiatoi must be
TABLE 12.5 Selected Values of Compiessibility
Factoi, Z
t, P,
Relative Humidity, U
ºC Pa 25% 50% 75%
20 101,325 0.99962 0.99960 0.99958
90,000 0.99966 0.99964 0.99962
80,000 0.99969 0.99967 0.99965
25 101,325 0.99966 0.99963 0.99961
90,000 0.99969 0.99967 0.99964
80,000 0.99972 0.99970 0.99967
theimometei coiiection
ief
t t,
© 2002 by CRC Press LLC
coveied with clean piotective gloves, and the ice cube tiays, the ice ciushei, ßask, and any othei objects
that come in contact with the ice oi watei must be thoioughly cleaned. An easy cleaning method is to
sciub the equipment with a glasswaie deteigent and watei mixtuie and then iinse with tap watei followed
by an ethanol (oi methanol) iinse. Then, aftei aii-diying, a distilled watei iinse and then steaming by
holding the inveited ßask ovei a boiling beakei of distilled watei follow. The ice ciushei and othei objects
can be cleaned by pouiing boiling distilled watei ovei them oi thiough them.
If the opeiatoi dedicates the iequiied equipment to this pioceduie and keeps it in clean stoiage when
not in use, futuie use iequiies less piepaiation. It is also wise to keep the ice cube tiays coveied with
plastic bags duiing theii stoiage in the fieezei duiing the ice-making piocess.
12.8.2.3 Trip!e Puint ul Water
The tiiple point of watei is achieved by fieezing a tiiple point of watei cell.
38
These cells aie available
commeicially and aie easy to use. The ieadei should iefei to Figuies 12.2 and 12.3. The cell should be
piechilled in a containei flled with a mixtuie of oidinaiy tap watei and ice as should the theimometei
to be tested. An ice mantle is foimed aiound the theimometei well by keeping the theimometei well
packed with chips of solid CO
2
(diy ice). Once an adequate mantle is foimed, the ice adjacent to the
theimometei well is melted (innei melt) by flling the well with chilled watei fiom the soaking bath and
then intioducing a ioom tempeiatuie, loose-ftting solid glass iod into the well. Aftei 15 to 20 s the cell
is held veitically and given a biisk quaitei-tuin iotation. If an innei melt has been obtained, the ice
mantle will iotate fieely about the theimometei well; if not, the waiming piocess should be iepeated
until the mantle can iotate. Once the iotation is achieved, the cell is accuiate to about 0.0002ºC, moie
than adequate foi the puipose heie.
The cell can be maintained in this state foi many days when stoied in a theimally insulated mixtuie
of packed ice and watei. The cell usei should always be suie that the ice mantle is fiee of the theimometei
well befoie each use as it can iefieeze, and also that the mantle diametei is not allowed to inciease to
wheie the cell is destioyed. The opeiatoi can pievent the destiuctive foimation of an ice biidge inside
the cell (neai the uppei end) by holding the cell with a baie hand duiing the mantle-foimation pioceduie.
FIGURE 12.1 Melting point of ice setup.
© 2002 by CRC Press LLC
12.8.2.4 Steam Puint
The watei vapoi above the suiface of boiling puie watei, when not confned, condenses at a well-
chaiacteiized tempeiatuie neai 100ºC. This phenomenon is also iefeiied to as the boiling point of
watei. The ielationship between the steam condensation tempeiatuie and atmospheiic piessuie
39
is
given in Table 12.6. The aii piessuie in mmHg must be measuied immediately adjacent to the fiee suiface
of the watei. The iesulting tempeiatuie is expiessed on IPTS 68 and must be conveited to ITS 90.
39
The
conveision foimula is
FIGURE 12.2 Tiiple point of watei cell.
FIGURE 12.3 Detail of tiiple point of watei cell.
© 2002 by CRC Press LLC
The diffeience between the two tempeiatuie scales is neaily constant space foi sea level and foi an aii
piessuie of 600 mmHg. The steam point tempeiatuie has an unceitainty (at k · 1) of 0.0015ºC when
the piessuie unceitainty is 0.1 mmHg.
