JKAU: Eng. Sci., vol. 7, pp. 81-92 (1415 A.H. / ]995 A.D.

An Empirical Correlation for the Solubility of Paraffin

Waxes in Base Oils

MALIK ALAHMAD, TARIQ F. AL-FARISS,

MOHAMAD E. ABASHAR and HISHAM M. ETIOUNEY
Chemical Engineering Department, College of Engineering,
King Saud University, Riyadh, Saudi Arabia

ABSTRACf. An empirical correlation was developed for the prediction of

solubility of three types of paraffin waxes in various base oils. The correla-
tion was developed using the superposition approach and was proposed as
A
Ts = e ( B In C + D)

The range of melting points for the three waxes was 43 to 60°C which corres-
pond to molecular weight range of 336 to 441.

Development of the empirical correlation and the experimental data was

performed as a function of: Type of wax and base oil; the concentration of
the wax in the base oil; and solubility temperature.

The developed correlation shows satisfactory agreement with the experi-

mental data with a average deviation of 5.9 % and a maximum deviation of
17.9%.

1. Introduction
The deposition of paraffin waxes in crude oils and petroleum products causes severe
problems in pipelining and storage. These and other problems are described by many
researchers (Shock et ale [1], EnDean[2], Newberry[3 f, AI-Ahmad et all [4] and Siffer-
man[5]. Determination of the solubility temperatures of paraffin waxes in crude oils
and petroleum products are extremely essential in order to obtain the proper operat-
ing conditions to control the paraffin wax deposition for efficient and economical
pumping and storage. Extensive research has been carried out to investigate the sol-
ubility behaviour of paraffin wax in pure solvents (hydrocarbons) and defined mix-

81
H2 Malik Alahmad et al.

tures (AI-Ahmad et ale (4 1). However, only a limited amount of work has been pub-
lished on the solubility of waxes in undefine mixture (Berne and Work[6], Nathan[7],
Gudmundsson and Bott[8 1).
Solubility of paraffin wax (melting point of 43°C) in base oils with molecular
weight range between 317 - 447 was studied by AI-Ahmad et ale [4]. Measurements
were made with wax concentrations up to 30 wt% . The study resulted in an empirical
correlation of the prediction of the paraffin wax solubility. Further analysis was
made by AI-Ahmad et ale [9] to cover higher wax concentration (up to 50 % ). Two re-
gimes (lower and higher than 15 wt% wax solutions) were found for solubility data
where two empirical correlations have been developed.
The solubility of refined paraffin waxes in petroleum fractions was studied by
Berne and Work[6]. They concluded that the solubility of any paraffin wax in any non-
aromatic saturated petroleum could be determined from their empirical correlation.
Further analysis of this problem was presented by Nathan[7], where solubility mea-
surements were performed for five paraffinic waxes with melting of (66.5 - 84.5°C)
in a number of solvents (including, hexane, heptane, mineral oil, crude oil (MW
220), kerosene, ... etc.). He found that solubilities of the waxes in hydrocarbon sol-
vents can be calculated using ideal solubility relations, however, in chlorinated and
oxygenated solvents large deviations from ideality was found (due to nonideal be-
haviour of the solvent).
The study Gudmundsson et ale [8] focused on measuring of a paraffin wax solubility
in a commercial kerosene. They performed comparison between data and the Berne
and Work correlation[6] and the ideal solubility equation used by Nathan[7]. They
concluded that perfect fit between the data and both correlation and the ideal solu-
bility equation was found for wax concentration less than 10 wt%. However, for wax
concentrations between 10 and 50 wt%, the theoretical equation gave more accurate
results.
Wardhaugh and Boger[lO] show that the crystallization of wax in crude oils produce
non-Newtonian characteristics.

McClafin and Whitfil[llj show that paraffin deposition can be prevented or greatly
12
retarded by using chemical surfactants. AI-Fariss[ 1 studied the parameters which af-
fect the formation and strength of the gel structure in pi~elines. Study of flow im-
fS
provers for transportation was done by Van Engelen et ale 13], Sifferman ]. Burger et
ale [14] stated that wax deposition is due to lateral transport by diffusiop., shear disper-
sion and Brownian diffusion.

