J. Chem.

Thermodynamics 1997, 29, 1157–1174

Measurement and correlation of the ( p, r, T )

relation of nitrogen
II. Saturated-liquid and saturated-vapour densities
and vapour pressures along the entire coexistence
curve

P. Nowak, R. Kleinrahm, and W. Wagner a

Lehrstuhl für Thermodynamik, Ruhr-Universität Bochum, D-44780 Bochum,
F.R.G.

Comprehensive and accurate measurements of the saturated-liquid and saturated-vapour

densities together with the vapour pressure of pure nitrogen were carried out from the
temperature T = 64 K (triple-point temperature Tt = 63.151 K) to 0.022 K below the critical
temperature. The typical values of the total uncertainties of the measurements are: Q=2·10−4·ps =
for the vapour pressures, Q=1.2·10−4·r'= for the saturated-liquid densities, and Q=3·10−4·r0=
for the saturated-vapour densities. The critical constants (Tc = 126.192 K, rc = 313.3 kg·m−3,
pc = 3.3958 MPa) and the isothermal compressibilities in the critical region close to the phase
boundary have also been determined from these measurements. Comparisons with experimental
results of previous workers are presented. Based on the new values of this work, new correlation
equations for the vapour pressure, the saturated-liquid density, and the saturated-vapour
density have been established. 7 1997 Academic Press Limited

KEYWORDS: nitrogen; vapour pressures; saturation densities; critical constants;

two-sinker-densimeter

1. Introduction
This paper is the second part of a project which deals comprehensively with the
( p, r, T ) behaviour of pure nitrogen. The first part,(1) where the reasons for these
new comprehensive measurements are given, describes the experimental results in
the homogeneous gas and liquid region, including the largest part of the critical
region in the temperature range from 66 K to 340 K at pressures up to 12 MPa.
This second paper presents the measurements of the saturated-liquid and the
saturated-vapour densities and the vapour pressure of nitrogen along the entire
coexistence curve from T = 64 K (triple-point temperature Tt = 63.151 K) up to
slightly below the critical temperature Tc = 126.192 K. The aim of this work was to
provide a comprehensive and reliable set of measurements, which accurately
describes the coexistence curve and allows the establishment of new correlation
equations for the thermal saturation properties.

a
To whom correspondence should be addressed. E-mail: Wagner.thermo.ruhr-uni-bochum.de.

0021–9614/97/101157 + 18 $25.00/0/ct970231 7 1997 Academic Press Limited

1158 P. Nowak, R. Kleinrahm, and W. Wagner

The experimental results described in reference 1 and in this work represent an

essential part of the ( p, r, T ) surface of nitrogen. Based on these new results
and on reliable results of other experimentalists, a new equation of state for
nitrogen is being developed by Span et al.(2) This equation will describe the entire
thermodynamic surface of nitrogen within the experimental uncertainty of the most
reliable results and it will be of reference quality.

2. Experimental
The essential details of the apparatus for accurate ( p, r, T ) measurements in the
entire fluid region of pure substances have been described in a previous paper.(3)
The operational range of this apparatus covers a density range from 1 kg·m−3
to 2000 kg·m−3 at temperatures from 60 K to 340 K and pressures up to 12 MPa.
The principle of the method for density measurements (‘‘Two-Sinker-Method’’)
was briefly described in the first part of this work.(1) Nevertheless, in order to
understand the measuring principle for the application considered in this paper, the
measuring procedure of saturated-vapour and saturated-liquid densities is briefly
explained in this section.
As well as the saturation densities of methane,(3,4) carbon dioxide,(5) argon,(6) and
ethylene,(7) the saturated-vapour densities of nitrogen along the largest part of
the coexistence curve were determined in the following way. In the lid of the
measuring cell (see the schematic drawing in reference 3), an especially designed
small reference cell was included which was filled with nitrogen in such a way that
the fluid was in the phase equilibrium (vapour + liquid). A differential-pressure
indicator (uncertainty Q10 Pa) was installed between the measuring cell and the
reference cell. With this arrangement, the values of the saturated-vapour density
r0 were determined by measuring three or more ( p, r, T ) values very close to
the phase boundary {pressure difference ( p − ps ) = −5 kPa to −0.4 kPa} on the
corresponding isotherms in the homogeneous gas region outside the critical region.
The measured densities were then extrapolated to the coexistence curve by means of
a linear equation fitted to the ( p, r, T ) values. In the critical region, however, a linear
extrapolation is no longer suitable for the curved isotherms. Therefore, to determine
the saturated-vapour densities in the critical region for T e 125 K, several densities
were measured on the corresponding isotherm at pressure differences p − ps between
the pressure p in the measuring cell and the vapour pressure ps ranging from
about −3.0 kPa to −0.15 kPa. The saturation densities were then determined by
extrapolating these isothermal ( p, r, T ) values to ( p − ps ) = 0 with:
= p − N1 = = N2 ·=r − N3 =d, (1)
which is mathematically identical with the corresponding power-law to describe the
critical isotherm around the critical density. For this extrapolation, the constants
N1 , N2 , N3 , and d were non-linearly adjusted to the experimental values on each
isotherm in the critical region.
Very close to the critical temperature (for nitrogen from 0.112 K below
Tc up to Tc ) a refined procedure was applied to determine saturated-vapour
 ( p, r, T ) measurements on nitrogen 1159

densities. This new procedure is illustrated in figure 1 and explained in the

following. For the isotherm considered, the saturated-vapour densities were
determined by measuring several ( p, r, T ) values in the homogeneous gas
phase close {30 hPa E ( ps − p) E 1.5 hPa} and, in addition, extremely close
{1.5 hPa E ( ps − p) E 0 hPa} to the (vapour + liquid) phase boundary. In order to
measure the densities extremely close to the phase boundary, marked by special
symbols in figure 1, it was necessary to fill the measuring cell in such a way that the
nitrogen was just in the two-phase state. Due to the disappearing surface tension in
this state the saturated-liquid phase gathered only on the bottom of the cell and
no liquid droplets on either of the two sinkers affected the density measurements.
Thus, ( p, r, T ) values could be measured in the gas phase extremely close to
the phase boundary (see figure 1). These values were sequentially measured in a
quasi-equilibrium state while the fluid in the measuring cell was slowly condensing
within a period of about (20 to 60) h. The saturated-vapour density was then
determined by extrapolating all ( p, r, T ) values of this isotherm by means of
equation (1) to ( p − ps ) = 0 with an extraordinarily small uncertainty. Knowledge
of the course of the quasi-equilibrium values and the very small pressure difference
Dphyd (see figure 1) is very important for the accuracy of the extrapolation.
As described in reference 3, the saturated-liquid densities outside the critical region
were measured when the cell was filled with liquid up to a certain level above the
sinkers. This level was checked with a specially designed liquid-level indicator. In the
critical region, however, the measurements were carried out in a way similar to

