Journal of Petroleum Science and Engineering, 2 (1989) 1-12 1 Elsevier Science Publishers B.V., Amsterdam -- Printed in The Netherlands

FAST CRUDE-OIL HEAVY-COMPONENT CHARACTERIZATION USING COMBINATION OF ASTM, HPLC, AND GPC METHODS

KOSTA J. LEONTARITIS and G. ALI MANSOORI*

Department of Chemical Engineering, Chemical Engineering Building, University of lUinois, P.O. Box 4348, Chicago, IL 60680 (U.S.A.)

(Received March 30, 1987; revised and accepted March 12, 1988)

Abstract

Leontaritis, K.J. and Mansoori, G.A., 1989. Fast crude-oil heavy-component characterization using combination of ASTM, HPLC, and GPC methods. J. Pet. Sci. Eng., 2: 1-12.

Resins and asphaltenes are considered important culprits of the asphaltene deposition problem and understanding their phase behavior is crucial to its solution. Predictive models of asphaltene deposition would normally require as input a minimum number of physical, chemical, and thermodynamic properties of the resins and asphaltenes.

The main objective of the paper is to propose an improved technique for quick determination of the resin (i.e., asphaltene peptizing agent) content of a hydrocarbon mixture that simultaneously separates the resins for further analysis. The technique is a combination of ASTM standard and HPLC methods. The results of this technique were checked satisfactorily against those of another equivalent conventional method, which is more time consuming and less reproducible. Furthermore, molecular size distributions of asphaltenes and resins were determined for two crude oils of widely different asphaltene and resin content and place of origin using GPC. A relationship between the asphaltene-number average molecular weight (relative to polystyrene) and the flocculant paraffin carbon number is suggested.

Introduction

Asphaltene deposition is the cause of major economic and safety problems during produc- tion and processing of both natural and syn- thetic hydrocarbons as reported by various re- searchers and engineers working on the subject (Katz and Beu, 1945; Haskett and Tartera, 1965; Lichaa, 1977; Adialalis, 1982; Tuttle, 1983). A detailed discussion of the field expe- riences and laboratory research related to the asphaltene deposition problem was presented by the authors earlier (Leontaritis and Man- soori, 1988). Also earlier a thermodynamic-

*Author to whom correspondence must be addressed.

colloidal model was proposed (Leontaritis and Mansoori, 1987) for predicting asphaltene flocculation during petroleum production and processing.

Experimental data suggest that asphaltenes and resins are very important components of the crude oil because they play a crucial role in the asphaltene deposition problem. Yet, not enough is known about their physical and chemical properties, and especially thermody- namic properties required by mathematical models predicting the phenomenon of asphal- tene flocculation. In the model mentioned above (Leontaritis and Mansoori, 1987) resins are a component of the asphaltene-free crude oil re- sponsible for peptizing the asphaltenes (see Fig.

0920-4105/89/$03.50 © 1989 Elsevier Science Publishers B.V.

9, to be discussed later). As a result they are included in the phase calculations of the oil at different state conditions. For these calcula- tions various properties of each oil constituent, such as critical point data and molecular weight, are necessary.

The experimental work presented here is geared towards determining the above proper- ties of the asphaltenes and resins. It should be kept in mind that the fraction resins in this pa- per refers to those compounds of the crude oil that are responsible for keeping the asphal- tenes in colloidal suspension. Henceforth, the terms resin and asphaltene peptizing agent are used interchangeably in this paper. The results of the work presented here by no means are all inclusive and/or conclusive. However, they suggest how an improved fast technique using a combination of ASTM standard, HPLC, and GPC methods may be used for determination of oil resin content, molecular weight distribu- tions, and other properties of resins and as- phaltenes. The main objective of the paper is to show how a relatively simple ASTM/HPLC hydrocarbon group separation technique can be used to determine the resin or asphaltene pep- tizer content of a crude. Also, since the pro- posed method separates the resin fraction from the rest of the crude it can be used as a sepa- ration step for methods seeking to determine the properties of resins. Furthermore, experi- mental work is presented correlating the num- ber average molecular weight of asphaltenes of a certain crude oil with the paraffin flocculant carbon number.