Based on previous studies[4,9] it was found essential to continue the work with var-
ious paraffin waxes having different melting points.
The objective of this study is to develop a generalized empirical correlation to de-
termine the solubility of paraffin waxes in base oils which takes account of the type of
wax as well as type of oil.
 An Empirical Correlation for the Solubility ... 83

2. Experimental Apparatus
The three paraffin waxes used in this study have melting points of 43, 56 and 60°C
which correspond to average molecular weights of 336, 386 and 441 respectively
(measured by vapour pressure osmometry according to ASTM-D2503). The waxes
were referred to as wax 1, 2 and 3 respectively. The wax solubility was measured in
three base oils having molecular weights of 317, 346 and 447, which were referred to
as oil 1, 2 and 3 respectively. Physical properties of the three oils are given in Table 1.

TABLE 1. Properties of base oils *.

Test Oil2 Oil3

Property Oil 1
procedure

Molecular weight, gmlgm mole 317 346 447

Color ASTM-D1500 1.5 3.0 4.5

Pour point, °C (max.) ASTM-D97 -18 -9 -6

Flash point (COC °C, min) ASTM-D92 199 218 232

Mean average boiling point, °C 351 385 455

Viscosity at 40°C ASTM-D445 18.5/21 55/61 128/139

Viscosity index 100 300 700

Density, gm / cm 3 0.84 0.88 0.89

·Obtained from Riyadh Refinery.

To investigate the solubility behaviour of various waxes in the base oils, the cloud
point was determined using cloud point test cabinet (Fig. 1). The measurement pro-
cedure was performed according to ASTM-D97 ~IP15). Details of experimental set
up and procedure were given by AI-Ahmad et ale 4]. The performed experiments in-
cluded nine sets, which include the solubility of the three waxes in the three base oils.
In each set wide range of wax concentrations were investigated, i.e., from 5 to 30
wt% .

3. Development of the Proposed Correlation

The parameters included for the analysis of solubility data were: wax concentra-
tion and its molecular weight, and oil molecular weight. Superposition technique was
implemented for the development of the correlation(15] and regression analysis was
used to estimate the values of the constants in each step, leading finally to the follow-
ing form:
Ts = eA ( B In C + D)
where: Ts is the solubility temperature in °C,
A = 0.434 M o - 0.752 M w'
84 Malik Alahmad et al.

Thermomete r s

Corlt

Test Jar
Jacket

Felt Gasket

Felt Disk

Section of Analysis
Test Cabinet

FIG. 1. Experimental set-up for solubility temperature determination (ASTM-D97).

B 8.0 X 10- 5 Mo 1.738,

D 1.946 X 10- 10 M o 3.632,
C is the wax concentration in wt%,
M o is oil molecular weight,
M w is wax molecular weight,

4. Results and Discussion

Results of solubility measurement are shown in Fig. 2, 3 and 4 for waxes 1, 2 and 3
in base oils, 1, 2 and 3 covering wax concentrations from 5 to 30 wt%.
The results show that for the three oils, the solubility temperature increases with
the increase of the wax concentration and molecular weight (Fig. 2-4). The rate of
change in solubility temperature seems higher in the lower range of concentration.
The rate of change in solubility temperature is particularly small above the melting
point of wax itself. This shows that in the present range of concentration~,the phys-
ical properties of wax dominate those of oil, so that the wax starts to solidify near its
melting point.
 An Empirical Correlation for the Solubility... 85

100.0 . . . - - - - - - - - - - - - - - - - - - - - - .

BASE OIL 1

80.0
-
u
0......,

..e
~
::;,
60.0 c
GI
a.
E
~
~
~ 40.0
:0
::;,
<5
en,
20.0 0

000 WAX 1
000 WAX 2
0.0.0. WAX 3

OD
0.0 5.0 10.0 15.0 200 25.0 30.0 35.0
Wax Concentration (wt ./.)