FIGURE 1. Schematic illustration of the new procedure to determine the saturated-vapour densities
very close (in this case at T = 126.170 K) to the critical temperature Tc = 126.192 K. A: filling of the
measuring cell; B: Phase equilibrium (vapour + liquid) in the measuring cell; q, ( p, r, T ) measurements
in the homogeneous gas region close to the phase boundary (vapour + liquid); w, ( p, r, T ) measurements
in the homogeneous gas region extremely close to the phase boundary (vapour + liquid); these values were
measured in a quasi-equilibrium state while the fluid in the measuring cell was slowly condensing; w, =
( p, r, T ) value measured after phase equilibrium (vapour + liquid) had been achieved in the measuring
cell; (, saturated-vapour density r0 determined by extrapolating all measured ( p, r, T ) values in the
gas-phase to ( p − ps ) = 0 by means of equation (1). Also, Dphyd corresponds to the hydrostatic pressure
of the gas column between the phase boundary (vapour + liquid) a few mm above the bottom of the
measuring cell and the reference height of the two sinkers (see figure 1 in reference 3); this distance only
amounts to a few cm and is very well known.
1160 P. Nowak, R. Kleinrahm, and W. Wagner

the measurements of the saturated-vapour densities. This means that several

liquid densities were measured at various small pressure differences ( p − ps ) on
an isotherm. These pressure differences, ranging from about 2.5 kPa to 0.2 kPa,
were achieved by setting different filling levels above the sinkers. Then, the
saturated-liquid density was determined by extrapolating these isothermal liquid
densities to ( p − ps ) = 0, again by non-linearly fitting equation (1) to the ( p, r, T )
values.
The vapour pressures were determined together with the saturated-liquid densities
at a filling level of the measuring cell of about 90 per cent. Furthermore, in the critical
region from T = 125 K to T = 126.17 K vapour pressures were also measured at a
filling level of the measuring cell of about 5 per cent after the isothermal density
measurements in the homogeneous gas region had been carried out for determining
the saturated-vapour densities.
The influence of impurities in the nitrogen used for the measurements was also
checked. For this purpose, the measuring cell was filled with nitrogen so that the
liquid phase took up about 90 per cent of the cell volume at different temperatures.
After equilibrium had been reached, the vapour pressure was measured. Afterwards,
the vapour phase of the measuring cell was exhausted and, when equilibrium had
once again been reached at the same temperature as before, the vapour pressure was
measured again. This procedure was repeated several times until the liquid level in
the measuring cell had decreased to about 20 per cent. If there had been significant
impurities in the nitrogen they would have caused a change of the vapour pressure
measured by this procedure, but considering the different hydrostatic pressures,
these test measurements did not show any significant changes. Consequently, the
influence of impurities on the measuring values is negligible.

3. Results
The nitrogen used for the measurements was supplied by Messer-Griesheim,
Germany. According to the gas analysis of the producer the purity was
x(N2 ) q 0.999999, where x denotes mole fraction {impurity: x(Ar) Q 1·10−6}. The
experimental results are listed in tables 1 and 2. The saturation densities and
the vapour pressures were measured from T = 64 K (triple-point temperature
Tt = 63.151 K) up to 0.022 K below the critical temperature Tc = 126.192 K in steps
of, generally, 3 K; near the triple-point temperature and near the critical temperature
the steps in temperature were reduced to 1 K and 0.02 K, respectively. In order to
check the reproducibility, some measurements were repeated at different times.
The results of the measurements in the critical region are listed in table 2 and
are partly shown in figure 2. The measurements were carried out very close to
the saturation line from T = 125.00 K to T = 126.17 K for the vapour, and
from T = 125.25 K to T = 126.15 K for the liquid. The values were measured on
isotherms at pressure differences ( p − ps ) ranging from about 23 kPa to 20.1 kPa
to determine the saturation densities by extrapolation to ( p − ps ) = 0 by means of
equation (1). In order to determine the isothermal compressibilities k'T and k0T
at saturation and also to check the consistency of the measurements in the
 ( p, r, T ) measurements on nitrogen 1161

TABLE 1. Experimental results for the vapour pressure ps , the saturated-liquid density r', and the
saturated-vapour density r0 of nitrogen together with their total experimental uncertainties Dps , Dr', and
Dr0, respectively, and T is the temperature (ITS-90)

