1612

Anal. Chem. 1981, 53, 1612-1620

Chemical Class Separation and Characterization of Organic Compounds in Synthetic Fuels

A separatlon method Is descrlbed for the Identlficatlon of organlc compounds In synthetlc fuel products. Prefractlonatlon of crude synfuel materlals Into discrete chemical classes was performed by adsorption column chromatography uslng small quantities of neutral alumlnum oxlde and slllcic acld. Subsequent high-resolutfon separatlon of lndivldual components was achleved by uslng caplllary column gas chromatography, and speclflc compound types were determlned by gas chromatographlc retentlon data and combined gas chromatography-mass spectrometry. The prlnclpal chemlcal classes lnvestlgated In a solvent-reflned coal llquld were allphatic hydrocarbons, polycycllc aromatlc hydrocarbons, polycycllc aromatlc sulfur heterocycles, nltrogen polycycllc aromatlc compounds, and hydroxyl polycycllc aromatlc hydrocarbons. The nltrogen-contalnlng aromatlc compounds were further separated Into secondary nltrogen polycyclic aromatlc heterocycles, amlno polycycllc aromatlc hydrocarbons, and tertlary nitrogen polycycllc aromatlc heterocycles to facilitate their ldentlflcatlon.

Current emphasis on the development of alternate energy sources has stimulated production of synthetic fuels derived from oil shale, tar sands, and coal. Although technology for producing liquid and solid fuels from these feed stocks has been available since the early 1900s ( I ) , the chemical characterization of these products has recently received increased attention. There are several reasons for determining specific compound types in synfuels. The identification and elimination of specific toxic and carcinogenic compounds would reduce the environmental and occupational health hazards associated with the production and combustion of these materials (2-4). Certain heteroatom species are known to decrease the efficiency of catalytic processes and contribute to the instability of liquid fuels during storage (5,6). Synthetic fuels are also a valuable source of chemical materials for the manufacture of pharmaceuticals, pesticides, herbicides, and dyes (3, 7). Because of the complex nature of synthetic fuel materials, fractionation according to chemical class is usually required before identification of individual components can be achieved. Traditionally, separation schemes have been developed by using solvent partitioning methods and column chromatography. The organic analytical group at the National Bureau of Standards has isolated neutral oils, acids, and bases using a solvent extraction method with subsequent high-performance liquid chromatographic (HPLC) and gas chromatographic analysis to separate and identify individual organic compounds in synfuels (8). The analytical group at Oak Ridge National Laboratory has developed a fractionation scheme using an acid-base extraction followed by solid-liquid chromatography on Sephadex LH-20, silicic acid, and basic alumina (I,9 , I O ) . They reported the separation of aliphatic and aromatic hydrocarbons as well as nitrogen heterocyclic and polar aromatic compounds. Wilson et al. (11)used a similar
'On leave from the Department of Physical Chemistry, The University of Leeds, Leeds LS2 9JT, U.K.

