Unsaturated hydrocarbons in crude oils

Unsaturated hydrocarbons in crude oils

Org. Geochem. Vol. 21, No. 2, pp. 189-208, 1994 Elsevier Science Ltd Printed in Great Britain 0146-6380/94 $6.00 + 0.00 Pergamon Unsaturated hydroca...

1MB Sizes 0 Downloads 13 Views

Recommend Documents

No documents
Org. Geochem. Vol. 21, No. 2, pp. 189-208, 1994 Elsevier Science Ltd Printed in Great Britain 0146-6380/94 $6.00 + 0.00

Pergamon

Unsaturated hydrocarbons in crude oils E. B. FROLOVand M. B. SMIRNOV Institute of Petrochemical Synthesis, Russian Academy of Sciences, Moscow B-71, Leninski prospect 29, Russia (Received 2 June 1992; returned for revision 15 September 1992; accepted in revisedform 20 May 1993) Abstract--The widespread occurrence of olefin-containing crude oils from various basins in the U.S.S.R. has been established. More than 900 crude oils were analysed by TLC of which nine (as crudes, and as separated olefin fractions) were analysed by gas chromatography, infra-red spectroscopy and ~H NMR spectroscopy: two of the oils were similarly analysed by '3C NMR. The carbon skeletons of the olefins, and coexisting saturates, correspond closely. The location of the double bonds in the olefin molecules is regarded as near statistical. A new concept of olefin generation in crude oils, by natural radiolytic dehydrogenation of saturated hydrocarbons is proposed. Different degrees of the subsequent natural thermocatalytic degradation of olefinic double bonds have been observed. Key words---crude oil, olefins, hydrocarbon transformation, radiolysis, NMR of olefins INTRODUCTION

It is generally accepted that crude oils do not contain olefins in significant amounts and this acceptance may account for the fact that little attention has been paid to this class of hydrocarbon in crudes. One example in which considerable quantities of alkenes were found (Bradford field, Pennsylvania) has been described by Hoering (1977), and Kushnarev et al. (1989) detected olefinic hydrogen atoms in a number of East Siberian crude oils ( > 10 samples) using ~H NMR. Although the class of unsaturated compounds was not defined, the latter authors assumed that they were olefinic hydrocarbons. These two papers stimulated our interest in the problem of olefin-containing crudes with respect to their composition, and to the structures of the olefins. Preliminary data have already been published (Frolov and Smirnov, 1990). In the present paper a rapid TLC method is described for detecting olefins in crude oils and results are given for their detection, in about 900 petroleums from different U.S.S.R. basins. Also a chromatographic method for their isolation from crudes is used for their subsequent investigations using gas chromatography, infrared spectroscopy, IH and 13C N M R spectroscopy. Hydrogenation of olefins, isolated from some Russian crude oils, and of the crudes themselves, is reported.

EXPERIMENTAL

All solvents were distilled before use. The silica gel used for chromatography (L, 40/100, Chemapol, Czechoslovakia) was washed with an ether-methanol mixture (2:1) and activated at 120°C in vacuum before use.

TLC-detection o f olefins in crude oils The TLC method for detecting olefins in crude oils is based on them being revealed selectively by means of the fluorescein-bromine reaction (Kirchner, 1978) after separation on silica gel. The method is as follows: the oil (0.3-0.5 #1) is applied to a Silufol plate (Kavalier, Czechoslovakia: isooctane as developer) on which the saturated and olefinic hydrocarbons (Rf= 0.9-1.0) are separated from monoaromatic hydrocarbons (R r < 0.85). The plate is visualized using an aqueous-alcoholic solution of sodium fluoresceinate (0.05%), drying, and treating with bromine vapours (avoiding excess bromine). Olefins are confirmed by the presence of a distinct yellow colour against the rose-coloured background of the plate. This reaction gives a clear, positive, result if the crude contains more than 0.5% olefins. Isolation o f UHCs from crude oils Olefins were isolated from the crude oils by liquid chromatography. Firstly, a fraction containing the saturates plus olefins was separated from the crude on a column of silica gel by elution with hexane. Separation from monoaromatic hydrocarbons was monitored by TLC on Silufol plates (isooctane as developer). Olefins were then isolated from the aliphatic concentrate on a column containing silica gel (ratio 1:4) impregnated with 7-8% AgNO 3 (Kirchner, 1978). Saturates were removed by development with hexane (until the appearance of pure solvent). The olefins were then eluted with a 4:1 hexanediethylether mixture. Separation was monitored on Silufol plates previously treated with a saturated solution of AgNO3 in methanol and dried at 80-90°C for 20--30 min. Using hexane as eluant, the olefins had Rf values in the range 0.1-O.7 (the variation in Rf

189

190

E.B. FROLOVand M. B. SMIRNOV

~

"

x

1 .... 150

6

4

' ....

I .... 100

' ....

I .... 50

' ....

I 0

PPM Fig. 1. ~3CNMR spectrum (62.9 MHz) of the total olefin fraction of Verkhne-Chonsk crude oil obtained without NOE.

values is due to differences in the structure of the olefins). Detection was by means of phosphomolybdic acid (Kirchner, 1978). The absence of saturated hydrocarbons in the olefin fractions was proved by TLC on AgNO3-impregnated plates. The detection limit was about 0.5%. The amount of aromatic hydrocarbons, in the olefin fractions was assessed from IH and 13C N M R data (see later). All olefin fractions were free of saturated hydrocarbons but contained about 3-6% of aromatic hydrocarbons. This procedure, developed by us, has a clear advantage to that used by Hoering (1977), in which mercuration-demercuration reactions of olefins in an HCl-acid medium were used. Hydrogenation of the separated olefin fractions was achieved, at room *A structural element (SE) is a specified fragment of a molecule, a list of which is shown in Table 6.

temperature, in the presence of Pt02 and hydrogen (1.5 bar, 24 h). 13C and IH nuclear magnetic resonance measurements

The contents of saturated structural* elements were determined by ~3C NMR employing NOE with a 70° flip angle and 15 s pulse repetition rate (more than 3 times the maximum value of the spin-lattice relaxation time T~ for protonated carbons: Smirnov and Smirnov, 1985). The following conditions were used: pulse width= 13/~s; quadrature detection; number of pulses = 4000-8000; broadband decoupiing; sweep w i d t h = 6 3 2 9 H z for saturates and 10,870 Hz for olefins; filter width = 20 and 30 kHz; acquisition time--2.6 and 1.5 s; relaxation delay = 12.4 and 13.5 s, respectively. For measuring the content of carbon atoms at a double bond, and also structural elements containing