T ps r' r0
102·Dps /ps 102·Dr'/r' 102·Dr0/r0
K MPa kg·m−3 kg·m−3

64.000 0.014605 863.723 0.7770 a 20.075 20.010 20.075

65.000 0.017407 859.605 0.9132 a 20.068 20.010 20.068
66.000 0.020625 855.454 1.0673 a 20.063 20.010 20.063
68.000 0.028484 847.062 1.4360 a 20.055 20.010 20.055
70.000 0.038548 838.544 1.8961 a 20.049 20.010 20.049
72.000 0.051216 829.900 2.4616 a 20.044 20.010 20.044
75.000 0.076043 816.681 3.5404 a 20.039 20.010 20.039
78.000 0.109252 803.154 4.9440 a 20.036 20.010 20.036
81.000 0.152506 789.275 6.7314 a 20.035 20.010 20.035
83.000 0.187790 779.811 8.1679 a 20.033 20.010 20.033
86.000 0.251732 765.248 10.7416 a 20.029 20.010 20.029
0.251736 765.247 10.7415
90.000 0.360450 745.057 15.0796 20.025 20.010 20.025
0.360459 745.056
93.000 0.462086 729.230 19.1311 20.023 20.010 20.024
0.462097 729.228
96.000 0.583155 712.708 23.9887 20.022 20.010 20.023
0.583156 712.707
98.000 0.675647 701.247 27.7413 20.021 20.010 20.022
101.000 0.833567 683.267 34.2626 20.020 20.010 20.021
103.000 0.952575 670.675 39.2934 20.019 20.010 20.021
105.000 1.08330 657.512 44.9583 20.018 20.010 20.021
108.000 1.30282 636.488 54.8573 20.017 20.010 20.020
110.000 1.46577 621.437 62.5782 20.016 20.011 20.020
1.46581 621.437 62.5792
112.000 1.64292 605.354 71.4062 20.016 20.011 20.021
114.000 1.83505 587.982 81.5875 20.015 20.011 20.021
115.000 1.93701 578.714 87.2959 20.015 20.011 20.021
116.000 2.04303 568.973 93.4900 20.015 20.011 20.021
117.000 2.15327 558.689 100.248 20.014 20.011 20.022
118.000 2.26783 547.757 107.671 20.014 20.011 20.022
119.000 2.38688 536.051 115.891 20.013 20.012 20.023
120.000 2.51059 523.387 125.088 20.014 20.012 20.024
121.000 2.63914 509.498 135.529 20.014 20.012 20.025
122.000 2.77273 493.973 147.614 20.013 20.013 20.027
122.500 2.84152 485.402 154.474 20.013 20.014 20.030
123.000 2.91165 476.131 162.032 20.013 20.014 20.033
2.91171 476.118 162.034
123.500 2.98321 465.970 170.492 20.013 20.015 20.036
124.000 3.05623 454.634 180.140 20.013 20.017 20.039
124.400 3.11572 444.409 188.993 20.013 20.018 20.042
124.700 3.16100 435.802 196.554 20.013 20.020 20.045
125.000 3.20686 426.104 205.211 20.013 20.023 20.048

a
Saturated-vapour density r0 determined from a virial equation of state, truncated after the third virial
coefficient, by means of the vapour pressure ps (see section 3).

critical region, the values for k'T and k0T were calculated by means of equation (1)
for each isotherm at ( p − ps ) = 0. The vapour pressures determined in the measuring
cell at liquid-levels of about 5 per cent (in the course of the r0 measurements)
TABLE 2. ( p, r, T ) measurements close to the (vapour + liquid) phase boundary in the critical region of nitrogen,
where T is the temperature (ITS-90), p the pressure, ps the vapour pressure, r the density, Dr0 and Dr' the total
uncertainties of the saturation densities, and kT the isothermal compressibility

Vapour densities Liquid densities

T ps a ps − p r k0T c ps b ps − p r k'T c
102·Dr0/r0 102Dr'/r'
K MPa 10 ·Pa
2
kg·m−3 MPa−1 MPa 102·Pa kg·m−3 MPa−1

125.000 3.206857 2 0.00 205.211 d 0.048 3.22 2

0.000410 1.67 205.098 0.10
7.31 204.734
12.15 204.423
19.08 203.979
26.12 203.540
32.10 203.172
125.250 3.245578 2 0.00 213.631 d 0.054 3.94 2 3.245578 2 0.00 416.815 d 0.028 1.11 2
0.000410 1.57 213.500 0.20 0.000410 2.44 416.926 0.02
6.83 213.062 7.07 417.144
13.10 212.552 16.80 417.589
19.49 212.046 27.45 418.085
26.27 211.520
32.63 211.031
125.500 3.284751 2 0.00 223.680 d 0.070 5.18 2 3.284751 2 0.00 405.883 d 0.037 1.62 2
0.000420 2.03 223.438 0.30 0.000420 2.35 406.037 0.03
5.02 223.107 9.73 406.522
9.67 222.589 14.16 406.808
14.48 222.071 19.57 407.168
19.97 221.488
27.64 220.707
125.700 3.316416 2 0.00 233.615 d 0.080 7.2 2 3.316416 2 0.00 395.220 d 0.045 2.53 2
0.000420 1.97 233.286 0.4 0.000420 2.28 395.448 0.05
5.02 232.791 6.72 395.891
8.60 232.234 11.22 396.341
12.13 231.697 14.26 396.636
17.34 230.936 16.72 396.890
26.83 229.640
125.850 3.340385 2 0.00 242.922 d 0.10 10.1 2 3.340385 2 0.00 385.301 d 0.055 3.97 2
0.000430 2.54 242.315 1.0 0.000430 2.22 385.640 0.10
5.27 241.668 6.30 386.242
8.47 240.984 9.12 386.677
14.73 239.694 11.80 387.067
21.54 238.394 14.73 387.499
30.63 236.823 17.53 387.882
18.94 388.099
19.91 388.241
20.92 388.381
125.950 3.356473 2 0.00 250.766 d 0.13 14.3 2 3.356473 2 0.00 377.000 d 0.075 6.4 2
0.000430 1.34 250.299 2.0 0.000430 2.18 377.522 0.4
2.22 249.992 5.51 378.285
6.61 248.591 7.90 378.818
11.47 247.200 10.53 379.410
16.59 245.869 12.88 379.890
22.23 244.527 15.46 380.392
28.35 243.182 17.94 380.862
126.020 3.367800 2 0.00 257.615 d 0.16 20.5 2 3.367774 2 0.00 369.827 d 0.10 9.8 2
0.000430 1.38 256.919 5.0 0.000430 2.13 370.578 1.0
2.06 256.583 3.87 371.178
5.17 255.195 5.65 371.694
8.49 253.915 7.29 372.237
12.08 252.610 9.23 372.840
16.12 251.303 11.10 373.394
20.60 249.951 13.12 373.934
27.47 248.123 14.99 374.448
126.080 3.377552 2 0.00 265.200 d 0.25 32 2 3.377512 2 0.00 361.960 d 0.13 17 2
0.000430 −0.04 265.239 7 0.000430 2.09 363.180 3
0.75 264.659 3.53 363.898
2.52 263.289 5.28 364.721
4.57 261.962 7.05 365.599
6.90 260.613 8.98 366.472
9.60 259.272 11.26 367.339
12.57 257.929 13.42 368.168
17.29 256.053 17.29 369.484
23.70 253.871 21.71 370.860
29.72 252.005 26.66 372.240
 ( p, r, T ) measurements on nitrogen 1163