The metiologist can ieadily access the steam condensation tempeiatuie to check theimometeis
using the appaiatus shown in Figuie 12.4. The appaiatus is constiucted fiom a 3-cm-diametei glass
tube about 30 cm long and heated with the smallest possible ßame that will sustain boiling and
piovide a steady theimometei indication. The watei must be distilled and kept puie. It is impoitant
that a clean boiling bead of glass oi Teßon' be in the watei to pievent supeiheating. A fbei glass
insulation sleeve suiiounds the glass tube about 2 cm above the watei suiface. The sensoi tip is
placed in the centei of the tube about 2 cm above the boiling watei and will be in theimal equilibiium
within a few minutes aftei the watei condensation diop foims on the theimometei tip. The thei-
mometei coiiection is deteimined as desciibed above foi the ice point.
It is extiemely impoitant that the system be open to atmospheiic piessuie and that the theimometei
nevei touch the watei oi the containei duiing the calibiation piocess. The aii piessuie is deteimined
TABLE 12.6 Condensation Tempeiatuie of Steam
Piessuie in mmHg (standaid)
P 0 1 2 3 4 5 6 7 8 9
Tempeiatuie in degiees on the Inteinational Piactical
Tempeiatuie Scale of 1968
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 .8648 .9092
610 93.9534 .9976 ¯.0419 ¯.0860 .1300 .1740 .2179 .2617 .3055 .3493
620 94.3940 .4366 .4802 .5237 .5671 .6105 .6538 .6972 .7404 .7835
630 94.8266 .8698 .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 ¯.0122 ¯.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
680 96.9128 .9532 .9936 ¯.0339 ¯.0741 ¯.1143 ¯.1546 ¯.1947 ¯.2347 ¯.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
730 98.8753 .9134 .9515 .9896 ¯.0275 ¯.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 .5848 .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
¯ Indicates change in integei.
T T
90 68
0 026 . . C
© 2002 by CRC Press LLC
adjacent to the appaiatus and at the elevation of the watei suiface. A ßame shield pievents oveiheating
of the walls of the glass tube neai the theimometei, and iadiation heating. Howevei, the ßame shield has
a small hole to expose the glass to the ßame.
Like the ice point, cleanliness of the appaiatus is impoitant and the same cleaning techniques aie
applied heie. The theimometei must also be clean.
12.8.2.5 Re!ative Humidity
Foi giavimetiic measuiement, the ielative humidity in the laboiatoiy should be maintained between
30 and 60%. Low humidity (20 to 25%) will suppoit the foimation of electiic chaige on the balance,
opeiatoi`s clothing, and nonconducting objects being weighed such as glasswaie. The iesulting foice
imposed on the balance mechanism iesults in eiioneous balance obseivations. With high humidities
theie is the possibility of watei condensation on the balance mechanism and the object being
weighed. The iesult of this is a degiadation of balance piecision oi an eiioi in the mass of the object
being weighed. These limits aie somewhat aibitiaiy and can be extended when dictated by ciicum-
stances; howevei, the opeiatoi should be awaie that the above eiiois might be encounteied.
Although many laboiatoiies contiol ielative humidity, otheis must accept what theii enviionment
offeis. Rapid ßuctuations should be avoided as should the above-mentioned extiemes. Fluctuations
less than 5% RH pei houi is a ieasonable limit foi high-piecision weighing.
The most widely used instiuments foi measuiing ielative humidity aie tiansduceis sensitive to
the amount of watei vapoi in the aii by the use of thin flms and iequiie exposuie to known ielative
humidities foi calibiation. Dew-point meteis, accuiate to 1% RH oi bettei, aie less common, moie
expensive, and need only have theii miiiois cleaned and theii inteinal theimometeis calibiated.
FIGURE 12.4 Steam condensation tempeiatuie appaiatus.
© 2002 by CRC Press LLC
Wet- and diy-bulb theimometeis iequiie theimometei calibiation, a souice of distilled watei, and
clean wicks. The wet- and diy-bulb theimometei method is limited to an accuiacy of appioximately 6%
RH foi the best instiuments and is no longei in common use foi mass calibiation.
The following calibiation method utilizing satuiated salt solutions,
40
that is, distilled watei and puie
chemical compounds, is an excellent means to calibiate the tiansduceis that iequiie exposuie to known
ielative humidity. One oi two of the salts given in Table 12.7 aie each mixed sepaiately in chemically
nonieactive containeis (glass oi polyethylene) with distilled watei. Some of the chemical compound (salt)
must iemain visible aftei mixing and given time to cieate a satuiated solution. The containei is coveied
with a lid that allows inseition of the humidity sensoi and at the same time be heimetically sealed; see
Figuie 12.5. A iecommended 24 h should be given to establish equilibiium.