Unsaturated hydrocarbons in crude oils double bonds, the spectra were obtained without NOE as described by Shoolery (1977). A relaxation agent [Cr(acac)3, 0.08 mol/l] was added to the CDC13 solution of the samples, and an inversion gated decoupling pulse sequence was used with a 2.5 s delay ( > 5T l) during which proton decoupling was gated off. The following experimental conditions were used: pulse width = 17.5 #s; sweep width = 13,157 Hz; filter width = 35 kHz; acquisition time = 0.62 s; number of pulses = 20,000-75,000. The experimental conditions for IH N M R were: pulse width = 1 ks (flip angle = 18°); sweep width = 3600 Hz; filter width = 10 kHz; single channel detection with phase alternating pulse sequence; acquisition time = 2.3 s; relaxation delay = 1 s; number of pulses = 400-700 for olefin fractions and 700-5000 for crude oil samples. As a preliminary, the signal area ratios of aromatic plus double bond protons, to those of aliphatic protons were demonstrated to be the same by using a relaxation delay of 4.3 s. All N M R spectra were obtained on a Bruker WM-250 spectrometer at 250MHz (IH) and 62.9 MHz (t3C). All samples were dissolved in CDCI 3

191

to give 25% v/v solutions for crude oils and 35~,5% v/v for saturated and unsaturated fractions. The internal standard was TMS. RESULTS

We have examined almost 900 crude oil samples using the TLC-method described above. These oils were sampled in various petroleum-producing provinces of the U.S.S.R., in which secondary recovery methods had not been used. The results, summarised in Table 1, indicate that olefins are not rare in crude oils, as accepted earlier. Furthermore, since the occurrence of oils with a low olefin content exceeds those with high values, one supposes that the occurrence of petroleums with a low olefin content (say 0.1-0.4%) will preponderate in most of the provinces examined. Table 2 provides data on the olefin content and Inb/lt n ratios measured by ~H NMR. (See Table A1 for definition of terms.) Table 2 indicates that the olefin content in our crudes ranged from trace amounts to 10%. This is the first report of such high concentrations of olefins in

[ CHCI

--CH=CH--

~ C'l

.... 7

I .... 6

H2~ CH--

I .... 5

I''''1' 4

'''I

3

.... 2

I' i

I

I

!

I

PPM Fig. 2. JH NMR spectrum (250 MHz) of the total olefin fraction of Verkhne-Chonsk crude oil.

192

E . B . F g o L o v a n d M. B. SMmNOV Table 1. Occurrence of olefins in crude oils from some important U.S.S.R. oil provinces Number of felds Oil province (era of oil formation) 1 North Caucasus (KZ) (MZ) II West Siberia

(MZ)

III Dnieper-Donetsk (MZ, e z ) IV Byelorussia (PZ) V Volga-Urals (PZ) VI Timan-Pcchora (PZ) VII East Siberia~

Number of oil samples

Studied

With* olefins

Studied

With* olefins

Olefint rich

58

3

97

5

0

47

11

117

14

0

53

21

90

33

1

23

6

49

9

1

123

64

303

119

16

56

27

149

63

19

32

31

75

74

34

*Containing >0.3-0.5% of crude. tContaining > 1-1.5% of crude. SOils from Riphean-Vendsk and Lower Cambrian formations of the Siberian platform. KZ = Cainozoic; MZ = Mesozoic; PZ = Palaeozoic.

a wide range of oils. The Tetersk oil with 21% olefins is unique (Frolov and Smirnov, 1990). We emphasise that the olefins have been found in oils of diverse chemical composition: for example paraffin-rich (5,30); biodegraded, non-paraffin (16,21); heavy oil

(21,30); light oil (5,23); condensate (1). The content of aromatic hydrocarbons in the oils ranged from 4 to 70% (21,30) and had sulphur contents ranging from 0.1 to 2.5% (3 and 27 respectively). We have noted also the remarkable inconstancy in the olefin content

Table 2. Unsaturated hydrocarbons in some Russian crude oils

Spl. No.

Oil province

1 2

III III

3

V

4 5

V V

6 7 8 9 10

20 21 22 23 24 25 26 27 28

V V V V V VI VI VI VI VI VI VI VI VI VI VI VI VII VII VII VII VII VII

29 30

VII VIIIt

11

12 13 14 15

16 17 18 19

Field name Raspashnovsk VerkhneNovoselovsk Araslanovo

Well No.; age*; depth (m)

1~ff % (NMR)

8; 17;

MC; MC;

3576-3610 283%2851

0.02 0.02

31;

2216-2227 2137-2158 1688-1714 1666-1682 1618--1623 1732-1735 2012-2096 800-899 -920-923 999-1007 945-962 1106-1112 639-648 927-960 1390-1470 938-942 942-947 220 180 2525-2586 2360-2405 2528-2531 2648 1592 1726-1740

0.12 <0.01 0.10 0.14 0.12 0.06 0.08 < 0.01 :~ 0. I 1 0.16 0.26 0.12 0.10 0.20 < 0.01 :~ < 0.01 0.16 0.13 0.27 0.20 0.16 0.22 0.10 -0.23 0.17

Araslanovo

112;

Kamenologsk Kamenologsk Kamenologsk Rusakovsk Tikchovsk Tikchovsk Michayu Nibelsk Nizne-Omrinsk Niznc-Omrinsk Nizne-Omrinsk Pashnya Pashnya Pashnya Verkhne-Omrinsk Verkhne-Omrinsk Yarega Yarega Krivoluksk Kuymbinsk Mezdurechensk Sobinsk Verkhne-Chonsk VerkhneViluchansk Tetersk Minusinsk

288; 410; 243; 122; 22; 22; 582; 115; 337; 345; 57; 48; 69; 83; 176; 89; 16; --; 3; 2; 2; 32; 41; 622;

MC; MC; LC; LC; LC; LC; LC; LP; UP; MD; UD; UD; MD; UP; LP; UD; UD; UD; MD; MD; PC; PC; PC; PC; PC; PC;

278; 3;

PC; MD;

1940 1460

0.43 0.29

Olefin content (wt%) in crude oil

2.7 2.3 1

2.8

8.1

1

6.4

0.64 6.5 21.0 10.4

*Ages respectively: UP, LP--upper, lower Permian; MC, LC--middle, lower Carboniferous; UD, MD--upper, middle Devonian; PC--Precambrian. tTbe upper reaches of the Yenisei. :~Detection limit.

Unsaturated hydrocarbons in crude oils in oil samples from a single feld (even from a single formation), for instance in the Araslanovo (3,4); Tikhovsk (9,10); Pashnya (16,17) and Nizne-Omrinsk

(13,14) samples. The olefins isolated from samples 3, 5, 9, 13, 21, 27, 29, 30 were studied in some detail. GCs of the olefin fractions isolated by TLC from the Minusinsk, Tikhovsk, Tetersk and Yarega crudes, samples 30, 9, 29 and 21 respectively, shown in Fig. 4, are complex and do not contain any individual compounds in high concentration. GCs of the olefins isolated from the paraffnic Minusinsk and Tikhovsk oils (and, to a lesser degree, the Tetersk oil) show a number of homologous series, which from their retention indices, and co-injection with standards, appear to be trans- and cis-n-alkenes (Frolov and Smirnov, in preparation). This assignment appears to be confirmed, since hydrogenation of the olefin fraction gave a homologous series of n-alkanes. We have found that the content of n-alkenes in the total olefin fractions increases with an increasing content of

193

n-alkanes in the saturate fractions. A gas chromatogram for each olefin fraction indicated that the envelope maximum was shifted to higher values than for the co-existing saturated fraction of the same oil. Gas chromatograms of the hydrogenated olefin fractions, isolated from the Verkhne-Chonsk, Minusinsk and Yarega crude oils are shown in Fig. 5. The hydrogenation of squalene to squalane indicated that complete hydrogenation of trisubstituted double bonds took place under the conditions used. Despite this, a trace of olefins (possibly tetrasubstituted) was detected by silver ion TLC in the twice-hydrogenated olefin fractions. The hydrogenation products contain n-alkanes, pristane, phytane, and other iso-alkanes, and comparison of Fig. 5 with Fig. 6 demonstrates their near identity with the co-existing saturates. Moreover, when the saturates have any distinctive compositional features the corresponding hydrogenated olefins possess them too. An example is the Verhne-Chonsk oil which exhibits a high content of 12- and 13-methylalkanes, branched hydrocarbons

x32

""

....