TABLE 2—continued

Vapour densities Liquid densities

T ps a ps − p r k0T c ps b ps − p r k'T c
102·Dr0/r0 102Dr'/r'
K MPa 10 ·Pa
2
kg·m−3 MPa−1 MPa 102·Pa kg·m−3 MPa−1

126.120 3.384081 2 0.00 272.000 d 0.34 51 2 3.384043 2 0.00 354.950 d 0.19 30 2

0.000430 0.07 271.902 13 0.000430 2.05 356.908 9
0.52 271.361 3.84 358.331
0.54 271.343 5.73 359.619
2.22 269.389 7.47 360.783
4.38 267.434 10.07 362.278
6.94 265.484 13.19 363.760
9.98 263.610 16.85 365.310
12.80 261.992 21.18 366.932
15.46 260.645
18.52 259.295
18.65 259.309
21.85 257.969
25.45 256.633
126.150 3.388982 2 0.00 278.970 d 0.55 90 2 3.388966 2 0.00 347.920 d 0.30 61 2
0.000430 −0.04 278.886 25 0.000430 2.02 351.357 15
0.22 278.531 3.19 352.800
0.38 278.166 4.35 353.960
0.45 277.952 5.81 355.403
0.47 277.900 7.54 356.864
1.66 275.721 9.71 358.353
1.77 275.441 13.07 360.325
2.41 274.479 16.92 362.210
4.04 272.430 21.56 364.155
6.13 270.370 26.57 365.980
8.48 268.339
11.24 266.373
14.72 264.351
17.30 263.005
20.01 261.589
33.70 256.301
126.170 3.392249 2 0.00 285.700 d 0.90 180 2
0.000430 −0.06 285.270 50
0.20 284.550
0.33 284.141
0.43 283.895
0.44 283.449
0.57 283.183
0.58 282.925
0.72 282.625
0.76 282.551
0.80 282.606
0.84 282.580
0.87 282.018
0.89 282.149
0.93 282.059
1.54 280.647
2.20 279.215
2.86 277.912
4.65 275.082
7.18 272.216
9.51 270.162
12.24 268.126
14.68 266.534
17.47 264.912
20.61 263.305
23.91 261.713
120.192 e 3.3958 e 313.3 e

a
Vapour pressure, measured directly after measuring the vapour densities, followed by its total uncertainty.
b
Vapour pressure, measured directly after measuring the liquid densities, followed by its total uncertainty.
c
Isothermal compressibility kT = r−1(1r/1p)T for the saturated-vapour k0T or the saturated-liquid k'T followed by its estimated uncertainty.
d
Saturated-liquid density r' or saturated-vapour density r0 obtained by extrapolation of the ( p, r, T ) measurements to the saturation line
( p − ps ) = 0.
e
Critical constants not measured but determined in section 4.
1164 P. Nowak, R. Kleinrahm, and W. Wagner

FIGURE 2. The ( p, r, T ) measurements on isotherms close to the (vapour + liquid) phase boundary
of nitrogen, where ps is the vapour pressure, rc is the critical density, r' the saturated-liquid density, and
r0 the saturated-vapour density.

and about 90 per cent (together with the r' measurements) are listed separately in
table 2.

UNCERTAINTY

The total uncertainty of the vapour pressures and the saturated-liquid and
saturated-vapour densities can be calculated by applying the Gaussian
error-propagation formula. Apart from the experimental uncertainties of the
individual quantities p, r, and T, which were given in the first part of this work(1)
and discussed in a previous paper,(3) one additional uncertainty must be taken
into account, namely the uncertainty of the extrapolation in the critical region
for the determination of the saturation densities from the isothermal ( p, r, T )
measurements. The influence of impurities on the vapour pressure and on the
saturated densities is negligible, as discussed in section 2. In general, each of the
 ( p, r, T ) measurements on nitrogen 1165

mentioned uncertainties consists of a random part and several superimposed

systematic parts, some of them depending on temperature, pressure, and time.
By taking into account all these effects appropriately, the total uncertainty of the
vapour pressures, the saturated-liquid densities, and the saturated-vapour densities
were calculated and listed in tables 1 and 2. The reliability of the given total
uncertainties was estimated to be about 95 per cent.
Since the saturated-vapour densities at T = 83 K and below are small
(Q8.2 kg·m−3 ), these densities can be calculated more accurately than they can be
measured. Therefore, the saturated-vapour densities in the temperature range from
83 K to 64 K were calculated from a virial equation of state, truncated after the third
virial coefficient, by means of the measured vapour pressures ps . The r0 values
determined in this way are also listed in table 1. The virial equation of state used
for the calculations covers a temperature range from 64 K to 340 K at densities up
to 30 kg·m−3. It was fitted to the ( p, r, T ) values published in the first part of this
work(1) and additionally to the highly accurate speed of sound measurements of
Ewing and Trusler(8) at low pressures in the temperature range from 80 K to 373 K.
The total uncertainty of these calculated saturated-vapour densities corresponds to
the uncertainty of the vapour pressure.