Experimental procedures

Conventional methods for characterizing as- phaltenes and resins (i.e., finding their physi- cal, chemical, and thermodynamic properties) consist primarily of physically separating these compounds from the original oil and then sub- jecting them to different analyses using a host of available analytical techniques. The sepa- ration step is apparently the most difficult and

TABLE 1

Crude oil properties a

Property Brookhaven Oil DK-107 Oil

Appearance dark black brown

Density at 25°C 0.883 0.863 at 15.56°C 0.892 0.875

API gravity 27.13 30.21

Surface tension (dynes/cm), 25°C 24.89 23.29

Viscosity (centipoise), 25°C 41.24 9.53

aFrom Vuong ( 1985 ).

error-prone in these methods. In the technique presented here the separation of resins is ac- complished by HPLC and the results are sat- isfactorily reproducible, as many researchers using the HPLC have shown (Jewell et al., 1972; Sautoni and Swab, 1975; Drushel, 1978; Dark, 1982). Two asphaltenic crude oils were used in these studies, the heavier 27.13 API gravity Brookhaven oil and the lighter 30.21 API grav- ity DK-107 oil. Crude oil physical properties are shown on Table 1. The experimental methods for asphaltenes and resins are discussed below.

Asphaltenes

The asphaltenes were separated from the crude oil using the method ASTM D-3279. This method does not prescribe the micron rating of the filtration apparatus needed. A 0.1/~m cel- lulosic membrane filter was used in these ex- periments. The dried asphaltenes were dis- solved in tetrahydrofuran (THF) to make a 0.1% by weight solution. 100/zl injections were made into the GPC to obtain the asphaltene distributions. Figure 1 shows a flow chart for the gel permeation chromatogram which is manufactured by Waters. Three equal size (3.9 mm × 30cm) ultrastyragel columns in series of

Solvent Reservoir

RI Detector

l I Data Processor 1

Fig. I. GPC flow chart.

61 Columns r I

Disposal

500 A, 10 ~ ~, and 103 .~ rating were used to ob- tain the distributions.

Resins

The resins were separated from the crude us- ing two different methods, one using "adsorp- tion" and another using HPLC fractionation. The term "adsorption" is used here to desig- nate the entire resin separation method as dis- cussed below.

Resin separation by "adsorption" This procedure is similar to the one dis-

cussed by Sachanen (1945) and Koots and Speight (1975). It mainly consists of passing asphaltene-free oil through a bed of a common adsorbent, e.g., fuller's earth, to separate the resins from the oil. A light paraffin, e.g., n-pen- tane, is then passed through the bed to prefer- entially desorb any non-resins that were ad- sorbed in the bed. In turn, a strong desorbent, e.g., chloroform, is used to desorb the resins from the bed. The resins are recovered after evaporation of the desorbent. The separated resins were dissolved in tetahydrofuran (THF)

to make a 0.1% by weight solution. 100/~1 in- jections were made into the GPC to obtain the resin molecular size distributions. The column set-up described above for the asphaltenes was also used for the resins.

Resin separation by HPLC [ractionation The procedure described above is laborious,

time consuming, and has high risk of inaccu- racy. However, there exist other methods (Jewell et al., 1972; Sautoni and Swab, 1975; Drushel, 1978; Dark, 1982) for splitting the crude oil into four fractions, namely, the as- phaltenes, paraffins, aromatics, and polars. The method by Dark (1982) or the "literature" method (as it will be called from now on), as well as some other methods, is similar to the recommended ASTM clay-gel chromato- graphic method for crude oil hydrocarbon group separation, the ASTM D-2007. An important difference between the two methods is that the literature method takes, at the most, one hour and the results are easily reproducible, whereas the ASTM method may take anywhere be- tween 2 to 4 days.

The key oil fraction here that is of major in- terest is the "polar" fraction. Another name given to this fraction in the literature is "het- eroatoms" (Sachanen (1945) and Hirschberg et al. (1984)). By the description of each method it is evident that the words "resins", "polars", and "heteroatoms" refer to very sim- ilar but not necessarily identical compounds. In our terminology, as mentioned earlier, res- ins represent those compounds of the crude oil that are responsible for peptizing the asphal- tenes and keep them in suspension. The pur- pose of the proposed method is to effect sepa- ration of these peptizing agents from the crude oil and make them available for further analy- sis, i.e., for determination of critical properties, molecular weight, gravity, etc. However, fur- ther improvements of the technique are possi- ble and these are discussed later.

