COPYRIGHT,

1942

AND

1943,

BY THE

AMERICAN INSTITUTE OF MINING AND METALLURGICAL ENGINEERS

(INCORPORATED)

Density of Natural Gases

By

MARSHALL

B.

STANDING* AND DONALD

L.

KATZ,* MEMBER

(New York Meeting,February

DENSITY data are reported on 16 saturated
hydrocarbon vapors at pressures ranging from
1000 to 8220 lb. per sq. in. and at temperatures
ranging from 35 to 250F. These data have
been used to extend the compressibility-factor
chart for natural gases up to ro.ooo lb. per
sq. in. The relatively large quantity of highboiling constituents present in high-pressure
vapors in equilibrium with crude oils makes it
necessary to include in the analysis of the gas
the molecular weight, density, and possibly
boiling range of the heptanes and heavier
fraction. Relationships have been presented by
which the density of gases may be obtained
directly from the temperature, pressure, and

A.I.M.E.

1941)

is the use of the ideal gas law corrected by a

compressibility factor. The method proposed by Kay? of correlating compres
sibility factors for gaseous mixtures on
pseudocritical properties" appears to be
satisfactory for natural gases." The meth
ane-compressibility factor has been shown
to deviate systematically from the behavior
of natural gases? and a chart giving these
corrected factors is available.
A method of computing specific volume
of gaseous mixtures from partial molal
volumes has been reported by Sage and
L acey. 8 Ciotmp e e d a t a for th e compu t a

tiIOn

gas
gravity provided the gases have a
common
source or are similar in composition.
INTRODUCTION

The densities of natural gases are neces

sary in many engineering computations in
petroleum production and utilization, Gas
reserves, changes in reservoir pressure, gra
dients in gas wells, metering of gases, pipe
line flow, and compression of gases are
typical problems requiring the density of
the gas. A decade ago, engineering compu
tations used ideal gas laws with deviations
up to 500 lb. per sq. in} Recent discoveries
of pools having pressures up to 7500 lb. per
sq. in., and installation of pressure-mainte
nance and recycling plants, have increased
the need for data on gas density at high
pressures.
METHODS OF COMPUTATION

The accepted method of computing the

density or specific volume of natural gases
Manuscript received at the officeof the Institute
Dec. 31, 1940. Issued as T.P. 1323 in PETROLEUM
uly 194I.
* University of Michigan.Ann Arbor, Mich.
, Referencesare at the end of the paper.

TECHNOLOGYJ,

are not yet available for all hydrocarbon

gases.
Most of the reported data on gas densi
ties have been in the single-phase region
removed from the dew point or saturation
condition. Since most gases in the reservoir
or while on contact with liquid during flow
are saturated, the determination of densi
ties under conditions of saturation is
particularly important.
Although there is
no reason to believe that the saturated
gases differ in behavior from those consider
ably removed from their dew points, the
usual experimental determinations find
difficulty on approaching the dew-point
conditions.
This paper reports the density of 16
saturated gases in equilibrium with crude
oils. The relationships developed to com
pute the density of gases to high pressure!
have been modified and enlarged. A simpli
fied method of obtaining gas density is
presented in the form of charts, which
apply for specific groups of gases.

MARSHALL B. STANDING AND DONALD

DATA OBTAINED

In the process of obtaining equilibrium

data on mixtures of natural gas and crude
oils, the density and analysis of the satu
rated vapor phases were obtained. These
data were in the range of 35 to 250F. and
1000 to 8220 lb. per sq. inch.
The vapors in equilibrium with crude oil
were transferred, by mercury displacement,
at constant temperature and pressure from
the equilibrium cell into pycnometers of
2s-ml. capacity.
The contents
of the
pycnometer were discharged into a low
temperature fractionating column, to ob
tain the analysis of the gas. In all cases the
pycnometer was weighed before and after
discharging the gas and any mercury ac
companying the gas was collected and
weighed. The combined error on weighing
is estimated as 0.04 gram out of 1.35
gram for the gas of lowest density and 11.2
grams for the gas of highest density.
The presence of relatively large quanti
ties of heptanes and heavier required refineTABLE

I.-Typical

L. KATZ

ments over the usual gas analysis. The

heptane-plus fractions were collected and
weighed as a liquid. The density and
molecular weight were measured if suffi
cient quantities were available.
A smooth curve relationship exists be
tween the liquid density and molecular
weight of the heptanes plus and was used if
the data for either were lacking. It was
found that there was a permanent hold-up
in the fractionating column packing as high
as 0.5 gram. Differences between the
weight of gas out of the pycnometer and the
weight of the gas accounted for in the
analysis varied up to this quantity. Accord
ingly, the weight of the heptanes and
heavier was taken as the difference between
the pycnometer weight of the gas and the
weight of the hydrocarbons of molecular
weight lower than heptane.
A typical analysis of the saturated vapor
is given in Table I. The absence of iso
butane is due to the method of blending a
butane-free gas with the Arkansas crude"
Analysis of Vapor

Pressure, 4310 Lb. per Sq. In. Absolute. Temperature, 250F.