4. New correlation equations and new critical constants

Based on the experimental results given in tables 1 and 2, new correlation equations
for the vapour pressure ps , the saturated-liquid density r', and the saturated-vapour
density r0 have been established. The equations represent the measured values along
the entire coexistence curve within their reproducibilities.
The equation for the vapour pressure ps of nitrogen has the form:

ln( ps /pc ) = (Tc /T )·(N1 t + N2 t 1.5 + N3 t 2.5 + N4 t 5 ), (2)

with t = (1 − T/Tc ), the critical temperature Tc = 126.192 K, the critical pressure

pc = 3.3958 MPa, and the coefficients N1 = −6.12445284, N2 = 1.26327220,
N3 = −0.765910082, and N4 = −1.77570564. For computer program verification
the following test values are given: ps = 0.014604762 MPa at T = 64 K;
ps = 3.36450851 MPa at T = 126 K. {Equation (2) is the optimized four-coefficient
equation which represents the experimental values best. However, as the
Wagner vapour pressure equation in the 3,6 form, namely ln(ps /pc ) = (Tc /T )·
(N1 t + N2 t 1.5 + N3 t 3 + N4 t 6 ), has been used for a great number of substances and
is still incorporated in numerous thermophysical-property calculation packages, the
corresponding coefficients of nitrogen are also given: N1 = −6.10959526,
N2 = 1.17270056, N3 = −1.12707288, and N4 = −1.62300202. The 3,6 form is only
slightly inferior to equation (2) where the deviations from the new measurements
are E=1.0·10−4·ps =.}
The equation for the saturated-liquid density r' has the form:

ln(r'/rc ) = (N1 t 0.3294 + N2 t 4/6 + N3 t 16/6 + N4 t 35/6 ), (3)

1166 P. Nowak, R. Kleinrahm, and W. Wagner

with t = (1 − T/Tc ), the critical temperature Tc = 126.192 K, the critical density

rc = 313.3 kg·m−3, and the coefficients N1 = 1.48654237, N2 = −0.280476066,
N3 = 0.0894143085, and N4 = −0.119879866. For computer program verification
the following test values are given: r' = 863.731200 kg·m−3 at T = 64 K;
r' = 371.982106 kg·m−3 at T = 126 K.
The equation for the saturated-vapour density r0 has the form:
ln(r0/rc ) = (Tc /T )·(N1 t 0.34 + N2 t 5/6 + N3 t 7/6 + N4 t 13/6 + N5 t 28/6 ), (4)
with t = (1 − T/Tc ), the critical temperature Tc = 126.192 K, the critical density
rc = 313.3 kg·m−3, and the coefficients N1 = −1.70127164, N2 = −3.70402649,
N3 = 1.29859383, N4 = −0.561424977, and N5 = −2.68505381. For computer
program verification the following test values are given: r0 = 0.777045028 kg·m−3 at
T = 64 K; r0 = 255.574511 kg·m−3 at T = 126 K. The values 0.3294 and 0.34 for the
exponent of t in the first term of equations (3) and (4), respectively, have nothing
to do with the critical exponent with regard to the saturation densities.
The structure of equations (2) to (4) has been obtained using the optimization
method, as developed by Setzmann and Wagner.(9)
The critical constants Tc , rc , and pc used in equations (2) to (4) have been
determined by means of the saturation densities and vapour pressures reported in
this work. For this purpose, the parabolic equation:

{(r/rc ) − 1} = N1 (1 − T/Tc ) 2 N2 (1 − T/Tc )b, (5)

was fitted simultaneously to various selected sets of saturated-liquid and

saturated-vapour densities in the critical region listed in tables 1 and 2; these selected
densities are shown in figure 3 together with the results of other workers.(10,11)
The optimum combination of Tc , rc , b, N1 , and N2 was obtained by a
non-linear least-square fit. Only the saturation densities between T = 125.7 K
and T = 126.17 K were finally used for the determination of Tc and rc . The results
of these investigations have led to the following values of the critical constants
and their uncertainties: Tc = (126.192 2 0.010) K, rc = (313.3 2 0.4) kg·m−3, and
pc = (3.3958 2 0.0017) MPa. The critical pressure was calculated from equation
(2) for the determined critical temperature. Its uncertainty consists of the
uncertainty of the vapour pressure equation (2) and the uncertainty of the critical
temperature.
In a further evaluation, the critical density was checked by means of the
rectilinear diameter r d = (r' + r0)/2. The values of rd could be calculated simply
because the saturated-liquid and saturated-vapour densities were measured up to
T = 126.15 K at exactly the same temperatures. Figure 4 shows the deviation of the
experimental values of the rectilinear diameter from values calculated from the
equation:
{(rd/rc ) − 1} = 0.770788(1 − T/Tc ), (6)

with Tc = 126.192 K and rc = 313.3 kg·m−3. Equation (6) corresponds to a

straight line through the critical density and the density r td at the triple-point
 ( p, r, T ) measurements on nitrogen 1167

FIGURE 3. Experimental densities r' of the saturated liquid and r0 of the saturated vapour in the
critical region of nitrogen: W, this work; t== , Goldman and Scrase;(10) q, Zozulya and Blagoi.(11)

FIGURE 4. Relative deviations of the rectilinear diameter r d = ( r' + r0)/2 of nitrogen, calculated from
the experimental results of this work for r' and r0. The zero line corresponds to equation (6), a straight
line through the critical density rc and r td at the triple-point temperature. The uncertainty of the r d values
corresponds to about one third of the size of the symbols.
1168 P. Nowak, R. Kleinrahm, and W. Wagner

FIGURE 5. Relative deviations of the rectilinear diameter r d = ( r' + r0)/2 from values calculated from
equation (7) in the critical region of nitrogen. A: uncertainty of r d values; B: reproducibility of r d values;
C: uncertainty of the critical density rc .

temperature. The maximum deviation is about −7.7·10−3·r d. Figure 5 shows an

enlarged section of figure 4 for the temperature range (Tc − 3.7 K) E T E Tc .
Taking into account the reproducibilities of the measured saturation densities,
the probable course of r d in the critical region is plotted; the course of rd turns
down significantly for temperatures above 125.7 K. Taking into account our
experiences based on the investigations of the critical region of methane,(4)
carbon dioxide,(5) argon,(6) and ethylene,(7) the illustrated course of r d is the
probable one. But it should be noted that either a straight course or a stronger
curvature of r d might also be possible. Thus, the uncertainty of rc , which is also
shown in figure 5, was conservatively estimated in such a way that both cases are
included.