FIG. 2. Variation in solubility temperature of waxes (1, 2, 3) in oil 1, as function of wax concentration and
type.
100.0 - - - - - - - - - - - - - - - - - - - - .

BASEOIL 2

80.0

E
..
~
~

.a 60.0
~

a.
A

E
~
~

:= 40.0
:0
::;,
0
en
20.0
000 WAX 1
000 WAX 2
A 0.0. WAX 3

OD 1J"J"L.u...L.I~.L.U.J....I.oI.oI.lo ...........I.U.oI.~~...........~~"""""""'............~~~...........1ooI.I.IoI
0.0 5.0 10.0 15.0 200 25.0 35.0
Wax Conc~ntration (wt 0/.)
FIG. 3. Variation in solubility temperature of waxes (1, 2,3) in oil 2, as function of wax concentration and
type.
86 Malik Alahmad eta/.

100.0
BASE OIL 3

80.0
-
.'u"-
...~a
::J
60.0
•
~

0..
E
~
>-
~ 40.0
~
::J
0cJ)

20.0
000 WAX t
000 WAX 2
l::J.6l::J. WAX 3

OD
0.0 5.0 10.0 15.0 2QO 25.0 30.0 35.0

Wax Concentration (wt .,.)

FIG. 4. Variation in solubility temperature of waxes (1, 2, 3) in oil 3, as function of wax concentration and
type.

The solubility temperature of wax increases with the increase of the molecular
weights of both the oil and the wax. This behaviour is clearly displayed in Fig. 2, 3
and 4. For example, for wax 1 at concentration of 20 wt%, the solubility tempera-
tures were 35, 39 and 44°C in oils 1, 2 and 3, respectively. Also for oil 1 at the same
wax concentration, the solubility temperatures were 35, 54 and 61.5°C for waxes 1,2
and 3, respectively. The increase in solubility temperature with the increase in wax or
oil molecular weights might be due to larger molecular volume, where higher temp-
eratures are required in order to increase the molecular spacing[9] .
Assessment of the validity of the developed correlation was achieved by compari-
son of the measured data and the predictions by the correlations. The comparison is
shown in Fig. 2-10.
In Fig. 2-4 the calculated values of solubility temperature as function of wax con-
centration, wax molecular weight and oil molecular weight, represented by solid
curves, are plotted against the measured data. Showing reasonable accuracy of cor-
relation predictions are as shown, obtained over the considered concentration range.
The calculated values of the solubility temperature are also plotted against the
measured data in Fig. 5-7 to show the error scatter around the 45° line.
 An Empirical Correlation for the Solubility ... 87

60.0
WAX 1

6.