The HPLC method consists of basically three steps: ( 1 ) a 0.5 g sample of crude oil is dissolved

Solvent Reservoir

r ,

..... i

Disposal

Fig. 2. HPLC flow chart.

in 20 ml of n-C6; (2) the mixture is then filtered to remove the precipitated asphaltenes; and (3) the hexane solubles are run through the HPLC. The paraffins elute first and are detected by a refractive index detector. The aromatics elute next and the polars elute after the column is backflushed and they are both detected by a UV absorbance detector.

This method gives the composition of the crude in terms of the above-mentioned four fractions and also separates each group for fur- ther analysis. Figure 2 shows a flow chart for the high performance liquid chromatography system used, most of which is manufactured by Waters. The column used was an energy anal- ysis (NH2) column (3.9 mm×30 cm). Sample concentration was 0.5 g per 20 ml n-hexane and dosage at 10 ml. The carrier solvent was hexane at a flow rate of 2 ml/min.

Resul t s and d i scuss ion

The results of the analysis are separated into two categories; asphaltenes and resins.

Asphaltene results

The asphaltenes were flocculated from Brookhaven oil and DK-107 oil using the fol- lowing six paraffin flocculants: n-pentane; n- hexane; n-heptane; n-decane; n-dodecane; and n-hexadecane. Figures 3 and 4 show the rela- tionship of the amount of asphaltenes floccu- lated versus the flocculant carbon number for each crude oil. The typical drop of the floccu- lated amount of asphaltenes as the carbon number increases is clearly evident from these plots. In fact if butane or lighter paraffins had been used the flocculated amounts would be even higher. This has been discussed by a num- ber of researchers on the subject (Yen, 1972; McKay et al., 1978; Boduszynski et al., 1980; Speight, 1981; Mansoori, 1988). The explana- tion given for this phenomenon is that more and more resins are flocculating together with the asphaltenes, as the paraffin flocculant be- comes lighter. Consequently, a pertinent ques- tion is which paraffin separates only and all the asphaltenes? The answer is probably none, even though from our preliminary data indications are that n-pentane comes close. That is why an efficient and accurate separation method for both asphaltenes and peptizing agents is required.

Samples of n-pentane, n-hexnae, and n-hep- tane asphaltenes flocculated from both crude oils were prepared, as described previously, and run through a Waters GPC. The particle size distributions of all the samples were deter- mined and are shown in Figs. 5 and 6. Samples of the n-decane, n-dodecane, and n-hexade- cane asphaltenes were not run through the GPC because of experimental difficulties in the oven drying purification step. The asphaltene par- ticle size distributions if compared to each other show a trend that is a function of the flocculant carbon number. The n-C5 asphaltenes have more of the larger particles than the n-CG as- phaltenes which in turn have more than the n- C7 asphaltenes for both crude oils. It should be

1.2-

1 , 8 -

B

1.6-

1.4

1.2-

1.0-

0.8 • s

0.6

0.4 | i w

4 6 8 10 112 14 116 118

Paraffin Carbon Number

Fig. 3. Brookhaven oil asphaltene contents.

1.0

0.4-

0.8

o.s.

0.2 4

| |

6 8

[ ]

|

10

Paraffin Carbon Number

Fig. 4. DK-107 oil asphaltene contents.

w

i

1 2

!

1 4

,¢:

3=

6 '

4,

2 '

'~ Pentane

• Hexane

" Heptane

O'

11 12 13 14

Retention Tlrrm

Fig. 5. Molecular weight distributions of Brookhaven oil asphaltenes.

15 16 17

.¢

2

,~ Pentane

• Hexane

u Heptane

Retention Time

Fig. 6. Molecular weight distributions of DK-107 oil asphaltenes.

0 ' 10 15 20 25

u

J o

3.3 =

3.2.

3.1

3.0 4

L,, DK-107

• Brookhaven

5 6 7 8

Fig. 7. Brookhaven and DK-107 asphaltenes.

Carbon Number

noted that in GPC the heavier compounds elute first.