A salt solution is chosen fiom Table 12.7 with a ielative humidity neai the woiking ielative humidity;
otheiwise two solutions aie selected that closely suiiound the woiking iange. If iequiied, the entiie iange
of the sensoi can be calibiated in this mannei. A sensoi coiiection is obtained in the same mannei
desciibed foi the theimometei.
This method depends on puie chemical compounds, Ameiican Chemical Society ACS giade, foi exam-
ple, puie distilled watei, and a veiy clean containei and sensoi. Techniques similai to those desciibed above
foi the ice point can be used to clean the containei. Caie must be taken not to immeise the sensoi in the
liquid solution as it may be damaged. The unceitainty of this method is 1% RH (at k · 2). We note that
lithium chloiide undeigoes an exotheimic ieaction when mixed with watei and caution is advised.
12.8.2.6 Pressure
Aii piessuie (baiometiic piessuie) is easily measuied by modein sensois, but like Foitin and aneioid
baiometeis, the sensois iequiie calibiation. Unlike meicuiy baiometeis and deadweight piessuie gauges,
they do not defne piessuie.
Most metiologists aie unable to geneiate the known piessuies necessaiy to peifoim this calibiation
in-house and must tianspoit the instiument elsewheie foi this expensive calibiation. Unfoitunately,
upon ietuin of the instiument, one can nevei be ceitain the calibiation iemains valid. It has been
demonstiated that piessuie sensois can be calibiated giavimetiically by weighing two neaily identical
known masses of widely diffeiing but known volumes.
35
The metiologist can choose to calibiate the baiometiic sensoi giavimetiically, simply test the validity
of a calibiation peifoimed elsewheie, oi peifoim a diiect deteimination of aii density. In any case, the
method is the same. Adequately calibiated weights of known mass and volume aie available commeicially.
The measuiements aie easily peifoimed using a 1-kg weight fabiicated fiom aluminum and anothei
FIGURE 12.5 Appaiatus foi calibiation of humidity sensois.
© 2002 by CRC Press LLC
fiom stainless steel. The mass diffeience between these two weights when they aie compaied on a balance
and accounting foi aii buoyancy is
(12.50)
Rewiiting Eq. (12.50) to expiess
g
, explicitly:
(12.51)
wheie
g
(g/cm
3
) is the giavimetiically deteimined aii density, AL is the mass of a low-density weight
such as an aluminum weight, SS is the mass of a stainless steel weight nominally equal to that of the
aluminum weight, and V
A
and V
S
aie theii iespective displacement volumes (cm
3
). C
1
and C
2
aie the
iespective balance indications when SS and AL aie weighed.
Equating the iight-hand membeis of Eq. (12.51) to the iight-hand membeis of Eq. (12.47), one can
wiite an expiession foi the aii piessuie at the time of weighing and compaie it to the piessuie indicated
by the piessuie sensoi duiing the weighing
35
:
(12.52)
TABLE 12.7 Relative Humidities Established by Satuiated
Salt Solutions
Lithium chloiide, LiCl ·H
2
O
Temp., ºC 0.23 9.56 19.22 29.64 39.64 46.76
RH, % 14.7 13.4 12.4 11.8 11.8 11.4
Magnesium chloiide, MgCl
2
· 6H
2
O
Temp., ºC 0.42 9.82 19.53 30.18 39.96 48.09
RH, % 35.2 34.1 33.4 33.2 32.7 31.4
Sodium dichiomate, Na
2
Ci
2
O
7
· 2H
2
O
Temp., ºC 0.60 10.14 19.82 30.01 37.36 47.31
RH, % 60.4 57.8 55.3 52.4 50.4 48.0
Magnesium nitiate, Mg(NO
3
)
2
· 6H
2
O
Temp., ºC 0.39 9.85 19.57 30.47 40.15 48.10
RH, % 60.7 57.5 55.8 51.6 49.7 46.2
Sodium chloiide, NaCl
Temp., ºC 0.92 10.23 20.25 30.25 39.18 48.30
RH, % 75.0 75.3 75.5 75.6 74.6 74.9
Ammonium sulfate, (NH
4
)
2
SO
4
Temp., ºC 0.39 10.05 20.04 30.86 39.97 47.96
RH, % 83.7 81.8 80.6 80.0 80.1 79.2
Potassium nitiate, KNO
3
Temp., ºC 0.62 10.17 20.01 30.70 40.35 48.12
RH, % 97.0 95.8 93.1 90.6 88.0 85.6
Potassium sulfate, K
2
SO
4
Temp., ºC 0.54 10.08 19.81 30.44 39.94 48.06
RH, % 99.0 98.0 97.1 96.8 96.1 96.0
SS AL C C V V
g A S 1 2
.
g S A
C C V V SS AL
1 2
,
P TZ Ue
g s
0 0034836 0 0037960 . . .