I .... 8

' ....

I ....

' ....

6

I .... 4

' ....

I ....

' ....

2

PPM

Fig. 3. ~H NMR spectrum (250 MHz) of the Yarega crude oil. OG 21/2--F

I-0

194

E.B. FROLOVand M, B. Star.Nov

typical of oils from the East-Siberian platform (Makushina, 1978). Also, saturates in the Minusinsk oil contain no pristane or phytane, neither does the hydrogenated olefin fraction. A similar correspondence is evident in the biodegraded Yarega oil which is free of n-alkanes (Figs 5, 6). The i.r. spectra of the olefin fractions of all the oils studied are similar to each other as exemplified

for the Tikhovsk oil, and compare with the coexisting saturates except for the presence of an intense absorption band due to trans-disubstituted double bonds (970cm -I) in the olefin fractions (Fig. 7). Other olefin groups such as CH2=-CH-- and CH2~---CRR' with strong absorption bands at 910, 990 and 890cm -~ appear to be absent, or very weak.

(a)

2O

(b) 15 25

20

Y Time

Fig, 4(a) and (b). Caption on facing page.

Unsaturated hydrocarbons in crude oils

195

(c)

25

20

15

(d)

25

Time

20

~

'

~

x

,

Fig. 4. Gas chromatograms of the total olefin fractions isolated from (a) Minusinsk, (b) Tikhovsk, (c) Tetersk, and (d) Yarega crude oils. Capillary column (25 m x 0.25 ram) coated with Apiezon L; carrier gas hydrogen. Column temperature programmed from 100 to 250°C at 4°C/min. Peaks coinciding in retention time with n-alkanes correspond to n-trans-2-alkenes: several are arrowed in the figure.

196

E.B. FROLOVand M. B. SmRNOV

(a)

17

18

Toc

(b)

18 17

T~C -.~

Fig. 5(a) and (b). Caption on facing page.

Unsaturated hydrocarbons in crude oils

197

(C)

Pr

Ph

T°C

Fig. 5. Gas c h r o m a t o g r a m s o f hydrogenated olefins of (a) Verkhne-Chonsk, (b) Minusinsk and (c) Yarega crude oils. o = 12- and 13-methylalkanes (see Makushina et al., 1978). Conditions: column 25 m x 0.25 m m coated with SE-30, programmed from 100 to 280°C at 5°C/min; carrier gas hydrogen.

Characterisation of olefins in petroleum by IH and "C NMR Average molecular structural parameters of the olefin f r a c t i o n s c a l c u l a t e d f r o m N M R d a t a a r e g i v e n in T a b l e 3. M o r e t h a n 9 0 % o f t h e olefins w e r e s h o w n to be m o n o e n e s (Kdb = 1.0) at l e a s t f o r h y d r o c a r b o n s b e l o w C3o.

Analysis of the NMR spectra enables the composition, a n d s t r u c t u r e , o f t h e olefins to be s t u d i e d in s o m e detail. F i n g e r p r i n t s in t h e 4 . 5 - 5 . 8 p p m r e g i o n o f t h e ' H s p e c t r a are s i m i l a r f o r all olefin f r a c t i o n s as e x e m p l i f i e d in F i g s 2 a n d 3. T h i s is s o for t h e s p e c t r a o f t h e c r u d e oils also (the Hdb s i g n a l s h a p e is welld e f i n e d f o r c r u d e s , w i t h In/ltn>O.lO-O.15: see T a b l e 2).

Table 3. Values for the average molecular structural parameters for the olefin fractions of some crude oils [C

rJo)"

Crude oil

Fraction

N*

Q~b

K~'b

VerkhneChonsk

total Ci0~12 C 16~Ci9 C24~27

30.5 I 1.2t 17.7t 25.5t

2.85 2.70 2.80 2.80

-0.96 1.04 0.97

2.06 5.70 3.69 2.38

6.55 17.1 I 1.8 7.60

Kamenoiogsk

total

28.0

2.75

--

2.44

7.05

Minusinsk

total

39.0

2.40

--

2.12

5.15

NizneOmrinsk

total C16-C19:~

26.0 17.7t

2.65 1.85

-0.98

2.61 5.98

7.66 I 1.1

Tikhovsk§

total

24

3.0

--

2.16

8.3

Yarcga

It

It

total 30.0 2.95 -1.94 6.63 *See Table 8 for definition of terms, tDetermined from GC data. :~Subfraction which formed a complex with urea (see text). §For this sample the error of measurement is about two times that for the others (about 5-7% for Qdb, 7-10% for Cdb/Ct a n d 10-15% for Nu measurements) because the spectrum of this sample was obtained with NOE only.

198

E. B. FgOLOVand M. B. S~mtNov

For the olefin fractions, weak signals (H,~b) due to protons in CH2~------CRR' structures centred near 4.65 ppm are apparent (Chamberlain, 1977; Kalabin et aL, 1986). The ratio of H,]b/Hdb is about 0.01-0.03 for the olefin fractions of different erudes. This corresponds to less than 2% of hydrocarbons containing these groups per fraction. The signal intensity of protons in CH2~------CH-- groups is very low in all the olefin fractions; their abundance cannot be determined from ~H N M R spectra because of the presence of tri-substituted double bonds (Kalabin et al., 1986), but could be determined from ~3C N M R spectra using the signal at l l 4 . 1 2 p p m due to the terminal carbon atom in the structural element CH2~----CH----CCC--: the values determined ranged from 0.2 to 1.0%. Hence almost all Hdb signals are due to di-substituted ---CH~-----CH-- and tri-substituted C-------CH-moieties which, unfortunately, cannot be distinguished by ~H N M R because of overlapping proton resonances. The strong signal centred at 5.39ppm (Fig. 2) is due to the ---CH~----CH-- group in the carbon chains (Kalabin et al., 1986). Typical ~3C N M R spectra of olefin fractions, and corresponding saturated hydrocarbon fractions, are shown in Figs 8-12. Sharp signals due to C--------C resonances in the 116--151 ppm region were assigned based on the following data:

Table 4. Assignment of carbon signals (for carbon atom underlined) in double bond-containing structural elements (SEs); (~3C N M R spectra in C D C I 3 solution; 6 from TMS)