Figure 7 shows a plot of the average molecu- lar weight of the asphaltenes, as determined by GPC, versus the carbon number of the paraffin flocculant. These molecular weights were de- termined relative to narrow dispersity polysty- rene standards and no claim can be made as to whether they are "absolutely" true or not. However, Fig 7 demonstrates a functional re- lationship which appears to be true for both crudes, namely, the logarithmic average molec- ular weight of the asphaltenes is linearly re- lated to the carbon number of the paraffin floc- culant. These experimental points were fitted to a polynomial of first degree and the following equations were developed:

Brookhaven asphaltenes:

loglo (MWa~ph.)

---- 3.561 -- 0.0640 × (Carbon No. ) (1)

D K-107 asphaltenes:

loglo (MWa~ph.)

=3.465--0.065× (Carbon No.) (2)

These two linear equations have nearly iden- tical slopes. The usefulness of these equations is still to be shown, as research on the subject continues. The main thrust of the research will be to collect more data with higher carbon number paraffins to verify the validity of the above equations.

Resin results

The Brookhaven oil resins were run through the GPC and their molecular weight size dis- tribution was generated as shown in Fig, 8. This is a s tandard normal distribution and, as a re- sult, it indicates that, with respect to size, res- ins distribute evenly around their mean in the crude oil. Their number average molecular weight (relative to polystyrene) is ~ 430 + 30.

Table 2 presents a summary of the hydrocar- bon group separation analysis data for both

4

a

.D 2 ,

o a.

[]

El

0

E

E

E

[ ]

0

E

BB [] 13

[] m la

[] []

0 | ! !

4 5 6 7

Resin LogMolecular Weight

Fig. 8. Molecular weight size distribution of Brookhaven oil resins.

[ ]

[ ]

[ ]

m

E

m

[ ]

[ ]

m m

! !

8 9

TABLE 2

Hydrocarbon group separation by HPLC method

Component Brookhaven 0il DK-107 Oil

Asphaltenes (wt%) 1.25 a _+ 0.03 b 0.53 _+ 0.03 Saturates (wt%) 70.57 _+ 0.42 47.00 _+ 0.28 Aromatics (wt%) 22.94 _+ 0.33 43.22 + 0.17 Polars (wt%) 5.24 + 0.06 9.25 _+ 0.08

aA minimum of three measurements for each data point was made to determine the precision, which was taken as the average distance from the mean. bThese are n-C6 asphaltenes.

Brookhaven oil and DK-" 107 oil using the HPLC method. The precision of the data was computed by using at least three experimental measurements for finding the arithmetic mean and then calculating the average distance from the arithmetic mean. The results of the modi- fied HPLC method for determining resin con- tent are compared with those of the adsorption method on Table 3. The same comments for computing precision mentioned above apply here as well.

TABLE 3

Comparison of crude oil resin (or asphaltene peptizer) content by "adsorption" method and HPLC method

Crude oil Resin Potars content Resin content by by HPLC content" by adsorption (wt%) HPLC (wt%) (wt%)

Brookhaven 4.59 +0.12 b 5.24_+0.06 4.84_+0.23 DK-107 8.30_+ 0.27 9.25 + 0:08 8.73_+ 0.14

"The resin content by HPLC is obtained by deducting the difference of the n-C5 and n-Cs asphaltenes from the polars content by HPLC: The n-Cs asphaltenes are 1.65_+ 0.05 and 1.05 _+ 0.03 wt% respectively for Brookhaven and DK-107 oils. The n-Cs asphaltenes are given on Table 2. bA minimum of three measurements for each data point was made to determine the precision, which was taken as the average distance from the mean.

Discussion o/results

As mentioned earlier the ultimate objective of this experimental work is to generate data and/or suggest methods for obtaining the physico-chemical, and especially thermody- namic, properties of asphaltenes and resins.

Effective and consistent separation of asphal- tenes and especially peptizing agents from the crude oil for further analysis is a major stum- bling block in the characterization process.