© 2002 by CRC Press LLC
12.8.2.7 Lncertainty
The paitial deiivatives foi Eq. (12.51) aie
(12.53)
(12.54)
(12.55)
(12.56)
(12.57)
(12.58)
The paitial deiivatives
41
foi the aii density, Eq. (12.47), that utilizes the sensoi obseivations of tem-
peiatuie, piessuie, and ielative humidity aie
(12.59)
(12.60)
(12.61)
Sample data and values foi ancillaiy teims such as e
s
and Z with unceitainty estimates can be found
in Table 12.8. These data weie used in the calculated aii densities, the giavimetiically deiived aii
piessuie, and eiioi piopagation that follows in Tables 12.9 and 12.10.
A guideline foi estimating unceitainties is given in Ref. 42. The guideline is abbieviated heie:
(12.62)
Foi Eq. (12.62), RSS (unceitainty at k · 1) is the squaie ioot of the sum of the pioducts of the squaie
of the paitial deiivatives and the squaie of the SDs foi each vaiiable (vai) of the functional ielationship,
f (x).
12.8.2.8 Mass Ca!ibratiun
Having obtained an aii density and the associated unceitainty eithei giavimetiically oi by measuiing the
paiameteis T, P, and U, it iemains to piopagate the unceitainty thiough a typical weight calibiation.
The calculated aii density and unceitainty found in Table 12.9 have been used to piopagate uncei-
tainties foi a 1-kg weight of class E
1
when used as a standaid to calibiate a 1-kg weight of class E
2
. The
data piovided by Jones and Schoonovei
43
(see Chaptei 16) with the above aii density exception have been
g A S
V V AL 1
g A S
V V SS 1
g A S
V V O
1
1
g A S
C V V
2
1
g A A S
V C C V V AL SS
1 2
2
g S A S
V C C V V AL SS
1 2
2
.
a a
T T
a a
P P
a s a
U e P 0 0037960 . .
RSS vai SD f x
2 2
1 2
© 2002 by CRC Press LLC
TABLE 12.8 Sample Data Set
Vaiiable Value SD
Aii Density Calculated fiom P, T, and U
T 296.28 0.1 K
(23.13ºC) (0.1ºC)
P 98,917 Pa 3.3 Pa
(741.94 mmHg) (0.02 mmHg)
U 26.6 2
a
calculated fiom Eq. (12.47) 1.1601 kg/m
3
Intiinsic unceitainty 0.0001 kg/m
3
Combined unceitainty 0.0005 kg/m
3
Giavimetiic Aii Density
SS 1000.01444 g 0.0001 g
AL 1000.02638 g 0.0001 g
C
1
-C
2
-0.00424 g 0.00005 g
V
A
333.736 cm
3
0.034 cm
3
V
S
126.418 cm
3
0.013 cm
3
a
calculated fiom Eq. (12.51) 1.1600 kg/m
3
Unceitainty 0.0006 kg/m
3
Ancillaiy Data and Equivalents
Z 0.99964 1.7 10
-5
e
s
2831 Pa 0.03 %
Equivalents: 101,325 Pa · 760 mmHg; 1.2
kg/m
3
· 1.2 mg/cm
3
· 0.0012 g/cm
3
.
Aii Piessuie by Weight, Eq. (12.52) · 98,908
Pa.
TABLE 12.9 Aii Density Eiioi Piopagation
Vaiiable Value SD / vai.