SE _C--~CCCG-~ccc--

Isomer

cc_-----cc--

trans

-ClS

c_c----cccc--

trans ClS

cc-----~_ccc--

trans ¢1s

cc_c---ccc--

trans ¢ls

ccc----cccc--

trans ¢1s

cc~ccc--

trans ClS

6 (ppm) 114.12 139.04

1

2

17.88 12.72

(dl) (d2)

124.46 123.52

3 4

131.72 130.90

5 6

25.74 20.65

(d3) (d4)

131.89 131.49

7

129.44 129.36

9 10

8

_c_ccc----c-CCCC------CC--

trans trans

13.68 34.89

(dS) (d6)

ccc_c-----cccc--

trans cis

130.I 1 129.60

11 12

CCCC-----CCCC--

trans cis trans trans trans

130.65 130.12 22.35 31.61 29.03

13 II (d7) (dS) (d9)

trans

32.79 32.74 27.40 27.28 130.40 130.33 129.90 129.83

(cliO) (dlO) (dll)

15.86

(d12)

CCCCC----C-CCC_CCC.---~C-CCCCCCC----C--

----chemical shift values of n-octenes (Dorman et aL, 1971), --spectra of the model compounds: -2, -3, -4 and -5 ---CCCC_C------CC-n-undecene; n-octadec-9-ene and 6-methylundec-5ene (all as mixtures of cis- and trans-isomers), --spectrum of the olefins isolated from the C16""C19 --CCCC-----CCCC-fraction of the Nizne-Omrinsk crude, after adduc---CCC----CC-tion with urea, --spectra of olefin fractions recorded using GASPE and CSE pulse sequences (Cookson and Smith, 1981). Signals in the Cal resonance region (5-65 ppm) in the olefin fractions were assigned similarly. Chemical shifts for each of the above model compounds were measured separately in CDCI3 solution, and mixed with the olefin fraction (5% model compound/95% olefin fraction in CDCI3). Data are given in Table 4 and Figs 8, 9 and 11. Normal alkene structures with double bonds in positions 1-, 2-, 3-, 4-, (trans- and cis-isomers) as well as 5-, 6- and 7- (trans-isomers) have been identified together with branched trans compounds of structure R--CC----C(Me)CC--R' (see Table 5). The proportion of unbranched to total olefins differs in different crudes but, as a rule, their abundance increases with that of the n-alkanes in the saturated fraction of the oil. Compounds with structures C-----CCCC--R, CC----CCCC--R, CCC----CCCC--R, CCCC----CCCC--R and R'--CCCC-----CCCC--R make up about one-third of all olefins in the Minusinsk crude.

--

Signal No. in Fig. 8 (Figs9 and 11)

cis trans cis trans

14 15 16 17

The more thermodynamically stable trans-isomers predominate over cis-isomers in all of our samples. The cis-/trans- ratio differs in different petroleums and depends on the double bond position but is greatest for bonds in the 2 position and least in position 3. Most double bond carbons give signals that appear in the ~3C N M R spectra as an unresolved "hump" (Fig. 1). Signals forming the "hump" are due to ring double bond carbon atoms and to those in ~-, fl-, ~-, or 6-positions to rings. Hence, one supposes that cyclic compounds predominate in the olefin fractions of petroleums (see Appendix). As a rule, the ratio of "hump" area to narrow signal area, measured for Cob resonances in the olefin spectra correlates with the same ratio measured for Ca resonance of the same olefin fraction, and in the spectra of the coexisting saturates. Furthermore, the above ratios generally increase with an increasing content of n-alkanes in the oils.

199

Unsaturated hydrocarbons in crude oils Table 5. The content of olefins in the three crude oils shown. Measured from lac NMR data (units per 100 "average olefin molecules") Total olefin fraction of: Type of hydrocarbons

Isomer

Minusinsk

Kamenologsk

Verkhne-Chonsk

--

0.30

0.30

0.70

CC------CCCCR

trans cis

4.0 0.95

2.6 0.65

2.0 0.75

CCC----CCCCR

trans cis

3.3 0.45

2.1 0.30

1.6? 0.30?

CCCC----CCCCR

trans cis

3.5 0.6

1.8 0.4

1.5? 0.4?

CCCCC-----CR CCCCCC----'CR CCCCCCC-----CR

trans trans trans

3.6 3.1 2.9

1.7 1.8 1.6

----

R 'CCCC------CCCCR

trans cis

18 3~0

8.4 1.7

4.3 0.90

R "CCC------CCR *

trans

4.0

1.2

0.5

CC----CCR

trans

5.5

4.0

4.6

~CCCR

*Note that the signal at 15.86 ppm can overlap the CH3-group signal of ethyl substituted benzene rings. Thus the value measured may be high. ?The error here is about 20%, because of greater signal overlap than for other oils. The error for the other two oils is about 10% (Smirnov and Smirnov, 1985). T h e 13C N M R s p e c t r a o f total olefins a n d s a t u r a t e s , isolated f r o m the M i n u s i n s k , K a m e n o l o g s k , Tikchovsk, Nizne-Omrinsk, Verkhne-Chonsk, Tetersk a n d Y a r e g a c r u d e oils were c o m p a r e d in the Cal

r e s o n a n c e region, as exemplified f o r the first t w o oils in Figs 9 - 1 2 a n d T a b l e 6. T h e n u m b e r o f alkyl, alkenyi a n d s o m e m o n o cycloalkyl s t r u c t u r a l e l e m e n t s p e r m o l e c u l e were

(a)

17

18 Ph

i

Fig. 6(a).

Caption overleaf.

2OO

E. B. FROLOVand M. B. SMIRNOV (b)

17 I$

& (c)

Pr

Ph

Fig. 6. Gas chromatograms of saturated hydrocarbons isolated from (a) Verkhne-Chonsk, (b) Minusinsk and (c) Yarega crude oils. Conditions as for Fig. 5.

U n s a t u r a t e d h y d r o c a r b o n s in crude oils

calculated using narrow 13C NMR signal areas. The calculations can be interpreted in two ways. Firstly, for all given SEs, the calculated value shows the number of that particular SE per "average molecule" (or, as in Table 6, per 100 molecules). Thus, in the Minusinsk oil saturates there are, on average 1.91 terminal n-C4 and 1.62 terminal n-C7 alkyl groups per molecule. The values for the Kamenologsk oil saturates are 1.15 and 1.01 respectively (Table 6). The difference between these oils means that the Minusinsk saturates contain more n-alkanes, and other compounds with n-alkyl fragments, in the molecule.