5. Discussion
A comparison of the experimental results of this work with selected measurements
on nitrogen of other experimentalists(10–21) is presented in figures 6 to 8. A summary
of the selected sets of measurements together with their uncertainties given by the
authors is listed in table 3; a comprehensive compilation of experimental ( p, r, T )
values published prior to 1986 is given by Jacobsen et al.(22) Figures 6 to 8 show the
deviations between the measured values and values calculated from the new
correlation equations for the vapour pressure, equation (2), the saturated-liquid
density, equation (3), and the saturated-vapour density, equation (4). In the case of
very comprehensive sets of mesurements, not all values have been plotted for clarity.
Moreover, the figures show a comparison with values calculated from the correlation
equations published by Jacobsen et al.(22) Figure 6 additionally shows a comparison
with values from the vapour pressure equation of Wagner.(12) This vapour pressure
equation has been adopted in the IUPAC book on nitrogen;(23) however, separate
correlation equations for the saturated-liquid and saturated-vapour densities are not
given in this book.
 ( p, r, T ) measurements on nitrogen 1169

TABLE 3. Summary of the selected experimental results for the vapour pressure ps , the saturated-liquid density r', and
the saturated-vapour density r0 together with their total experimental uncertainties given by the authors

Number
of points Temperature Relative uncertainties
range
Source Year ps r' r0 T/K 2
10 ·Dps /ps 102·Dr'/r' 102·Dr0/r0

Armstrong(14) 1954 74 64 to 78 0.002

Moussa et al.(15) 1966 32 63 to 78 0.03
Weber(13) 1970 47 65 to 126 0.04 to 0.1
Wagner(12) 1973 68 63 to 126 0.04 to 0.1
Zozulya(16) 1976 28 84 to 126 0.06
a
van Itterbeek and Verbeke(17) 1960 9 63 to 91
a
Cockett et al.(18) 1968 10 80 to 125
Goldman and Scrase(10) 1969 30 79 to 126 0.03 to 0.16
Zozulya and Blagoi(11) 1975 14 14 120 to 126 0.06 to 0.25 0.06 to 0.25
Haynes et al.(19) 1976 19 95 to 120 0.12 to 0.18
Orrit and Laupretre(20) 1978 19 79 to 111 0.15
Straty and Diller(21) 1980 9 80 to 120 0.15
This work 1997 57 52 50 64 to 126 0.015 b 0.012 b 0.025 b

a
Not specified by the authors.
b
In the critical region and near the triple-point temperature the uncertainties exceed this value, see tables 1 and 2.

VAPOUR PRESSURES
Figure 6 shows the comparison between the experimental vapour pressures ps and
the values calculated from the new vapour pressure equation (2). For temperatures
T E 80 K nearly all measuring points confirm each other within the experimental
uncertainty of this work. For T q 80 K, the ps values of Weber(13) and Wagner(12)

0.08
102·{pσ – pσ(calc)}/pσ

Tt Tb Tc
–0.08
60 70 80 90 100 110 120 130
T/K
FIGURE 6. Relative deviations of experimental vapour pressures ps and of calculated vapour pressures
of nitrogen from values ps (calc) calculated from equation (2). W, this work; q, Wagner;(12) t, = Weber;(13)
y, Armstrong;(14) , Moussa et al.;(15) r, = Zozulya;(16) – · – · –, values calculated from the vapour pressure
equation of Jacobsen et al.;(22) – ·· –·· –, values calculated from the vapour pressure equation of Wagner.(12)
The grey area corresponds to the experimental uncertainty of this work. The temperature Tb = 77.355 K
indicates the normal boiling temperature at ps = 0.101325 MPa.
1170 P. Nowak, R. Kleinrahm, and W. Wagner

show, compared with the values in this work, a slightly rising course. Taking into
account the scatter of these two measuring runs, their maximum deviation is about
4·10−4·ps at T 1 120 K which corresponds to their uncertainty of E=4·10−4·ps =. The
measurements of Zozulya(16) agree very well with the new values for T q 120 K
(deviations Q =5·10−5·ps =) but they show increased deviations, up to −8·10−4·ps and
more, for T Q 100 K. These deviations are greater than the claimed uncertainty of
Zozulya’s(16) values of E=6·10−4·ps =.
The vapour pressure equations of Wagner(12) and of Jacobsen et al.(22) yield values
that mostly deviate by not more than the experimental uncertainty of this work. Due
to the previous data situation for T q 105 K the two equations run slightly outside
the experimental uncertainty of the new values with a maximum deviation of
5·10−4·ps .

SATURATED-LIQUID DENSITIES

Figure 7 presents the comparison between experimental saturated-liquid densities

and values calculated from the saturated-liquid density equation (3). For T Q 80 K
the data situation was very poor because only the values of van Itterbeek
and Verbeke(17) were available. These measurements, for which no uncertainty was
given, deviate by more than 2·10−3·r' from the new values. The two measuring runs
of Cockett et al.(18) and of Goldman and Scrase(10) mostly deviate systematically
by about −2.5·10−3·r' in the temperature range between 80 K and 120 K.
While the values of Goldman and Scrase(10) exceed their claimed uncertainty of
E=1.6·10−3·r'=, no uncertainties are given by Cockett et al.(18) The saturated-liquid
densities measured by Zozulya and Blagoi(11) for T e 120 K also exceed the

0.3
102·{ρ' – ρ'(calc)}/ρ'