Aii Density - Giavimetiic
AL 1000.026 g 0.0001 4.7 10
-3
SS 1000.014 g 0.0001 -4.7 10
-3
V
A
338 cm
3
0.0338 -3.6 10
-7
V
S
126 cm
3
0.0126 3.6 10
-7
C
1
0.000 g 5 10
-5
4.7 10
-3
C
2
-0.004 g 5 10
-5
-4.7 10
-3
a
· 0.00116 g/cm
3
RSS · 0.00000074 g/cm
3
Aii Density - Calculated fiom T, P, U
T 296.28 K 0.1 3.9 10
-6
P 98917 Pa 3.3 1.2 10
-8
U 26.6 2 1.3 10
-7
a
· 0.00116 g/cm
3
RSS · 0.00000047 g/cm
3
With intiinsic · 0.00000048 g/cm
3
© 2002 by CRC Press LLC
used. The E
1
weight has mass of 1000.0002 g and a density of 8.067 g/cm
3
. The E
2
weight has a density
of 7.810 g/cm
3
and its mass is deteimined to be 999.9994 g. See Table 12.11 foi the unceitainty piopa-
gation. As iequiied by R111, the unceitainty of a class E
2
1-kg weight must not exceed 0.5 mg. The
unceitainty foi the example, 2 97 10
-6
g, meets this iequiiement.
12.8.2.9 Summary
A piactical infoimation souice foi the deteimination of aii density and its unceitainty (at k · 1) along
with calibiation pioceduies foi the associated instiumentation have been piovided.
12.9 Test ul Air Density Equatiun at Dillering A!titudes
12.9.1 Intruductiun
A seiies of measuiements of the mass of aluminum and tantalum aitifacts compaied against mass
standaids of stainless steel was undeitaken at the National Buieau of Standaids (NBS) and the iesults
weie published in 1975.
44
Inconsistencies weie iepoited that seemed to coiielate with baiometiic piessuie.
The aii density equation of Jones
2
was subsequently developed. Measuiements foi inteicompaiison
between two objects of diffeient density weie made to deteimine the mass diffeience between the objects
when weighed in aii.
45
The inteicompaiison measuiements could be used to test the Jones aii density equation.
2
The agiee-
ment between the measuiements and the equation was within the expeiimental unceitainty, 600 ppm
in aii density.
An expeiiment was then undeitaken to attempt to ieconcile the measuiements, which weie consistent
with the equation, with those in Ref. 44.
12.9.2 Experimenta! Detai!s
A seiies of measuiements with a selection of kilogiam aitifacts was made at NBS, Gaitheisbuig, MD.
Similai measuiements weie made with the same aitifacts at Sandia Laboiatoiies at Albuqueique, NM.
TABLE 12.10 Eiioi Piopagation foi
Giavimetiically-Deteimined Aii Piessuie
Vaiiable Value SD P/ vai.
a
1.16 kg/m
3
0.0005 85,019
T 298 K 0.01 320
Z 0.99964 1.7 10
-5
94,826
U 26 2 10.74
e
s
2831 Pa 85 0.10
RSS · 49 Pa. Piessuie Calculated by Eq. (12.52) ·
98908 Pa.
TABLE 12.11 Eiioi Piopagation
foi a Class E
2
1-kg Weight
Vaiiable E
2
/ Vai. SD
Mass of E
1
1 83 10
-6
Balance obseivation -1 50 10
-6
Density of E
1
0.014 8 10
-6
Density of E
2
-0.0135 5 10
-4
Aii density -2.16 5 10
-7
RSS · 97 10
-6
g; mass of E
2
· 999.9994 g.
© 2002 by CRC Press LLC
The NBS, Gaitheisbuig, laboiatoiies aie neai sea level, the Sandia Laboiatoiies aie at about 1600 m
(about 5249 ft) above sea level. The aitifacts included the aluminum and tantalum kilogiams used in
Ref. 44, as well as seveial othei weights designed to elucidate suiface effects.
At both NBS and Sandia Laboiatoiies, mass measuiements weie made on commeiically available
single-pan kilogiam balances of conventional design, each with a piecision of 25 to 50 g.
Ten diffeient 1-kg weights weie used in the expeiiment. In Table 12.12, the chaiacteiistics and desig-
nations of the weights aie listed.
B1 and D2, the most conventional weights, weie used as standaids. They weie of single-piece stainless
steel constiuction, of neaily minimum suiface aiea, and with knobs foi ease of handling.
H1 and H2, of stainless steel, weie designed to have density neai that of aluminum. They weie hollow,
iight ciiculai cylindeis of minimum suiface aiea (diametei equal to height), with an inteinal centeipost
foi iigidity, and flled with helium at about 1 atmospheie piessuie.