201

Secondly, if the molecules contain only one given SE (as is probable for cyclic SEs, 3-, 4-, 5-, methylsubstituted SEs, ethyl-, propyl- and butyl-substituted SEs; signals 21, 22, 23, 8, 9, 11, 16, 18, 19 respectively, in Table 6), the calculated value is nothing but the total proportion of all compounds containing the SE in question in their structure. For example, the saturates in Kamenologsk crude contain 3.8% monosubstituted cyclohexanes, 2.0% monosubstituted and 1.6'/o dialkylsubstituted cyclopentanes. Differences between the composition of the Kamenologsk and Minusinsk saturates is evident from Table 6. Kamenologsk crude is typically

Table 6. The content of some structural elements (SEs) in the total saturated and unsaturated hydrocarbon fractions of Minusinsk and Kamenologsk crude oils. Measured from ~3C NMR data (units per 100 "average molecules") Signal No. Figs 8--I I

Minusinsk

Kamenologsk

6 (ppm)

saturates

olefins

saturates

olefins

1 2 3 4 5 6

14.13 22.89 32.18 29.62 29.92 39.34

191 187 166 162 830 6.2

183 170 143 129 760 4.5

115 116 104 101 460 4.2

90.5 81.5 60.5 59.0 360 1.8

7

27.67

5.2

4.0

3.7

1.2

8

19.29

7.3

4.9

7.7

5.0

9

13.44

4.7

--

6.4

3.0

10

20.33

5.3

4.4

5.6

3.2

l1

23.24

5.3

--

4.5*

2.9*

12

33.02

20.5

15.0

20.5

13.0

CCCCCCC-C C --CCCCCCC-C C

13

25.03

3.0

1.7

8.8

4.2

14

24.70

4.7

3.3

13.0

9.7

CC c~CCC-C

15

25.74

0.75

0.55

0.45

0.2-0.3

--CC

16

26,18

5.7

5.5

2.0

1.7

17

14.58

5.9

7.0"

1.9

1.5"

18

20.06

4.2

4.9

1.5

--

19

33.54

5.2

5.7*

1.4"

1.1"

20

34.03

22.5

31.5

2.3

2.7

--C--

21

26.69

4.8

3.5

3.8

1.7

~

---C--

22

25.38

0.40

0.35

2.0

1.0

a(~

--CC--t --CC--

23

24.04

0.2-0.3

0.3

1.6

1.0

SEs CCCC-C CCCC-CCCCCC-CCCCCCC---CCCCCCC-CCCCCC-C CCCCCCC-C CCCCC-C CCCCC-C CCCCCC-C CCCCC-C --CCCCCCC--

cc/X"ccc ---C

-ccc/x'cC--CC CCC ~N'CCC--CC

cccc

xCCC--

-~%CC-/ --CCCC_"

*The error is about 20%. The error for other values is 7-10% (Smirnov and Smirnov, 1985). tTrans-isomer; cis-isomer was not found.

E. B. FROLOVand M. B. SmRNOV

202

(cm-I) 1500

1000

700

i

]

10

30 970

50

(a) 70

30

50

(b) 70

Fig. 7. i.r. spectra of (a) olefin and (b) the saturated hydrocarbon fraction isolated from Tikhovsk crude oil (UR-20), film, d ~ 0.035 mm.

h 14'

5

9 7

6

-1

3

16

19

.... I . . . .

' ....

I ....

132

'

130

4 l

I

. . . .

I

135

. . . .

I

. . . .

I

130

125

'

'

'

I 120

.

.

.

.

f

115

PPM Fig. 8. The region of double bond carbon signals in the 13C N M R spectrum (62.9 MHz) of the total olefin fraction isolated from Nizne-Omrinsk crude oil. Numbered signals are assigned in Table 5.

'

Unsaturated hydrocarbons in crude oils dlO

203

3

dll 12

d9 d8

I *

19

21 16 1t3

I

.

.

.

45

.

I

.

.

.

.

I

40

"

"

'"

"

35

I

.

.

.

.

d7

I

30

.

•. ,

18

.



d5

|.

25

. . . . ..

.

20

|

15

PPM Fig. 9. Region of aliphatic carbon signals in the ~sCNMR spectrum (62.9 MHz) of the total olefin fraction isolated from Minusinsk crude oil. The assignment of signals dl-dl2 and 1-23 is given in Tables 4 and 6 respectively, paraffinic with a not uncommon isoalkyl SE content but with some increase in alkylcyclopentanes (essentially dialkylsubstituted). Crude oils with a saturate composition, such as Minusinsk, are rarely found. This highly paraffinic oil contains no isoprenoids. In addition, the content of "T-shaped" structures (i.e. with mid-chain ethyl-, n-propyl-, n-butylsubstituents: signals 20, 16, 18, 19 respectively in Table 6) is unusually high. The same compositional difference was also found for the Kamenologsk and Minusinsk olefin fractions. In all of the petroleums studied the skeletal structures of the olefins were similar to that of the

5

saturates. The most important differences between the olefins and saturates of the same petroleums are as follows. The olefin fractions contained a higher proportion of cyclic compounds. The proportion of compounds with methyl branches was about 1.5-2 times greater for the saturates than for the olefins. DISCUSSION

This detailed study of the composition, and molecular structure, of olefins in a wide range of Russian oils allows one to draw the following conclusions. Firstly, there is a close structural relationship between

.11

20 12

I

6

I t

~ 119 j s

15

I

45

. . . .

I

40

. . . .

I

35

• .

|

1

2

4

. . . .

30

' 9

x

1314 |

25

. . . .



20



|





,

15

PPM Fig. 10. ~3C NMR spectrum (62.9 MHz) of the total saturates isolated from Minusinsk crude oil. Designations are the same as in Fig. 9.

,

E. B. FROI..OVand M. B. SMIRNOV

204

dl0 3

12

ili

*

' ~

I

.

.

.

.

I

45

.

.

.

.

2o

I

40

i,,1

.

.

.

.

I

35

.

.

.

.

I

30

.

.

.

.

25

I

.

.

.

.

20

I

.

.

.

.

15

PPM Fig. 11. Region of aliphatic carbon signals in the ~3C NMR spectrum (62.9 MHz) of the total olefin fraction isolated from Kamenologsk crude oil. Designations are the same as in Fig. 9. the olefins and coexisting saturates in the oils. This is clear by comparing GC fingerprints of the indigenous saturates with the hydrogenated olefins, as well as |3C N M R spectra of olefins with the co-existing saturates. We are also of the opinion that the location of the double bonds in the olefins is near-statistical. Of course, determination of the exact location of all double bond positions is very difficult due to the complexity of the mixture of homologues and isomers. Although resolution of the above problem needs further investigation, one can advance arguments for a near-statistical distribution of the double bonds.

Such a distribution for the unbranched olefin isomers in the urea adduct of the C I 6 - - - C I 9 fraction of the Nizne-Omrinsk oil was demonstrated from ~3C N M R data (ratios of the sum of the cis- and transisomers of l, 2, 3, 4, 5, 6, and 7-enes are about 0.30:1.0:0.7:0.55:0.5:0.45:0.40). Hoering (1977) came to the same conclusion after ozonolysis of the olefins isolated from the Bradford oil. During our study of the NMR spectra of olefin fractions, we have found no preferred double bond positions. Thus, the strongest, sharp signals in the olefin region of our 13C NMR spectra correspond to structural elements with in-chain double bonds. The marked carbon atoms of

4

2

12 x~

21

.%

13 14

6

*

,



~

I

45

.