Figure 9 depicts how the peptization of as- phaltenes, which are considered particles col- loidally suspended in crude oil, is effected by the "adsorbed" resins. This process is the basis of the thermodynamic-colloidal model pro- posed earlier by Leontaritis and Mansoori, 1987. A number of experimenters {Swanson, 1942; Katz and Beu, 1945; Yen, 1972; Lichaa, 1977; Speight, 1981) demonstrated that once the adsorption equilibrium of the resins be- tween the solid and liquid phases is disturbed, i.e., by adding a paraffin solvent, the asphal- tene particles flocculate irreversibly. This means that, in order for the asphaltenes to floc- culate, they grow in size and as a result their apparent molecular weight changes. If the floc- culated asphaltenes are redissolved in a good solvent, i.e., benzene or toluene, after some time passes they begin to grow in size again. How- ever, if the asphaltenes are redissolved in ben-

I. C

C~

J

Legend :

O resin molecules

• aromatic molecules

ell molecules of different size and parafflnic nature

asphaltene particles of different sizes end shapes

Fig. 9. Physical model of asphaltene peptization; the as- phaltenes are surrounded by the peptizing agents that keep them in suspension.

zene (or another appropriate solvent) that contains enough resins for complete peptiza- tion then the particles remain permanently peptized and do not grow in size (Swanson, 1942). This suggests that any experimental work that is destined for characterizing as- phaltenes as they exist in the original oil is bet- ter to be performed while the asphaltene par- ticles are still peptized as Fig. 9 depicts, if the data are to correspond to the original asphal- tene particles.

All the experimental work presented in this paper regarding the asphaltenes has been per- formed after the asphaltenes were separated from the original crude oil and, as a result, the data may not necessarily apply to the original asphaltenes. The determination of the original asphaltene particle properties has been ex- tremely difficult. Yen (1972) and Boduszynski et al. (1980) interpreted some of their data to apply to asphaltenes in their original non-ag- gregated state. The model proposed earlier by Leontaritis and Mansoori (1987) does not re- quire any asphaltene properties and thus the problem of acquiring true asphaltene proper- ties is bypassed. In the literature there have been proposed models (e.g., Hirschberg et al., 1984) that require input of asphaltene properties.

In hydocarbon mixtures (as in most mix- tures), size difference between the components is a major factor affecting their solubility. In this case, resins are expected to be more soluble in heavier than lighter paraffins. As a result, it is expected to require smaller amounts of the heavier flocculants for dissolving the same amount of resins. Consequently, the required amount of flocculant for arriving at the onset will be inversely proportional to the flocculant molecular weight, provided that the amount of resins that needs to be desorbed from the as- phaltenes in order to arrive at the onset is con- stant regardless of the flocculant used (refer to Fig. 9). This trend has been observed experi- mentally (e.g., Hirshberg et al., 1984).

Furthermore, the amount of asphaltenes

10

flocculated with a given flocculant is affected by the amount of resins left in the asphaltene phase, i.e., left adsorbed on the asphaltenes after flocculation (Speight, 1981). Figures 3 and 4 show the typical drop reported for the amount of flocculated asphaltenes versus the carbon number (or molecular weight) of the paraffin floccutant for each of the two crudes. This drop corroborates the statement made in the above paragraph of the increase in solubility between resins and heavier paraffins, i.e., for the same amount of paraffin flocculant added more res- ins will be removed from the asphaltenes by the heavier paraffins. Hence, the amount of as- phaltene precipitate will decrease as the par- affin molecular weight increases.

The asphaltene size distributions deter- mined by GPC, shown on Figs. 5 and 6, are the ones the asphaltenes have in tetrahydrofuran (THF) solvent, and as a result they do not con- tradict the discussion made in the previous two paragraphs. These distributions or apparent molecular weights are a function of the GPC carrier solvent used (in this case THF), among other things. The fact that the distributions are different for each flocculant solvent indicates that the asphaltene flocculation process is af- fected by the flocculant. In other words, the strength of the bonds formed during floccula- tion relative to the THF solvent is a function of the flocculant solvent (i.e., n-Cs,n-C6,n-CT). It is even more remarkable that the logarithmic number average molecular weight (relative to polystyrene) of these distributions is linearly related to the carbon number of the flocculant.