R1 and R2, constiucted as companions to the hollow weights, weie thick-walled stainless steel tubes
with suiface aieas nominally equal to those of H1 and H2.
S1 and S2, of solid stainless steel with suiface aieas ioughly equal to twice those of the R weights, weie
each in the foim of two nested stainless steel tubes on a ciiculai stainless steel base. A centeipost welded
to the base allowed easy manipulation of the S weights.
A was a single-piece weight of aluminum, constiucted of bai stock in the foim of a iight ciiculai
cylindei of minimum suiface aiea.
T was a single-piece weight of tantalum, of neaily minimum suiface aiea, with a knob foi ease of
handling.
Weights A and T weie the same ones used in the expeiiment iepoited in Ref. 44.
All of the weights except A and T weie steam-cleaned piioi to these expeiiments. The weights with
stainless steel suiface weie also vapoi-degieased with 1,1,1-tiichloioethane. No fuithei cleaning was
attempted; all weights weie dusted with a soft, lint-fiee biush piioi to each use.
12.9.3 Ca!cu!atiun ul Air Density lur Buuyancy Currectiun
The values of aii density foi the measuiements weie calculated using the aii density equation of Jones.
2
12.9.4 Measurements ul Parameters in the Air Density Equatiun
12.9.4.1 Temperature
Tempeiatuie was the most elusive of all the measuiements made in the couise of the expeiiment. Type E
theimocouples weie distiibuted within the weighing chambei. A meicuiy-in-glass total immeision thei-
mometei was used as a backup to the theimocouples.
TABLE 12.12 Chaiacteiistics and Designations
of the Ten Weights
Aitifact Nominal Volume (cm
3
) Nominal Suiface
Designation Mass (kg) at 20ºC Aiea (cm
2
)
B1 1 127.385 145
D2 1 127.625 145
H1 1 337.381 270
H2 1 337.666 270
R1 1 126.395 270
R2 1 126.392 270
S1 1 126.549 660
S2 1 126.545 660
A 1 359.488 280
T 1 60.027 85
© 2002 by CRC Press LLC
12.9.4.2 Pressure
Aneioid baiometeis weie used to iead baiometiic piessuie at both NBS and Sandia. One of the aneioid
baiometeis coveied aii piessuies neai sea level at NBS, and the othei coveied atmospheiic piessuies at
an altitude of about 1600 m at Sandia.
The aneioid baiometei used at NBS was calibiated at NBS, and it was also checked at ambient piessuie
twice daily against a cistein-type meicuiy manometei.
The aneioid baiometei used at Sandia was calibiated fist by a piivate calibiation laboiatoiy and then
in situ at Sandia. In addition to the meicuiy manometei, a piston gauge and a sensitive quaitz piessuie
tiansducei weie taken to Sandia.
The quaitz piessuie tiansducei was used to calibiate the aneioid baiometei and the tiansducei was
calibiated twice daily against the piston gauge. The calibiated quaitz tiansducei agieed with the meicuiy
manometei to within 30 ppm.
The value of the local acceleiation due to giavity, g, which was needed to make accuiate piessuie
measuiements using the piston gauge and the meicuiy manometei, was supplied by peisonnel of the
Sandia Laboiatoiies.
12.9.4.3 Re!ative Humidity
A Dunmoie-type humidity-sensing element was mounted in one dooi of the balance case to measuie
ielative humidity. The humidity elements used weie calibiated at NBS. Peiiodic checks of calibiation
weie made by immeising the element in aii above a satuiated salt solution, which pioduced an atmospheie
of 34% ielative humidity (RH), close to ambient humidity at both NBS and Sandia.
12.9.4.4 Carbun Diuxide Cuntent
The caibon dioxide content of the aii in the balance case was tested twice daily foi both NBS and Sandia.
Aii samples weie diawn into evacuated spheies, sealed with vacuum-type stopcocks, and then analyzed
at NBS.
12.9.5 Weighings
All of the weighings weie made using a so-called foui-ones pattein in which foui objects of nominally
equal mass weie inteicompaied in each of six possible combinations (see Chaptei 8).
A least-squaies ft of the weighing data iesulted in the assignment of mass to each of thiee of the
objects when the mass of the fouith object was known. B1 and D2 weie used in eveiy measuiement.
Since the mass of B1 and D2 was known, theii sum was used in the least-squaies solution.