.

.

.

I

40

.

.

.

.

t

20

I

35

.

.

.

.

i'"" " 30

"

"



I 25

. . . .

I 20

. . . .

| 7

15

PPM Fig. 12. ~3C NMR spectrum (62.9 MHz) of the total saturates isolated from Kamenologsk crude oil. Designations are the same as in Fig. 9.

Unsaturated hydrocarbons in crude oils their saturated analogs also give strong signals in the aliphatic region (see Tables 4, 6 and Figs 8-12). Gas chromatograms of the olefins, and their hydrogenation products (Figs 4 and 5) indicate that for each n-alkane and isoprenoid peak in the hydrogenated fractions there corresponds a series of lowintensity precursor peaks in the original olefin fractions. We consider that this direct genetic relationship of the olefins to saturates is most likely due to the non-selective dehydrogenation of the saturates to give a near statistical distribution of double bonds in the olefins. We do not believe that olefins of such structure can be formed by the natural thermal cracking of petroleums as proposed earlier (Hoering, 1977) for a number of reasons. It is well known that the cracking process leads to a considerable decrease in chain length and hence molecular weight of the olefins formed. Conversely, the crude oil olefins are generally of higher molecular weight than the coexisting saturates. Also, olefins formed by thermal cracking contain mainly terminal (CH2~-----CH--) and exomethylene (CH2~---CRR') double bonds. Acid-catalysed (mineral) isomerization of such olefins cannot explain the observed nearstatistical distribution of double bonds. At each stage of migration of a double bond, from the end to the middle of the carbon chain, there should be an accompanying sharp decrease in the content of each subsequent isomer due to many secondary reactions (isomerization of the skeleton, eyclization, etc). On the contrary, according to our ~3C N M R data, isomers with double bonds in mid-chain positions prevail over alk-l-enes and alk-2-enes to a great extent, this regardless of the concentration of olefins in the oils. Indeed, it is difficult to explain olefin contents as great as 10 and 21% by thermal cracking. The composition and the structures of the olefins in the oils can be explained, if one supposes that olefins are products of the radiolytic dehydrogenation of saturated hydrocarbons. The radiolysis of formation waters, by radiation from radium and radon, is a known geochemical process (Vovk, 1979). We propose that radiolysis of the hydrocarbons in the oil, initiated by the or-particles of Ra, and most likely by Rn also (radon is readily soluble in oil), generate the olefins. High concentrations of these radioactive elements are known to occur in oil fields. The general rules of radiolysis of saturates are briefly as follows (Cseret et al., 1981). During the initial stage of the radiolysis of alkanes, monoolefins with the same skeleton with a near-statistical location of double bonds are the predominant products. Molecules with this type of structure amount to 50-80% of all the unsaturated products, and from 30 to 60% (discounting CI-C4) of all the hydrocarbon products. Subsequent dehydrogenation of the monoolefins to diolefins during radiolysis is unusual. The yield of monoenes is proportional to the irradiation dose, and does not depend on temperatures ranging

205

up to 250-300°C. In general, the above rules hold for all types of ionizing radiation, since the alkanes are dehydrogenated mainly by low-energy (20-80eV) electrons (about 3-5 x l04 electrons per g-particle) which are generated in the hydrocarbon phase by high-energy (103-107 eV), ~t, fl, and y bombardment. The radiolysis of saturated hydrocarbons in conditions modelling oil formation has not been studied. However, since the major proportion of the monoolefins are formed through non-selective monomolecular elimination of hydrogen from highly excited alkane molecules, we assume that this mechanism is insensitive to the influence of other factors such as the presence of water, salts, minerals, etc, and can take place in the reservoir and/or source rock. General rules for the radiolytic dehydrogenation of alkanes allow us to explain certain important experimental facts. As noted, the olefins in crude oils are generally of higher molecular weight than the coexisting saturates. This we attribute to the fact that an exciting electron has a greater probability of encountering a larger, rather than smaller alkane, other conditions being equal. We have ascertained that the crude oil olefins are mainly monoenes, as are those generated experimentally by radiolysis. Again, as stated before, the content of branched and cyclic olefins is higher than similar skeletons in the saturates. This appears to be due to the fact that the yield of double bonds, generated by radiolytic dehydrogenation at branched positions and, in particular, in cyclic structures is roughly 3-5 times greater than at unbranched positions. Also, olefins are probably the most reactive constituents of petroleum. Obviously, thermocatalytic degradation of the olefins which may subsequently take place will do so with a preferential decrease of olefins with the least substituted double bonds. This is due to their isomerisation into thermodynamically more stable tri- and tetrasubstituted olefins, and also to their greater rate of transformation (they are more reactive owing to steric accessibility) into other classes of petroleum compounds by cyclisation to saturates, dehydrogenation to aromatic hydrocarbons, formation of sulphur-compounds, etc. The C----CCCC--R group has the highest free energy of the open chain olefin structures detectable by ~3C N M R (see Tables 5 and 6), and in comparison with more stable structures (for example cis- and trans- non-terminal olefins) must react to a greater extent in the degradation process. Table 7 shows that the C-----CCCC--R/CC----CCCC--R ratio (for Verkhne-Chonsk and Nizne-Omrinsk oils) is considerably higher than equilibrium values (Zhorov et al., 1977) and approximates to that observed for the radiolysis of n-aikanes (Rappoport and Gaumann, 1973). According to our reasoning, this ratio should be about the same for olefins formed in oils and those generated by radiolysis in model experiments. The ratio approximates to the equilibrium value for olefins in the Minusinsk and

206

E.B. FROLOVand M. B. SMmNov 7. Ratios of isomericolefinstructures Total olefinfractionof crudes: NizhneVerkhneOmrinsk Minusinsk Kamenologsk Chonsk 0.06 0.07 0.25 0.24 Table

Ratio C------CCCCR/CC-----CCCCR cis-/transof CC-----CCCCR 0.24 0.18 0.37 cis-/transof RCCCC.---'--CCCCR' 0.16 0.20 0.21 *For n-alkenesgeneratedby radiolysis.After Rappoport and Gaumann(1973). ?After Zhorov et aL (1977). ~This valueis dependenton R. Kamenologsk oils which seems to indicate that the olefins in these two oils are appreciably degraded. As seen in Table 7 the cis-/trans- CC-------CCCC--Rratios for the Minusinsk and Kamenologsk olefins are lower than the equilibrium value, but for the VerkhneChonsk and Nizne-Omrinsk oils it approximates the radiolytic value. The cis-/trans- ratio for the more sterically hindered R '---CCCC------CCCC--R structures is less changeable for all of the oils. Therefore the data in Table 7 not only demonstrate the thermocatalytic degradation of petroleum olefins but also the degree of degradation, from a detailed description of their molecular structure. Considering our data, and that of Kushnarev et al. (1989), we conclude that crude oils of the Riphean-Vendsk and Lower Cambrian formations of the Siberian platform are generally characterized by the presence of olefins. The petroleum geochemistry of the basin has been described by Kontorovich et al. (1986), and Drobot (1988). The oils in this basin occur on (or close to) the surface of the basement complex, in formations with a high level of natural radioactivity. The dissolved and free gases of the crude oils contain hydrogen (up to 0.5% and several wt% respectively). Ethylene was not observed. Note that hydrogen is the main product of the radiolysis of saturates, while ethylene is the main product of thermal cracking. The accumulation of olefins in the oils of this basin is clearly assisted by the abnormally low formation temperatures (10--40°C at 1600-2400 m). The data in Table 1 demonstrate that, in general, there is a tendency for olefins to occur more often in oils from older formations. It appears that the time needed for the natural radiolytic generation of olefins in oils is short on a geological time scale. An approximation is that about 5% of olefins can accumulate in about 107 yr if it is assumed that (a) no transformation of the olefins occurs; (b) the permanent Rn value is 10 -8 Ci per litre; (c) the total 222Rn~ 2°rpb disintegration energy is about 40 MeV; (d) two olefin molecules are formed by absorption of 100 eV. It has been shown that the amount of Rn can reach values of 10 -7 Ci per litre close to basement complex fractures (Baranov and Titaeva, 1973). In conclusion, we emphasize that the presence of olefins in crude oils should be regarded as the result of natural radiation; this conclusion does not appear to have been reached previously. We assume that