At this time no definite conclusions may be drawn about the experimental trend described above mainly because the number of experi- mental points is small (only three paraffin flocculants); more experimental points with heavier paraffins are needed. However, during the above experiments it was observed that, in the filtering step of the flocculated asphaltenes, plugging of the membrane filter becomes more severe (i.e., the required filtration time of the paraffin-crude mixture becomes larger for the

same amount of sample ) as the carbon number of the paraffin flocculant increases. This indi- cates that the original asphattenes aggregate in smaller size particles as the paraffin molecule size increases. It seems that because the larger paraffin solvent molecules tend to hinder tight packing of the original asphaltene particles, the asphaltene aggregates due to these larger sol- vent molecules contain a larger amount of the original, smaller particles. As a result, the ap- parent molecular weights of asphaltenes floc- culated by larger paraffins are expected to be smaller than those flocculated by smaller par- affins. This is supported by the trend in the dis- tribution curves of Figs. 5 and 6 and the mea- sured average molecular weights.

The resin situation is different because the resin molecules are in true solution in the orig- inal oil and so the problem of aggregation does not exist. Thus the resins can be separated from the original oil and subjected to reasonably se- vere experimental analysis without any alter- ation to their molecular size and structure. This is a bonus from nature for, in certain situa- tions, accurate resin thermodynamic proper- ties are needed for predicting asphaltene floc- culation (Leontaritis and Mansoori, 1987).

Table 3 indicates that, taking into account the experimental error involved in determining the resin content of a crude by the conventional adsorption method and the error involved in determining the polars content of a crude by the HPLC method, the fraction called "resins" becomes nearly equal to the fraction called "po- lars" if the difference between the amount of n- C~ and n-C6 asphaltenes is deducted from the amount of polars. This is not entirely unex- pected because any asphaltenes that are left be- hind in the oil by the HPLC hydrocarbon group separation method (during the asphaltene flocculation step using hexane ) will be retained in the column and afterwards elute with the resins in the backflush mode to be counted as polars.

Based on the results of Table 3, it may be concluded that it is possible to determine the

resin content of a crude, as defined by the tra- ditional adsorption method, through use of HPLC. One such method suggested by the re- sults of Table 3 would involve the following steps: (a) determine the n-C5 and n-C6 asphal- tene contents of the crude using the ASTM D- 3279 standard method; (b) use the HPLC method to find the polar content of the crude; and (c) deduct the difference of the n-C5 and n-Ce asphaltenes from the polar content to de- termine the resin content of the crude. It should be noted that the method proposed above for determining the resin content of a crude is based on the results of only two crudes. The credibil- ity of the method will be enhanced after more crudes are analyzed with the same satisfactory results. A possible improvement to the above method would be to develop an asphaltene sep- aration method that determines accurately the "total" asphaltene fraction of the crude instead of assuming that n-pentane asphaltenes are the total fraction. For instance, that might be ac- complished by successive fractionation/sepa- ration of asphaltenes using a combination of lighter and heavier paraffins or even other as- phaltene flocculants. The n-C6 asphaltene con- tent would be deducted from the true total as- phaltene content, and the result would subsequently be deducted from the polars con- tent, as determined by the HPLC method, to yield the resin or asphaltene peptizer content.

The HPLC method is more precise and faster than the adsorption method. Also the possibil- ity of losses using the HPLC method is nil. To the contrary, the losses of the adsorption method can be substantial during the adsorp- t ion-desorption steps. The fact that the resin content of each crude, as determined by ad- sorption, is consistently lower than that deter- mined by HPLC, see Table 3, indicates that some losses may be incurred when using the ad- sorption method.

Conclusions

A simple, rapid, and precise technique for de- termining the resin or asphaltene peptizer con-

11

tent of a crude was proposed. This technique involves high performance liquid chromato- graphy and ASTM standard methods and was satisfactorily tested against a conventional la- borious "adsorption" method. The proposed method may be used as a resin separation step for methods seeking to do further analysis of the resin fraction. Further possible improve- ments of the proposed method were also suggested.

A functional relationship was demonstrated between the asphaltene-number average mo- lecular weight ( relative to polystyrene) and the carbon number of the paraffin flocculant used to separate the asphaltenes from the crude. This indicates that the paraffin solvent takes part in or influences the asphaltene aggregation process.

Acknowledgements

This research is supported in part by the Na- tional Science Foundation Grant CBT-8706655 and in part by the Shell Oil Company.