12.9.6 Cunc!usiuns
Ovei a peiiod of seveial months, fve gioups of measuiements of the mass of the aluminum and the
tantalum kilogiam against stainless steel mass standaids weie caiiied out. Foui gioups of measuiements
weie made at NBS, Gaitheisbuig and one gioup was made at the Sandia Laboiatoiies.
The agieement between the measuiements and the equation of Jones
2
was within the expeiimental
unceitainty, 600 ppm in aii density.
The gioups of data exhibited signifcant diffeiences among them of the type iepoited in Ref. 44.
Howevei, the magnitude of these disciepancies was a factoi of 5 less than had been pieviously obseived
by the authoi of Ref. 44. The authois of the expeiimental woik iepoited in Ref. 45 weie unable to
iepioduce oi satisfactoiily explain the eailiei iesults.
Suiface effects weie consideied likely to play a iole in the disciepancies iepoited in this latei woik.
It was hypothesized that this obseived behavioi was due to weights not being in suffciently good
theimal equilibiium with the balance.
© 2002 by CRC Press LLC
Relerences
1. Davis, R. S., New assignment of mass values and unceitainties to NIST woiking standaids, }. Res.
Natl. Bur. Stand. (U.S.), 95, 79, 1990.
2. Jones, F. E., An aii density equation and the tiansfei of the mass unit, }. Res. Natl. Bur. Stand.
(U.S.), 83, 419, 1978.
3. Giacomo, P., Equation foi the deteimination of the density of moist aii (1981), Metrclcgia, 18, 33,
1982.
4. Balhoin, R., PTB Mit., 93, 303, 1983.
5. Davis, R. S., Recalibiation of the U.S. National Piototype Kilogiam, }. Res. Natl. Bur. Stand. (U.S.),
90, 263, 1985.
6. Davis, R. S., Equation foi the deteimination of the density of moist aii (1981/91), Metrclcgia, 29,
67, 1992.
7. List, R. J., Smithscnian Metecrclcgical Tables, 6th iev. ed., Smithsonian Institution, Washington,
D.C., 1951, 347.
8. Cohen, E. R. and Tayloi, B. N., }. Phys. Chem. Ref. Data, 2, 663, 1973.
9. Quinn, T. J., Colclough, A. R., and Chandlei, T. R. D., Philcs. Trans. R. Scc. Icndcn, A283, 367, 1976.
10. Gammon, B. E., The velocity of sound with state piopeities in helium at -175 to 159ºC with
piessuie to 150 atm, }. Chem. Phys., 64, 2556, 1976.
11. Gammon, B. E., Piivate communication.
12. Rowlinson, J. S. and Tildesley, D. J., Prcc. R. Scc. Icndcn, A358, 281, 1977.
13. Moldovei, M. R. et al., Measuiement of the univeisal gas constant R using a spheiical acoustic
iesonatoi, Phys. Rev. Iett., 60, 249, 1988.
14. Moldovei, M. R. et al., Measuiement of the univeisal gas constant R using a spheiical acoustic
iesonatoi, }. Res. Natl. Bur. Stand. (U.S.), 93, 85, 1988.
15. Inteinational Union of Puie and Applied Chemistiy, Inoiganic Chemistiy Division, Commission
on Atomic Weights, Pure Appl. Chem. 47, 75, 1976.
16. Machta, L. and Hughes, E., Atmospheiic oxygen in 1967 to 1970, Science, 168, 1582, 1970.
17. Gluekauf, E., in Ccmpendium cf Metecrclcgy, Malone, T. F., Ed., The Ameiican Meteoiological
Society, Boston, 1951, 4.
18. Hughes, E., Papei piesented at the Ameiican Chemical Society Fouith Mid-Atlantic Meeting,
Washington, D.C., Febiuaiy 1969.
19. U.S. Standaid Atmospheie, 1976, U.S. Goveinment Piinting Offce, Washington, D.C., 1976, 3, 33.
20. Hyland, R. W., A coiielation foi the second inteiaction viiial coeffcients and the enhancement
factois foi moist aii, }. Res. Natl. Bur. Stand. (U.S.), 79A, 551, 1975.
21. Wexlei, A., Metecrcl. Abstr., 11, 262, 1970.
22. List, R. J., Smithscnian Metecrclcgical Tables, 6th iev. ed., Smithsonian Institution, Washington,
D.C., 1951, 333.