Mol Frac-

tion

Compound

~~i:<
Heptanes

+ heavier

Heptanes

0.8058
0.0386
0.0197
0.0170

0.0280
0.0123

0.0148

0.0521

Mol
Fraction
X pp,

277.2

12.89
1. 16
0.51
1.14

13.5
59.0

542.3
275
7.2
10.8
8.2
13.6
5.5
6.4
18.7

442.7

640.2

28. SO

1132

weight,

139; specific gravity,

Gas gravity

= pp,
.!'_ =
_

Til - pT,

2850
2896

4310 = 673
640.2
.
250
460
442.7
= 1.603

PV = NZRT
PV = WZRT
M
Calculated density = lb. per cu. ft.

r,

21.2

7.8
IS. I

14. I

23.7
I I. I

1,000

molecular

PH

Mol
Fraction

673
712
617
551
483
485
451
435
360

344
549
666
766
829
846
900
914

0.0117

..................

and heavier:

Critical

Temperature, Critical
Pressure,
Deg. R
P"Lb.
per Sq.
In. Abs.

Mol
Fraction
X Mol
Weight

1.22
2.01

1.06
1. 27
7.24

69'60. 0.791.
0.985

.
Z from Fig. 2

0.933

N = ~ = pounds divided by mol weight

!V!: =cu.PMft . =

4310 X 2850
0.933 X 10.71 X
Experimental density = 0.277 X 62.43 = 1729 lb. per cu. ft.

ZRT

710

=
.

1735 lb

. per

DENSITY

OF NATURAL GASES

and varying quantities of natural gasoline?

that contained normal butane but no
isobutane. Data on the density and proper
ties of 16 saturated vapors in equilibrium
with a crude oil are given in Table 2.
PSEUDOCRITICAL
HEPTANES

CONDITIONS
AND

vised to be in agreement with Smith and

Watson's newly defined boiling points.
This mean average boiling point may be
converted to the molal average from the
knowledge of the initial boiling point of
about 200F. for the heptanes and heavier
fraction. The molal average boiling point
and gravity give the pseudocritical tempera
ture by Smith and Watson's relationship.
The densities and molecular weights of
heptanes and heavier of Table 2 have been
used to compute the pseudocriticals shown
in Fig. I. It should be noted that the

OF

HEAVIER

FRACTIONS

Computation of the pseudocritical tem

perature and pressure of a gas is given in
Table 1. Because of the high content and
high molecular weight of the heptanes and
TABLE

Run

2.-Properties of Saturated Gases

Temper ..
Pressure, Lb. at.u re,
Gravity.
Air
= I
per Sq.

Mol Per
Cent

Mol Per
Cent

Deg. F.

CR,

C,'

MolWt.
C7'

ExperiDensity
C,+
6%0

Experimental

mental
Com-

Gram

bility

Density. pressi-

No.
In. Abs.

Z
----

---

...
A-5 .....
1,000
1,600
A-4L ... ...
A-I. .. ....
3.185
A-2I. .. ...
5.270
A-3 ...
8,220
B-I. ....
......
2,920
C-2 .....
I,oro
C-I. .... ....
2.880
C-3I. ..
.....
5.330
D-1. . ....
......
4.330
E-1. ..... . . . . . . . . . . .
4.195
F-I. ....
3.185
F-2 ....
4.310
F-3 ... .... ....
5.530
G-2 ................
3.485
G-3 ...............
4.970

120
120
120
120
120
250
120
120
120
120
120
250
250
250
35
35

---

---

0.659
0.671
0766
0.840
0.899
0 766
0.629
0.736

---

09200

0.9158
0.8800
0.8588
0.8440
0.8750
09345
0.8895
1.040
0.7995
0.876
0.8340
0736 0.8920
0915 0.8450
0 8060
0985
1.070
o. 7840
0.8230
0926
0.7810
1.074