Tt Tb Tc
–0.3
60 70 80 90 100 110 120 130
T/K
FIGURE 7. Relative deviations of experimental saturated-liquid densities r' and of calculated
saturated-liquid densities of nitrogen from values r' (calc) calculated from equation (3). W, this work;
== , Goldman and Scrase;(10) r, = Zozulya and Blagoi;(11) r, van Itterbeek and Verbeke;(17) r, = Cockett
t
et al.;(18) u, Haynes et al.;(19) , Orrit and Laupretre;(20) w, Straty and Diller;(21) –·–·–, values calculated
from the saturated-liquid density equation of Jacobsen et al.(22) The grey area corresponds to the
experimental uncertainty of this work.
 ( p, r, T ) measurements on nitrogen 1171

claimed uncertainty of E=6·10−4·r'= at T 1 120 K (deviation −2·10−3·r'). The

measurements of Haynes et al.(19) and of Straty and Diller(21) which show a maximum
deviation of 1.5·10−3·r' and −1·10−3·r', respectively, are the only ones that can
be confirmed within their uncertainty range of about Q=1.5·10−3·r'=. The
saturated-liquid densities measured by Orrit and Laupretre(20) (uncertainty
E=1.5·10−3·r'=) agree well with the new values in the temperature range between
78 K and 106 K but, for T q 106 K, they show systematically rising deviations that
exceed 2·10−3·r'.
The values calculated from the correlation equation of Jacobsen et al.(22) confirm
the fact that the saturated-liquid densities of nitrogen were previously known only
to within the limits of Q=2·10−3·r'=. Hence, the new values presented in this work
describe the densities on the saturated-liquid line with an accuracy greater than ten
times that of the previous measurements.

SATURATED-VAPOUR DENSITIES
Figure 8 presents the comparison between experimental saturated-vapour densities
and values calculated from the saturated-vapour density equation (4). The r0
measurments of Zozulya and Blagoi(11) for T e 120 K deviate mostly by more
than the claimed uncertainty of E6·10−4·r'= at T 1 120 K and E=2.5·10−3·r'= at
T 1 126 K, respectively. Figure 8 illustrates clearly the unsatisfactory data situation
for nitrogen where no recent r0 measurements for T E 120 K are available. Since the
saturated-vapour densities for T E 83 K are very small (Q8.2 kg·m−3 ) they have been
calculated by means of a virial equation and the vapour pressures measured in this
work (see section 3 and table 1). The results are also plotted in figure 8. At T = 86 K
the calculated and the measured saturated-vapour density have nearly the same value
(see table 1), and consequently they are consistent with each other. Moreover, such

0.3
102·{ρ" – ρ"(calc)}/ρ"

Tt Tb Tc
–0.3
60 70 80 90 100 110 120 130
T/K
FIGURE 8. Relative deviations of experimental saturated-vapour densities r0 and of calculated
saturated-vapour densities of nitrogen from values r0(calc) calculated from equation (4). W, this work;
w, this work, calculated values (see table 1 and section 3); r, = Zozulya and Blagoi;(11) –·–·–, values
calculated from the saturated-vapour density equation of Jacobsen et al.(22) The grey area corresponds to
the experimental uncertainty of this work.
1172 P. Nowak, R. Kleinrahm, and W. Wagner

TABLE 4. Compilation of values for the critical temperature Tc (ITS-90) and its uncertainty DTc
determined by previous workers

Source Year Tc /K DTc /K Procedure

Voronel et al.(39) 1966 126.193 0.003 Evaluation of isochoric heat capacities

This work 1997 126.192 0.010 Evaluation of equations for r' and r0
Zozulya and Blagoi(11) 1975 126.205 0.003 Evaluation of equations for r' and r0
White et al.(40) 1951 126.259 0.040 Visual observation of meniscus

calculated values remain consistent with the newly measured values even up to
r0 1 19 kg·m−3 at T = 93 K which are, however, not shown in the figure.
Due to the previous data situation, the values calculated from the correlation
equation for r0 of Jacobsen et al.(22) show deviations up to =2.5·10−3·r0= and more
(for T q 122 K).

CRITICAL TEMPERATURE
Table 4 shows a compilation of selected published values for the critical temperature
of nitrogen; the estimated experimental uncertainties are also given. The Tc value of
Voronel et al.(39) agrees very well with the critical temperature determined in this
work within its experimental uncertainty, which was conservatively estimated.
The Tc value of Zozulya and Blagoi(11) can just be confirmed within the combined
experimental uncertainties. The critical temperature determined by White et al.(40) is
obviously too high.

TRIPLE-POINT TEMPERATURE
To complete our comprehensive measurements, we also measured the triple-point
temperature as Tt = (63.151 2 0.003) K and the vapour pressure at Tt as
pt = (12.523 2 0.010) kPa. By means of equations (3) and (4) the saturated-liquid and
-vapour densities at the triple-point temperature have been calculated with the
following result: r't = (867.21 2 0.09) kg·m−3 and r0t = (0.6744 2 0.0006) kg·m−3.
The triple-point temperature of nitrogen was determined by several authors with a
very small uncertainty. The temperatures measured by Lee et al.,(41) Ancsin,(42) and
Pavese et al.(43) only range between T = 63.1505 K and T = 63.1509 K with an
uncertainty E=5·10−4 K=. Therefore, the triple-point temperature determined in this
work is in excellent agreement with these values.

ISOTHERMAL COMPRESSIBILITIES
Figure 9 shows the derived isothermal compressibilities at saturation, namely k'T and
k0T for T e 125 K together with the estimated uncertainties. The determination of
the isothermal compressibilities is mentioned in section 3. As kT : a at the critical
point, 1/kT is plotted in figure 9. For comparison, figure 9 also shows the isothermal
compressibilities calculated from the equations of state from the IUPAC(23) and from
Jacobsen et al.(22) Neither equation is able to reproduce the experimental values
within their uncertainties.
 ( p, r, T ) measurements on nitrogen 1173

FIGURE 9. Derived isothermal compressibilities kT = r−1(1r/1p)T on the (vapour + liquid) phase

boundary, where k'T is the isothermal compressibility of the saturated liquid and k0T that of the saturated
vapour. W, this work (see table 2); ——, values calculated from the equation of state of Jacobsen et al.,(22)
– – –, values calculated from the IUPAC equation of state.(23) The uncertainty of the experimental values
is smaller than the size of the symbols.

6. Conclusion
The conclusions from these brief comparisons can be summarized by the
statement that the (p, r, T ) data situation for nitrogen was only satisfactory in a few
sections of the (vapour + liquid) coexistence curve. With the new, comprehensive,
and consistent set of measurements presented in this work the previous data
situation has been considerably improved along the entire saturation curve from
the triple-point temperature to the critical temperature up to a high level of
accuracy. On the basis of these measurements, including the new measurements
in the homogeneous region,(1) a new equation of state is being developed by Span
et al.(2) This equation will describe the entire fluid region of nitrogen in reference
quality.