References

Adialalis, S., 1982. Investigation of physical and chemical criteria as related to the prevention of asphalt deposi- tion in oil well tubings. M.Sc. thesis, Pet. Eng. Dep., Imperial College Univ. London, London.

American Society for Testing and Material. Method ASTM D 3279: standard test method for normal heptane insolubles.

American Society for Testing and Material. Standard method ASTM D-2007: clay-gel adsorption chromato- graphic method for hydrocarbon group analysis.

Boduszynski, M.M., McKay, J.F. and Latham, T.M., 1980. Asphaltenes, where are you? Proc. Assoc. Asphalt Pav- ing Tech. Tech. Sess., Louisville, Ky., 49: 23-143.

Dark, W.A., 1982. Crude oil hydrocarbon group separation quantization. J. Liq. Chromatogr., 5 (9): 1645-1652.

Drushel, H.V., 1978. HPLC techniques for analysis of re- sidual fractions. Adv. Chem. Set., No. 170, Washington, D.C., 1981, p. 17.

Haskett, C.E. and Tartera, M., 1965. A practical solution to the problem of asphaltene deposits - - Hassi Mes- saoud Field, Algeria. J. Pet. Technol., {April), pp. 387- 391.

Hirschberg, A., De Jong, L.N.J., Schipper, B.A. and Mei-

12

jers, J.G., 1984, Influence of temperature and pressure on asphaltene flocculation. Soc. Pet. Eng. J., 24 (3): 283-293.

Jewell, D.M., Albaugh, E.W., Davis, B.E. and Ruberto, R.G., 1972. Am. Chem: Soc., Div. Pet. Chem., Prepr., 17 (4): F81.

Katz, D.H. and Beu, K.E., 1945. Nature of asphaltic sub- stances. Ind. Eng. Chem., 37: 195-200.

Koots, J.A. and Speight, J.C., 1975. Relation of petroleum resins to asphaltenes. Fuel, 54: 179-186.

Lichaa, P.M., 1977. Asphaltene deposition problem in Venezuela crudes - - usage of asphaltenes in emulsion stability. J. Can. Pet. Technol., (June), Oil Sands, pp. 609-624.

Leontaritis, K.J. and Mansoori, G.A., 1987. Asphaltene flocculation during oil production and processing: a thermodynamic-colloidal model. Proe. Soc. Pet. Eng., Int. Syrup. Oilfield Chem., San Antonio, Tex., Pap. No. SPE 16258, pp. 149-158.

Leontaritis, K.J. and Mansoori, G.A., 1988. Asphaltene de- position: a survey of field experiences and research a p proaches. J. Pet. Sei. Eng., 1 (3): 229-239.

McKay, J.F., Amend, P.J., Cogswell, T.E., Harnsberger, P.M., Erickson, R.B. and Latham, D.R., 1978. Petro- leum asphaltenes: chemistry and composition. Adv.

Chem. Ser., No. 170; Washington, D.C., 1981; p. 129. Sachanen, A.N., 1945. The Chemical Constituents of Pe-

troleum. Reinhold Publ. Corp., 385 pp. Sautoni, J.C. and Swab, R.E:, 1975. Rapid hydrocarbon

group type analysis by high performance liquid chro- matography, J. Chromatogr. Sci., 13: 361-368.

Speight, J.G., 1981. Solvent effects in the molecular weights of petroleum asphaltenes. Symp. Characterization of Heavy Ends, Am. Chem. Soc., New York meeting, Au- gust 23-28, pp. 825-830.

Stevenson, R.J., 1971. J. Chromatogr. Sci., 9: 257. Swanson, J., 1942. A contribution to the physical chemis-

try of the asphalts. J. Phys. Chem., 46: 141-150. Tuttle, R.N., 1983. High pour-point and asphaltic crude

oils and condensates. J. Pet. Technot., (June), pp. 1192- 1196.

Vuong, L.T., 1985. Asphaltene deposition and its role in enhanced oil recovery (EOR) miscible gas flooding pro- cesses. M. Sc. thesis, Dep. Chem. Eng., Univ. Illinois, Chicago, Ill., October, 1985.

Yen, T.F., 1972. Present status of the structure of petro- leum heavy ends and its significance to various techni- cal applications. Am. Chem. Soc., Div. Pet. Chem., Prepr., 17 (4): 102.