++

0.0090
0.0124
0.0303
0.0400

0.0444
0.0268
0.0021

0.or80
0.0573
0.0375
0.0243
0.0495
0.0521

0.0588
0.0434
0.0631

--I lob
110b

131
127
144
IISb

--0.742'
0.742'
0.775'
0.788
0.802
O.753b

100b

O.72sb

113
151
123

0.750'
0 800
o. 788
0.771
0.768
O. 791
0 808
0.765
0.792

130b

130
139
157
128
145'

Series A and
7.000Bcu. ft. gas. per bbl. 50 vol. crude
50 vol. gasoline.
50 vol. gasoline.
Series C.. F3.500
and cu.
G ft. gas per bbl. 50 vol. crude
. 7.000 cu. ft. gas per bbl. 25 vol. crude
Series D
7.000
cu.
ft.
gas
per bbl. 75 vol. crude
Series E
.
Estimated.

Natural gas: gravity

+ 75
+ 25

----

---

0.0564
0.097
0 224
0.340
0.404
0.154
0.0556
0.203

0400
0.327
0.261

0.192

0.277
0.349
0.383
0.463

0.870
0.826
0.8ro
0970
1358
0.884
0.852
0.777
1.031

0.865
0.879
0.925
0.931

1.031

0.737

1.008

vol. gasoline.
vol. gasoline.

0.596; CH" 93.20 per cent; C2Hs, 4.25 per cent; CaHs, r.6r per cent;
N2, 0.43 per cent; C02, 0.5r per cent.

heavier fractions, the usual procedure of

using the critical properties of heptanes or
octanes for this fraction is no longer reliable.
Smith and Watson!" reported a method
of obtaining the pseudocritical pressure
directly from liquid density and molecular
weight. For the pseudocritical temperature,
the molal average boiling point must be
known in addition to the liquid density.
The mean average boiling point'? may be
obtained from the density and molecular
weight by a chart similar to that given by
Egloff and Nelson (ref. 3, p. 235) and re-

calculated values deviate somewhat from

the normal paraffin curve of critical tem
peratures and pressures as a function of
molecular weight. The difference between
the calculated pseudocriticals and the criti
cals of the normal paraffin hydrocarbons is
much smaller when using the molecular
weight of the heptanes and heavier than
when using the liquid density as the cor
relating factor. In no case would the critical
properties of heptanes have been repre
sentative of the heptanes and heavier frac
tion. Since the data given are for saturated

MARSHALL B. STANDING AND DONALD L. KATZ

vapors, it is indicated that the molecular

weight, liquid density, and boiling range if
possible, should be measured for gases at
equilibrium above 1000 lb. or with appreci-

I43

density of the saturated vapors at 8220 lb.

per sq. in. and 120F. checked the computed
density within the error of reading the
methane chart.

1300

~v;

.>'

Vv:;

v-)7
/~

/'

0
00

"

c.~..'~Ii ":"l--

,,~\o

~o"

o /

so 0

<,
.0.0

"'"

.,

<,

.0,

t--.......

~:r"''''e-s

I"'O~..~-!~
300

"
-of.~

"1""~/f'y~
"'''tJ,iOC

200

eo

FIG.

able quantities
present.

I.-VARIATION

100
120
MOL.CULAR

140

160

"'81"'3
180

200

WEIG~T

OF PSEUDOCRITICALSWITH MOLECULAR WEIGHT.

of heptanes and heavier

COMPRESSIBILITY FACTORS FOR NATURAL

GASES

The pycnometer density and the molecu

lar weight of the gases, gas gravity referred
to air, were used to compute the compressi
bility factors of the saturated vapors, Table
2. Excellent agreement was found between
these values and those from the chart for
natural gases- up to reduced pressures of
about 5. At higher pressures, the data
deviated from the natural-gas chart and
were closer to the values from the chart
based on pure methane. The measured

The data on gaseous mixtures available

to construct the chart for natural gases"
were meager above 3000 lb., and the chart
had been extended only to a reduced pres
sure of 8. Since it did not cover the pressure
range of this investigation, and was shown
to deviate consistently at pressures above
reduced pressures of 5, a revised and
extended plot was prepared (Fig. 2). This
compressibility-factor chart is identical
with the former chartt-" up to reduced
pressures of 4. Using the data of Table 2
and the methane chart as a guide, Fig. 2
was drawn up to a reduced pressure of IS or
9000 to 10,000 lb. per sq. inch.