The authors are grateful to all who contributed to this work, especially to S. Müller
for performing a part of the measurements. Above all, we thank the Deutsche
Forschungsgemeinschaft for financial support of our work.
1174 P. Nowak, R. Kleinrahm, and W. Wagner

REFERENCES
1. Nowak, P.; Kleinrahm, R.; Wagner, W. J. Chem. Thermodynamics 1997, 29, 1137–1156.
2. Span, R.; Lemmon, E. W.; Jacobsen, A. T.; Wagner, W. A reference Quality Equation of State for
Nitrogen. Submitted to Int. J. Thermophys.
3. Kleinrahm, R.; Wagner, W. J. Chem. Thermodynamics 1986, 18, 739–760.
4. Kleinrahm, R.; Duschek, W.; Wagner, W. J. Chem. Thermodynamics 1986, 18, 1103–1114.
5. Duschek, W.; Kleinrahm, R.; Wagner, W. J. Chem. Thermodynamics 1990, 22, 841–864.
6. Gilgen, R.; Kleinrahm, R.; Wagner, W. J. Chem. Thermodynamics 1994, 26, 399–413.
7. Nowak, P.; Kleinrahm, R.; Wagner, W. J. Chem. Thermodynamics 1996, 28, 1441–1460.
8. Ewing, M. B.; Trusler, J. P. M. Physica 1992, 184A, 415–436.
9. Setzmann, U.; Wagner, W. Int. J. Thermophys. 1989, 10, 1103–1126.
10. Goldman, K.; Scrase, N. G. Physica 1969, 44, 555–586.
11. Zozulya, V. N.; Blagoi, Y. P. Sov. J. Low Temp. Phys. 1975, 1, 562–571.
12. Wagner, W. Cryogenics 1973, 13, 470–483.
13. Weber, L. A. J. Chem. Thermodynamics 1970, 2, 839–846.
14. Armstrong, G. T. J. Res. Nat. Bur. Stand. 1954, 53, 263–266.
15. Moussa, M. R.; Muijlwijk, R.; van Dijk, H. Physica 1966, 32, 900–912.
16. Zozulya, V. N. Sov. J. Low Temp. Phys. 1976, 2, 185–188.
17. van Itterbeek, A.; Verbeke, O. Physica 1960, 26, 931–938.
18. Cockett A. H.; Goldman, K.; Scrase, N. G. The Density of Liquid Nitrogen from 85 to 120 °K and
from the Saturation Boundary to 200 atm. Proceedings of the Second International Cryogenic
Engineering Conference, Brighton, England, May 1968. ILIFFE Science and Technology
Publications. 1968, pp. 276–280.
19. Haynes, W. M.; Hiza, M.; Frederick, N. V. Rev. Sci. Instrum. 1976, 47, 1237–1250.
20. Orrit, J. E.; Laupretre, J. M. Adv. Cryog. Eng. 1978, 23, 573–579.
21. Straty, G. C.; Diller, D. E. J. Chem. Thermodynamics 1980, 12, 927–936.
22. Jacobsen, R. T.; Stewart, R. B.; Jahangiri, M. J. Phys. Chem. Ref. Data 1986, 15, 735–909.
23. Angus, S.; de Reuck, K. M.; Armstrong, B.; Jacobsen, R. T.; Stewart, R. B. International
Thermodynamic Tables of the Fluid State. Vol. 6: Nitrogen. Pergamon: Oxford. 1979.
24. Albuquerque, G. M. N.; Calado, J. C. G.; Nunes da Ponte, M.; Staveley, L. A. K. Cryogenics 1980,
20, 416–418.
25. Cath, P. G.; Kamerlingh-Onnes, H. Commun. Phys. Lab. Univ. Leiden 1918, 152 d, 45–53.
26. Clusius, K.; Schleich, K. Helv. Chim. Acta 1958, 41, 1342–1358.
27. Crommelin, C. A.; Kamerlingh-Onnes, H. Commun. Phys. Lab. Univ. Leiden 1915, 145 d, 28–32.
28. Dodge, B. F.; Davies, H. N. J. Am. Chem. Soc. 1927, 49, 610–620.
29. Friedman, A. S.; White, D. J. Am. Chem. Soc. 1950, 72, 3931–3932.
30. Giauque, W. F.; Clayton, J. O. J. Am. Chem. Soc. 1933, 55, 4875–4889.
31. Henning, F. Z. Phys. 1926, 40, 775–785.
32. Heuse, W.; Otto, J. Ann. Phys. 1932, 406, 185–192.
33. Keesom, W. H.; Bijl, A. Physica 1937, 4, 305–310.
34. Michels, A.; Wassenaar, T.; de Graaff, W.; Prins, C. Physica 1953, 19, 26–28.
35. Porter, F.; Perry, J. H. J. Am. Chem. Soc. 1926, 48, 2059–2060.
36. Rodosevich, J. B.; Miller, R. C. A. I. Ch. E. J. 1973, 19, 729–735.
37. von Siemens, H. Ann. Phys. 1913, 42, 871–888.
38. Terry, M. J.; Lynch, J. T.; Bunclark, M. Mansell, K. R.; Staveley, L. A. K. J. Chem. Thermodynamics
1969, 1, 413–424.
39. Voronel, A. V.; Gorbunova, V. G.; Chskin, Y. P.; Schekochikhina, V. V. Sov. Phys. JETP 1966, 23,
597–601.
40. White, D.; Friedman, A. S.; Johnston, H. L. J. Am. Chem. Soc. 1951, 73, 5713–5715.
41. Lee, H. K.; Kang, K. H.; Rhee, C. Physica 1991, 169B, 451–452.
42. Ancsin, J. Can. J. Phys. 1974, 52, 1521–1531.
43. Pavese, F.; Ancsin, J.; Bonhoure, J.; Astrov, D. N. Metrologia 1984, 20, 127–144.

(Received 20 January 1997; in final form 27 March 1997)

WE-089