U
50.0 6.
0_
6.
~ a
~ 0

E 40.0 0

"E
~
0
0

..." o 0
~

.~ 30.0 0
0

'.0 &
:J
0
VOl 0
0
'tJ 20.0
~"
:J
u
C
u 10.0 000 OIL 1
000 OIL 2
~~~ OIL 3

0.0
0.0 10.0 20.0 30.0 40.0 50.0 60.0
Exp~rim~ntal Solubilit y T~mp~rotu~(oC)

FIG. 5. Experimental and model predictions of variations solubility temperature as a function of wax con-
centration for wax 1 in oils 1, 2 and 3.

70.0
WAX 2 6

60.0 0
~
0
60
o~ 6-

~ 50.0 80
~
0
c 0 0
X itO. 0 ~ o 0
...E" 0 0
?: 6.

''a 30.0
::J
~
VOl

~ 20.0
c 0
"3v
C
u
10.0 000 OIL 1
000 OIL 2
6~~ OIL 3

0.0
0.0 laD 20.0 30.0 40.0 50.0 60D 7GO
Exp~rim~ntQl Solubility T~mp~rQture COC )

FIG. 6. Experimental and model predictions of variations solubility temperature as a function of wax con-
centration for wax 2 in oils 1, 2 and 3.
88 Malik Alahmtld et ai.

90.0
WAX 3
6
80.0 ori'-
0
~
u
e-
70.0 0
O
"~
eI 60.0 •
0 a
E 50.0
~ 0 Q

€
:.a
-,40.0 0
A

:J

~ 30~0
Ja
-; 20.0
~ 000 OIL 1
10.0 aaa OIL 2
AAA OIL 3

0.0
0.0 1QO 20D 300 40.0 50.0 SOD 70.0 80.0 900
Experimental Solubility Temperature CoC )

FIG. 7. Experimental and model predictions of variations solubility temperature as a function of wax con-
centration for wax 3 in oils 1, 2 and 3.

Analysis of data for these figures shows that the average deviation between the
measured and the calculated data is only 5.9% while the maximum deviation is
17.9 % . Meanwhile 82% of the whole data points have less than 100/0 error, which
means satisfactory predictions of the proposed correlation.
Further assessment of the model was performed by comparison with the results ob-
tained from the correlation developed in this work and that developed previously by
AI-Ahmad et ale [4].
The correlation in Ref. [4] was developed for wax 1 (of melting p.oint 42°C) in three
base oils (oils 1,2 and 3). The comparison is limited to wax 1 in the three oils.
The results of this test are shown in Fig. 8-10, where the solubility temperature is
drawn versus the wax concentration. The figures also include the fits of both correla-
tions.

The new model fits the data for oils 2 and 3 with average deviations of 5.9 % , 4.60/0
and 5.2% respectively and with maximum deviation of 12.2% for the whole set of
data points. The old model fits the data for the three oils with average deviations of
3.2 % ,4.20/0 and 5.5% respectively and with maximum deviation of 16.8% for the
whole set of data points. Although the predictions of both models are satisfactory,
and comparable the new model is superior and general, because it includes the effect
of the type of wax, while the old model is limited to single type of wax.
 An Empirical Correlation for the Solubility... 89

45.0
WAX 1
OIL 1
40.0 --- o

35.0
u
~
~ /
::» 30.0 /
/

IE I
/
/0
I

~
I
25.0 I
>. I
I
~ ,I
:0
:J I
20.0 I
~ d
o0 0 Ex perimentGl
_ _ Hew Equation
15.0
____ Old Equation

10.0 u.L"I,.~~""""""""""""~~~u...L.I.i~~~~~~u...L",L"I~u.I
ao 5.0 10.0 15.0 20.0 25.0 30.0 35.0
Wax Concentration (wt .,.)

FIG. 8. Comparison of prediction of solubility temperature as a function of wax concentration for wax 1 in
oil 1 by the new and old l4] models.

45.0
WAX 1
OIL 2

40.0
0
o~
,,-
u 35.0 0/ "
~
~ /
~ o/ I
:J /

IE
30.0 0/
I
I
/
/

~ I
I

~
>.
,I
~ 25.0 I
:J 0/
~
I

000 Experimfntal
20.0 _ _ Hew Equation
____ Old Equation

15.0 ........~.............................,...,..u...I.Ioo~~~L.L."J""L,ju...u..l~~~..L..LL.L..&..LL.L.~
0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0
Wax Concentration (wt .,.)

FIG. 9. Comparison of prediction of solubility temperature as a function of wax concentration for wax 1 in
oil 2 by the new and old l41 models.
90 Malik Alahmad et al.

55.0
WAX 1
OIL 3
50.0
0

45.0 0

u
~
---
~
40.0
i
~
E
35.0
.~
:.c:::J
30.0
~
o0 0 Ex p.rimentGl
_ _ N.w Equation
25.0