DENSITY OF NATURAL GASES

144

RELATIONSHIPS OF GAS GRAVITY AND

curves for miscellaneous gases.' All gases

whose pseudocritical temperatures and
pressures fit these gravity curves have the
same density for a temperature, pressure
and gas gravity independent of composi-

GAS DENSITY

The gas gravity, compared to air, and

pseudocritical conditions gave a smooth
relationship when compared with the

PSEUDO REDUCED PRESSURE

2

NO.

cr

0
IU

~
~

..J

III
III
III

1.4 N

cr

W
Q.

cr

0
I-

::l

0
U

1.3

..J

III

2~w

cr
Q.

::!:

1.1

1.0

10

II

12

PSEUDO REDUCED PRESSURE

FIG. 2.-COMPREsSmILITY FACTORS.

13

14

0
U

145
For the gases represented by Fig. 3,
density plots have been prepared. Fig. 4
gives the density of these gases with gravi
ties (air = 1.0) from 0.6 to 0.8, and Fig. 5
from 0.9 to I.I for pressures from 100 to
10,000lb. per sq. in. and temperatures from

MARSHALL B. STANDING AND DONALD L. KATZ

tion. This simplification makes it possible

to construct charts giving gas density as a
function of temperature, pressure, and gas
gravity or molecular weight.
The hydrocarbon gases used to construct
the curve of Fig. 3 contained more than 83

.... ~

~ tfi-

.. r-r-----

I-

.....

...
0:
I-

.50
0:

::l

1- ...

~i .00
\)Z

t~

_9V

0:.

o
o
a

J......- V

Q.

I~

I--tyI. .s-

!---

350

..."~

o wISCELL.ANEOUS DATA
AUTHORS' DATA

0_60

FIG.

3.-VAR1ATION

0.70

GAS

0.10
GRAVITY

0.90
IAIA-I)

1.00

1.10

OF PSEUDOCR1TICALSWITH GAS GRAVITY.

per cent of combined methane and heptanes

and heavier fraction. Gases containing over
2 or 3 per cent of nitrogen or high con
centrations of ethane, propane, and butanes
deviate from the curves given. In all such
cases, whenever the analysis is available,
the pseudocritical conditions should be
computed to establish the position of the
curve on Fig. 3.

o to 300F. Table 3 shows a comparison

between the measured density, the density
by the calculated pseudocriticals and Fig. 2,
and the density from Figs. 4 and 5 for
the authors' data. As would be expected,
the densities computed from individual
pseudocritical conditions and the compres
sibility-factor chart are superior to those
from the gravity-density charts and are in

DENSITY

OF NATURAL GASES

GAS

GAS

FIG. 4.-DEN5ITIES

OF NATURAL GASES.

MARSHALL B. STANDING AND DONALD

L. KATZ

100

FIG.

5 ..,--DENSITIES

OF NATURAL GASES.

I47

148

DENSITY

TABLE3.-Comparison
Pressure,
Lb. per
Sq. In.
Abs.

Run No.

A-5 ....
A-4I ..
A-I. .
A-2 ...
A-3 ..
B-I. .. :
C-2 ..

..

1,000

1,600
3,185
5.270

...

8,220
2.920
1,010

.......
C-31.
5,330
D-I. . ...
.....
4,330
E-I. .
.....
4,195
F-I.
........
3,185
F-2 ..
. . . . . . . . . . 4,310
F-3 ...
. . . . . . . . . . . . 5,530
G-2 .... . . . . . . . . . . . . .
3,485
G-3 ...
. . . . . . . . 4,970

OF

NATURAL

of Experimental and Calculated Densities

Temper- Gravity,
ature,
Air = I
Deg. F.

Experimental
Density,
Lb. per
Cu. Ft.
---

Pseudo
Reduced
Temperature

mental
Density,
Lb. per
Cu. Ft.

lated
Density.
Fig.2,
Lb. per
Cu. Ft.

Density.
Fig. 4,
Lb. per
Cu. Ft.

3.50
6.12
144
20.8

355
6. IS
148

0.659
0.671
0.766
0.840
0.899
0766
0.629

1.50
242
487
8.09
12.7
444
1.50

1. 55
1. 54
1.47
1.41
1. 37
1. 77
1. 59

352
6.06
14.0

120
120
120
250
250
250
35
35

1.04
0.876
0.736
0.915
0.985
1. 07
0.926
1.074

8.36
6.70
6.34
490
6.72
8.72
5.41
7.86

I. 29
1.37
1.48
1.69
1.60
1. 55
I. 14
1.08

250
204
16.3

TABLE4.-Comparison of Experimental
and Calculated Densities of Rio Bravo
Gas
Temperature,
Deg. F.