Radiogenic 0.54).6

Equilibrium? value 0.01

0.38

0.354).45

0.30

0.22

--

0.15-0.25~

n-alkenes*

natural radiolytic dehydrogenation can affect not only saturated hydrocarbons but also other classes of compounds containing hydrocarbon substituents. The natural radiolysis of petroleum is capable of playing an important role in reforming their composition in that the generated olefins can be transformed into other types of compounds due to thermocatalytic and/or addition reactions. Since the activation energy of radiolytic transformations is near zero (Cseret et al., 1981), while that of thermocatalytic reactions is much higher (150kJ/mol; Goldstein, 1983), the relative contribution of radiolysis should be regarded as significant during the low temperature transformation of sedimentary organic matter. Associate Editor--G. W. Lijmbach Acknowledgements--The authors are grateful to O. A. Arefev; S. V. Atanasayn; O. K. Bazhenova; E. G. Burova; V. A. Chekhmakhchev; D. I. Drobot; I. V. Goncharov; A. N. Guseva; G. A. Kalabin; D. F. Kushnarev; A1. A. Petrov; and V. Punanov for providing oil samples from their collections and to V. P. Panov; S. V. Shorshnev and S. V. Tarabakin for assistancein obtaining NMR spectra. Thanks are also due to V. A. Melikhov for help with the experimental work. We would like to thank the reviewers for their critical reading of the original text of the paper. We especially wish to acknowledge Dr A. G. Douglas for helpful discussionsand invaluable contributions by improving the original manuscript. Acknowledgement is made to Mrs P. Haselhurst for typing the final document. REFERENCES

Baranov V. I. and Titaeva N. A. (1973) Radiogeology. Moscow University, Moscow (in Russian). Chamberlain N. F. (1977) The Practice of NMR Spectroscopy with Spectra-Structure Correlations for Hydrogen--l. Plenum Press, New York. Cookson D. J. and Smith B. E. (1981) Improved methods for the assignment of multiplicity in ~3C NMR spectroscopy with application to the analysis of mixtures. Org. Magn. Reson. 16, 111-116. Cseret Gy, Gyorgy I. and Wojnarovits L. (1981) Radiation Chemistry of Hydrocarbons (Edited by Foldiak G.). Academiai Kiado, Budapest. Dorman D. E., Jautelat M. and Roberts J. D. (1971) Carbon-~3 nuclear magnetic resonance spectroscopy. Quantitative correlation of the carbon chemical shifts of acyclic alkenes. J. Org. Chem. 36, 2757-2766. Drobot D. I. (1988) Geochemistry of Crude Oils, Condensates and Natural Gases of Riphaen-Vendsk and Lower Cambrian Deposits of the Siberian Platform. Nedra, Moscow (in Russian).

Unsaturated hydrocarbons in crude oils Frolov E. B. and Smirnov M. B. (1990) Higher unsaturated hydrocarbons in crude oils. Neftekhimiya 30, 147-157 (in Russian). Goldstein T. R. (1983) Geocatalytic reactions in the formation and maturation of petroleum. Bull. Am. Assoc. Pet. Geol. 67, 152-159. Hoering T. C. (1977) Olefinic hydrocarbons from Bradford, Pennsylvania crude oil. Chem. Geol. 20, 1-8. Kalabin G. A., Polonov V. M., Smirnov M. B., Kushnarev D. F., Afonina T. V. and Smirnov B. A. (1986) Quantitative Fourier transform NMR spectroscopy for petroleum chemistry. Neftekhimiya 26, 435--463 (in Russian). Kirchner J. G. (1981) Thin-Layer Chromatography. Mir, Moscow (in Russian). Kontorovich E. A., Surcov V. S. and Trofimuk A. A. (1986) Nepsk-Botuobinsk Anticline as a New Promising Region for Oil and Gas Production in Eastern U.S.S.R. Nauka, Novosibirsk (in Russian). Kushnarev D. F., Afonina T. V., Kalabin G. A., Presnova R. N. and Boganova N. I. (1989) Investigation of the composition of petroleums and condensates of the south Siberian platform by IH and 13C N M R spectroscopy methods. Neftekhimiya 29, 435-440 (in Russian). Levy G. C. and Nelson G. L. (1972) Carbon-13 Nuclear Magnetic Resonance for Organic Chemists. WileyInterscience, New York. Makushina V. M., Arefyev O. A., Zabrodina M. N. and Petrov A1. A. (1978) New relict petroleum alkanes. Neftekhimiya 18, 847-854 (in Russian). Petrov AI. A. (1987) Petroleum Hydrocarbons. Springer, Berlin. Rabenstein D. L. and Nakashima T. T. (1986) Nuclear magnetic resonance spectroscopy and its application to trace analysis. In Chemical Analysis, Vol. 84, Trace Analysis: Spectroscopic Methods for Molecules (Edited by Christian G. D. and Callis J. B.). Wiley-Interscience, New York. Rappoport S. and Gaumann T. (1973) Radiolyse des hydrocarbures, n-Alkanes en phase liquide de l'heptane au dodecane. Helv. Chim. Acta 56, 531-542 (in French). Smirnov B. A. (1969) Comparative study of the composition of organic compounds in complex mixture by combined spectral micromethods of analysis. In Methods For Analysing Mixtures of Organic Petroleum Compounds and Their Derivatives (Edited by Smirnov B. A.). Nauka, Moscow (in Russian). Smirnov M. B. (1990a) Saturated petroleum hydrocarbons of T-shaped structure. Neftekhimiya 30, 158-165 (in Russian). Smirnov M. B. (1990b) 1,2-Dialkylcyclopentanes in higher molecular weight petroleum fractions. Neftekhimiya 30, 445-448 (in Russian). Smirnov M. B. and Smirnov B. A. (1985) Using t3C NMR spectroscopy for the quantitative analysis of high molecular weight petroleum hydrocarbons. Neftekhimiya 2 5 , 402-411 (in Russian). Volk I. F. (1979) The Radiolysis of Formation Waters and Its Geochemical Significance. Nedra, Moscow (in Russian). Zhorov U. M., Pancbencov G. M. and Volokhov G. S. (1977) Isomerisation of Olefins. Khimiya, Moscow (in Russian).