____ Old Equation

20.0
0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0
Wax Conctntration (wt -,_)

FIG. 10. Comparison of prediction of solubility temperature as a function of wax concentration for wax 1
in oil 3 by the new and 01d[4] models.

Attempts have been made to test the model with data from literature. When com-
pared with Gudmundsson[8] data, the model gave conservative predictions. This is
because of the difference in solvent used, where Gudmundsson used kerosene.
Other's data were completely unsuitable for comparison, because of lack of experi-
mental data and parameters to fit in our model.

s. Conclusion
A single satisfactory empirical correlation was developed for the determination of
solubility temperature of various paraffin waxes, with wide range of concentrations,
in various base oils. The developed correlation is based on experimental measure-
ment using superposition approach.
The proposed correlation gives an overall average obsolute deviation of S. 9% and
maximum absolute deviation of 17.90/0 from the experimental measurement of solu-
bility for three types of paraffin waxes in three types of solvents (base oils). It can
predict the solubility temperature in terms of wax molecular weight, wax concentra-
tion and solvent molecular weight..

References

[IJ Shock, O.A., Sudbury, J.D. and Crockett,J.J., Studies on the nlechanism of paraffin deposition and
its control. J. Pet. Tech., p. 23 (1955).
 An Empirical Correlation for the Solubility ... 91

[2J EnDean, H., The paraffin problems, parts 1-4, Champion News (1981).
[3J Newberry, M.E., Chemical effects on crude oil pipeline pressure problems, J. Pet. Tech., p .779
(1984).
[4] AI-Ahmad, M., AI-Fariss, T. and Obaid, R., Solubility studies on paraffin wax in base oil, Fuel, 69:
293 (1990).
[5] SitTerman, T.R., Flow properties of difficult- to-handle waxy crude oil,J. Pet. Tech., p. 1042 (1979).
[6] Berne, A.A. and Work, L., Refined paraffin waxes in petroleum fractions, Ind. Eng. Chem., 30(7):
806 (1938).
[7] Nathan, C.C., Solubility studies on high molecular weight paraffin hydrocarbons obtained from pet-
roleum rod waxes, Pet. Trans., p. 204 (1955).
[8] Gudmundsson, J.S. and Bott, T.R., Solubility of paraffin wax in kerosene, Fuel, 56: 1192 (1977).
[9] AI-Ahmad, M., AI-Fariss, T. and Obaid R., Solubility behaviour of a paraffin wax in base oils, Fuel,
72: 895 (1993).
[10J Wardhaugh, L. T. and Boger, D. V., Measurement of the unique flow properties of waxy crude oils,
Chem. Eng. Res. Des., 65:74 (1987).
[11] McClaflin, G .G. and Whitfill, D.L., Control of paraffin deposition in production operations, J. Pet.
Tech., 1965 (1984).
[12] AI-Fariss, T., Effect of wax on oil behaviour, Ind. Chem. Eng., 32(2): 9( 1990).
(13] Van Engelen, G.P., Kaul, C.L. and Aranha, R.P., Study of flow improvers for transportation of
Bomby high crude oil through submarine pipelines, J. Pet. Tech., p. 2539 (1981).
[14] Burger, E.D., Perkins, T.K. and Striegler, J.R., Studies of wax deposition in the Trans Alaska
pipelines, J. Pet. Tech., p. 1076 (1981).
[15] AI-Fariss, T., Viscosity - temperature - shear rate correlation for crude oils and polymers, J. Eng.,
Sci., 14(2) (1988).
92 Malik Alahmad et al.

~.,:J' rU-A J At J.J. J , ~JUJ' JJu, J , J.,I-~' ~l.

~Y'-" ~, ~~ '" ~~, ~ '" ~~, ~~, ~

~~~, ~rJ'~' - ~~)'

~~, ~, ~ t.!Y t ~~ ~4J.) ~ ~L..) ~.b~ J"~ ~ • ~\

~lAH jA:J1 ~ ,'~4 ~.)WI J"~ ~ -!iJ . ~L.,\11 ~~)I ~ ~~ ~

JWI ~I ~ (Superposition)
Ts = eA (b In C + D)
~:r:- ~j} otits::.. ~ ~.J.),. ~ tf ~ ~~J.) ~,.) ~ J:UU ~.bW'J
~)' ~J ~, tj ~yi. :4 .a,'->..:-1.' ~_bW,-, . t t' ~ ,.,,.,, ~ Gtr.!
~ ~.bWI 4;;lA... ~ -!iJ . ~, ~4J.)J ~)' J ~, j:S; ~..lSJ ' ~L., \1,
. ~Ll4 'V, ~ .JA tk> ~~ ~Ll4 0 ,~ tk4-\ .1a-,p ~lS'J ' ~I ~l::J1