Pseudo
Reduced
Pressure

120
120
120
120
120
250
120

good agreement with the experimental

values. However, for gases containing less
than 2 or 3 per cent nitrogen and more than
83 per cent of combined methane and
heptanes and heavier, Fig. 4 is a convenient
and quick method of obtaining densities for
general engineering calculations.

Pressure,
Lb. per Sq.
In. Abs.

GASES

Calculated

Density,
4,
Lb. per
Cu. Ft.

Density, Fig.
Fig. 2,
Lb. per
Cu. Ft.
---

---

2I.2

25.2
9.61
347

12.0

17.3
2I.8

23.9
28.9

252

9 55
3.31

2IoO

253
9.59
3.39

255

25.6

16.3
12.8
17.3
21.6
24.2
28.6

16.6
12.7
18.0

20.2

20.1

21.2

puted and experimental results should be

representative of gases of that type since
the data were not used in any of the
correlations.
A set of charts giving the density as a
function of gas gravity, temperature, and
pressure could be prepared for any series of
gases that have a smooth relationship be
tween gas gravity and pseudocritical condi
tions. Gases from a common source, such as
covered by the data of this paper, or gases
of similar composition may be expected to
have such a relationship.
CONCLUSIONS

I. Density data have been reported on 16

saturated hydrocarbon vapors in the range
roo
of 35 to 250F. and 1000 to 8220 pounds.
3,500
roo
5.000
2. The compressibility-factor chart for
500
190
1.44
1.44
1. 44
natural gases has been revised and extended
1,000
190
3.00
302
302
2,000
190
6.32
638
640
to 10,000 pounds.
190
10.8
10.8
3,500
109
5,000
190
14.0
14.0
14.0
3. The molecular weight, density and
500
280
1. 25
1. 24
1. 25
boiling
range of the heptanes and heavier
1,000
280
2.52
255
255
2,000
280
5.16
5.24
525
fraction has been shown to be an important
280
3,500
8.82
9.02
8.94
280
II.8
II.8
II.8
5,000
part of the gas analysis for saturated vapors
at high pressure and temperature.
The recent density data of Sage and
4. Relationships have been presented by
Reamer" on the separator gas from the Rio which the density of hydrocarbon gases
Bravo field are compared with the com containing more than 83 per cent of com
puted densities and compressibility factors bined methane and heptanes plus may be
in Table 4. The agreement between com- obtained from temperature, pressure, and
gas gravity (air = 1.0) directly.

500

1,000
2,000

100

I. 74
100
100

1. 74
378
8.54
13.9
16.8

1. 75
3.76
846
13.9
16.9

3.75
8.45
138
16.8

MARSHALL B.

STANDING AND DONALD L. KATZ

2.

ACKNOWLEDGMENT

The authors are indebted to the Humble

Oil and Refining Co. for the natural gas, to
the Phillips Petroleum Company for the
crude oil, and to the Indian Territory
Illuminating Oil Co. for the natural gaso
line used in the investigation.
REFERENCES
r. G. G. Brown: Nat. Gas. Assn. Amer., Proc, r oth
Annual Convention, 54. May 1940.

3.
4.
5.

6.

149

G. G. Brown and D. E. Holcomb: Pelr. Engineer

ing (1940) II (5),23.
G. Egloff and E. F. Nelson: Trans. Amer. Inst.
Chern. Eng. (1936) 32, 221.
T. W. Johnson and W. B. Bermald: U. S. Bur.
Mines Tech. Pub. 539 (1932).
D. L. Katz: Trans. A.I.M.E. (I938) 127, I59.
D. L. Katz and C. C. Singleterry: Trans. A.I.M.E.

(I939) 132, 103.

7. W. B. Kay: Ind. and Eng. Chern. (I936) 28,
IOI4:--

8. B. H. Sage and V\f. N. Lacey: Amer. Petro Inst.

Drill. and Prod! Practice (I939) 641.
9. B. H. Sage and H. H. Reamer: Trans. A.I.M.E.

(I94I) 142,179.
10. R. L. Smith and K. M. Watson: Ind. and Eng.
Chern. (I937) 29, I408.