APPENDIX

Calculation o f Average Molecular Structural Parameters o f Olefins The 13C NMR spectra of olefin fractions can be divided into regions based on the chemical shifts of double bond carbon atoms Cdb (105--155 ppm) and aliphatic (including naphthenic) carbon atoms C~I (5-65 ppm) (Fig. 1). In samples containing aromatic hydrocarbons, the former region also includes signals of aromatic ring carbons (C~) (Levy and

207

Table AI. Definition of terms appearing in the text, and tables X,r = amount of aromatic hydrocarbons in the olefin fraction X~= amount of olefins in the olefin fractions Cdb= double bond carbon atoms C,j = n- and cyclic alkane carbon atoms C~ = aromatic ring carbon atoms lib = integrated NMR signals of Cdb+ C,~ l~l = integration of all C,msignals I~ = I~b + l~, In a similar manner ltH= Idb 14 + / a ]H+ / a tH ~ab = average degree of substitution of isolated double bonds Kdb= average number of double bonds per olefin molecule Nu = average number of carbon atoms per olefin molecule Nar = average number of carbon atoms per aromatic molecule N~= average number of carbon atoms per alkane molecule

Nelson, 1972). The integrated intensities (areas) of the signal groups have been designated l~b (all signals of Cab and Car), c + laj. c I~c (all signals of Ca) and I ct _- ldb The ~H N M R spectra of the olefin fractions provided three groups of signals (Figs 2 and 3) namely those of aliphatic (and naphthenic) protons H,~ (0.7-2.5ppm), double bond protons Hdb (4.5--5.8 ppm) and aromatic protons H~ (6.6-7.1ppm) following Chamberlain (1977). Their integrated intensities were designated 1~, l~b and I,~ H + lax H _-- ItH. For convenience, respectively, so that I~ + Idb the above terms, and others used in calculating average molecular structural parameters are defined, and listed, in Table A1. We assume that the average number of carbon atoms per aromatic molecule (N~) and carbon atoms per olefin molecule (N~) are similar. If this assumption is incorrect and, for instance, Afaris two times more, or less, than Nu, then the calculated values of Qdb (average degree of substitution of isolated double bonds), Kdb (average number of double bonds per olefin molecule), and N~ will not differ by more than 5% of the measured value). In the ~H NMR spectra all aromatic signals were seen to be upfield from 7.1 ppm, corresponding to protons of di- and multi-substituted benzene rings (Chamberlain, 1977). Hence, there are six aromatic C-atoms and, on average, three aromatic H-atoms per molecule of the contaminating alkylbenzenes in the olefin fractions. Since there are no saturates present, average molecular structural parameters can be calculated as follows: Xar+Xu= 1

6Xar + 2KdbXu = I~bNJl~

(A1)

3Xar = I~N.(H/C)at/Itn and /db /db lar Qdb= 4-- 2 T~ It (H/C)at T6 It - 2 T~ It (H/C),t

(A2)

where (H/C)at is the H/C atomic ratio (about 1.8, from elemental analyses of all the olefin fractions). Qdb was calculated from 13C and IH NMR data using equation (2), but for estimates of N u (or Kdb) and X~ it was necessary to know the values of Kdb (or Nu) respectively. Equation (1) was used to calculate the Kdb values of fractions with a known N u (from Verkhne-Chonsk: Cl0~12, Ct6-CI9 and C24-C27 distillate fractions and from the Nizne-Omrinsk Ct6-CI9 n-alkene fraction. It was shown that Kdb = 1 for all of those fractions (see later). This Kdb value was used for calculating N u values for the total olefin fractions. The structural element content (SE) was calculated by the weight factor procedure described by Smirnov and Smirnov (1985). For spectra of the olefin fractions obtained by NOE (by broadband decoupling) the I F values were calculated as I~jNJ(Nu-2 ). The assignment of signals has been published previously (Smirnov and Smirnov, 1985; Smirnov, 1990a,b). For structural elements with

E. B. FROLOVand M. B. Slvmiuov

208

unknown weight factor values, their contents were estimated from spectra recorded without NOE [with Cr(acac) 3 addition]. If, in that spectrum, the signals of two C-atoms in two structurally similar SEs overlapped, the sum of those SEs was used. In this case their ratio was calculated from the spectrum obtained without Cr(acac)3 addition, on the assumption that Tl and NOE factors would be nearly identical in such structurally similar SEs. These calculations require Nu or N, values for the fractions in question. Assuming the existence of a wide carbonnumber range of olefins in the oils, a more exact determination of the Ns values of the distillate fractions Ct0-Ci2, CI3-C15, C16-Ci9 and so on to C27-C31 were obtained. Saturated, unsaturated and aromatic hydrocarbons were isolated from the three fractions CI0-C,2, Ct6-C19 and C24-C27. The amount of saturated hydrocarbons in other distillate fractions were believed to be equal to the average value taken from two neighbouring fractions. Saturates were isolated from residual fractions (>C27) and their N, values were measured by cryoscopy in benzene. The N s values for saturates boiling above 200°C were calculated by using the equation

Ns =

Zm,lY.(m,IN ~)

(A3)

where N~= Ns values for corresponding distillate (or residue) subfraction of the saturates and mi = mass of this subtraction.

In the 13C NMR spectra, a "hump" of double bond carbon signals covering about 30 ppm can be seen (AJh in Fig. 1). The minimum number of uarrow signals 'n' (that is to say, C-atom resonances of "individual" SEs, containing a double bond) needed to form this "hump" can be evaluated. The unresolved "hump" is formed when the difference between the chemical shift values of neighbouring signals is less than their width at half peak height (Li/2). In olefin spectra recorded without the addition of Cr(acac)3 the LI/2 value is equal to 1.2Hz or less than 0.02ppm. Thus n > nn~n= A6h/LI/2 = 1500. In reality n must have a much greater value than this since it was supposed, in calculating n, that all neighbouring signals had similar intensities. Thus the total amount of different structural elements containing a double bond must be greater than 1000. It has been established that the chemical shift of any carbon atom in an alkyl or alkenyl chain is unaffected by carbon atoms at a greater distance than 5 carbon-carbon bonds (Levy and Nelson, 1972), that is to say, any SE which contains no more than 9 C-atoms in its chain. Basic alkyl structures in petroleums are monomethyl-substituted and isoprenoid (Petrov, 1987). Theoretically, there can exist less than 100 SEs with such structures. Even if all theoretical dimethyl-substituted SEs are taken into account the total amount of such alkenyl SEs does not exceed 300. Therefore the unresolved signals which form a "hump" are, for the most part, signals of olefinic carbon atoms in rings or in the -, ~-, 3'- and 6 positions of alkenyl substituents.