23. Haiiison, L. P., in Humidity and Mcisture, Vol. III, Wexlei, A. and Wildhack, W. A., Eds., Reinhold,
New Yoik, 1964, 51.
24. Hyland, R. W., }. Res. Natl. Bur. Stand. (U.S.), 79A, 551, 1975.
25. Wexlei, A. and Gieenspan, L., Vapoi piessuie equation foi watei vapoi in the iange 0 to 100ºC,
}. Res. Natl. Bur. Stand. (U.S.), 75A, 213, 1971.
26. Besley, L. and Bottomley, G. A., Vapoui piessuie of noimal and heavy watei fiom 273.15 to 298.15,
}. Chem. Thermcdyn., 5, 397, 1973.
27. Ku, H. S., Statistical concepts in metiology, in Handbcck cf Industrial Metrclcgy, Ameiican Society
of Tool and Manufactuiing Engineeis, Pientice-Hall, Englewood Cliffs, NJ, 1967, chap. 3.
28. Koch, W. F., Davis, R. S., and Bowei, V. E., Diiect deteimination of aii density in a balance thiough
aitifacts chaiacteiized in an evacuated weighing chambei, }. Res. Natl. Bur. Stand. (U.S.), 83, 407,
1978.
© 2002 by CRC Press LLC
29. Glasei, M., Schwaitz, R., and Mecke, M., Expeiimental deteimination of aii density using a 1 kg
mass compaiatoi in vacuum, Metrclcgia, 28, 45, 1991.
30. Schoonovei, R. M., Measuiement Science Confeience, Anaheim, CA, Januaiy 2000.
31. Oiganisation Inteinationale de Metiologie Legale (OIML) Inteinational Recommendation R111,
Weights of Classes E
1,
E
2
, F
1
, F
2
, M
1
, M
2
, M
3
, OIML, 1994.
32. Oiganisation Inteinationale de Metiologie Legale (OIML) Inteinational Recommendation No. 33,
Conventional Value of the Result of Weighing in Aii, OIML, 1979.
33. List, R. J., Smithscnian Metecrclcgical Tables, Smithsonian Institution, Washington, D.C., 1958,
266-267.
34. Davis, R. S., Equation foi the deteimination of the density of moist aii (1981/91), Metrclcgia, 29,
67, 1992.
35. Schoonovei, R. M., A simple giavimetiic method to deteimine baiometei coiiections, }. Res. Natl.
Bur. Stand., 85, 341, 1980.
36. Schoonovei, R. M. and Tayloi, J. E., Some iecent developments at NBS in mass, Measuiement,
IEEE Trans. Instrum. Meas., 35, 418, 1986.
37. Wise, J., Liquid-In-Glass Theimometiy, NBS Monogiaph 150, U.S. Depaitment of Commeice,
1986, 8-9.
38. Cioss, J. L., Instiuctions, Tiiple Point of Watei Cells, Jaiiett Instiument Co., Wheaton, MD, Jan.
1964, 1-4.
39. Mangum, B. W. and Fuiukawa, G. T., Guidelines of Inteinational Tempeiatuie Scale of 1990 (ITS),
NIST Technical Note 1265, U.S. Depaitment of Commeice, 1990, 5.
40. Wexlei, A., and Hasegawa, S., Relative humidity-tempeiatuie ielationships of some satuiated salt
solutions in the tempeiatuie iange of 0º to 50ºC, }. Res. Natl. Bur. Stand. (U.S.), 53, 19, 1954.
41. Schoonovei, R. M. and Jones, F. E., Paiameteis that can cause eiiois in mass deteiminations, Cal.
Iab., 5, 26, 1998.
42. Tayloi, B. N. and Kuyatt, C. E., Guidelines foi Evaluating and Expiessing the Unceitainty of NIST
Measuiement Results, NIST Technical Note 129, 1994.
43. Jones, F. E. and Schoonovei, R. M., A compaiison of eiioi piopagations foi mass and conventional
mass, Measuiement Science Confeience 2000, Anaheim, CA, Januaiy 2000.
44. Pontius, P. E., Science, 190, 379, 1975.
45. Schoonovei, R. M., Davis, R. S., Diivei, R. G., and Bowei, V. E., A piactical test of the aii density
equation in standaids laboiatoiies at diffeiing altitude, }. Res. Natl. Bur. Stand. (U.S.), 85, 27, 1980.
© 2002 by CRC Press LLC