solvent fractionation scheme to analyze coal liquefaction products before and after hydrotreatment. They reported the occurrence of neutral polycyclic aromatic hydrocarbons and sulfur, oxygen, and nitrogen heterocyclic aromatic compounds. In another separation scheme developed for solvent-refined coal (12), nine fractions were isolated from silica gel on the basis of chemical functionality. Schiller and Mathiason (13) reported the development of a chromatographic procedure using neutral alumina to fractionate a wide variety of coalderived solids and liquids for analysis by mass spectrometry and other spectral methods. The principal compound types investigated were saturated hydrocarbons, aromatic hydrocarbons, benzofurans, ethers, nitrogen compounds, and hydroxyl compounds. A combination of ion exchange, coordination, and adsorption chromatography was employed by Jewel1 et al. (14) the separation of petroleum distillates. for This method has also been used to separate coal-derived products into five fractions: acids, bases, neutral nitrogen compounds, saturated hydrocarbons, and aromatic hydrocarbons. This method is commonly known as the SARA technique and has been widely used. Several problems exist with the previously described separation methods. First, most schemes are complicated, time-consuming, tedious, and require highly trained personnel to execute each step with precision and accuracy. Long periods of time are required for pretreatment, cleaning, preparation, and activation of resins and adsorbent packings. Many methods use expensive materials, and large volumes of highpurity solvents are necessary for each chromatographic or partitioning step. Solvent extraction and partitioning of polar types of compounds can cause formation of tars and stable emulsions (15)which contribute to the loss of components and inefficient separations. Finally, many separation schemes are hindered by incomplete chemical class separation and overlap of compound types into adjacent fractions. Improvement of existing separation technology is necessary if faster analysis and better evaluation of the numerous types of synfuel products are to be achieved. In this paper, the development of a two-step separation method using neutral alumina and silicic acid adsorption chromatography is described. This procedure minimizes the problems previously discussed and is rapid, reproducible, economical, and efficient. Crude solid and liquid synfuel products can be fractionated into seven chemical classes by this method. This separation scheme can be completed in less than 6 h and uses less than 500 mL of solvent and approximately 12 g of adsorbent packing. The applicability of this fractionation method is demonstrated by the characterization of a solvent-refined coal (SRC) liquid heavy distillate.

EXPERIMENTAL SECTION
An SRC I1 heavy distillate sample (boiling point range 260450 "C) from the Fort Lewis, WA, pilot plant (using West Virginia coal f o the Pittsburgh seam) was provided by Pacific Northwest rm Laboratory for analysis. This material is of pilot plant origin and should not necessarily be considered representative of commercial

scale production. A schematic diagram of the separation method is shown in Figure 1. First, neutral aluminum oxide (Brockman Activity I, 80-200 mesh, Fisher No. A950) was used as purchased to separate the raw material into four initial fractions: A-1,

0003-2700/81/0353-1612$01.25/0 0 I987 American Chemical Soclety

4ALYTICAL CHEMISTRY, VOL. 53, NO. 11, SEPTEMBER 1981

1813

Hexane

m-7-1
Benzene

INEUTRAL'ALUMINA

i1 SAMPLE

Chloroform

Tetrahydrofuran

.1

s.
(A-2'1

Efhrl Ethanol

.1

1-

(A-1)

q pr/p T - 1 p
I
Sulfur Separation
Method (A- 31
(A-4)

Silicic Acid

Hexanr, Benzene(1 1)

1
Benzene

Benzene Etherd 1)

pGq-qpq
(S-t) (S-2)

1
(S-3)

Figure 1. Chemical class separation scheme for synthetlc fuel products: neutral polycyclic aromatic compounds (PAC), nitrogen polycyclic

aromatic compounds (N-PAC), hydroxyl polycyclic aromatlc hydrocarbons (HPAH), polycyclic aromatic oxygen heterocycles (PAOH), polycyclic aromatic sulfur heterocycles (PASH), secondary nitrogen polycyclic aromatic heterocycles (2'-PANH), amino polycycllc aromatic hydrocarbons (APAH), and tertiary nltrogen polycyclic aromatic heterocycles (3O-PANH). aliphatic hydrocarbons; A-2, neutral polycyclic aromatic compounds (PAC); A-3, nitmlgen polycyclic aromatic compounds (N-PAC); and A-4, hydroxy polycyclic aromatic hydrocarbons (HPAH). [The acronyms u3ed in this paper are assigned according to the format defined by Bsirtle et al. (16).] Approximately 0.1-0.3 g of crude sample was dissoIved in a few milliliters of chloroform and adsorbed onto 3 g of neutral alumina. The solvent was removed from the alumina by vigorously stirring the mixture under a gentle stream of dry nitrogen gas. The alumina with sample was then packed on top of an 11-mm i.d. column which already contained 6 g of neutral alumina. The sample was then eluted with the following chromatographicgrade solvents fraction A-l,20 mL of hexane; fraction A-2,50 mL of benzene; fraction A-3,70 mL of chloroform (containing 0.75% ethanol preservative); fraction A-4,50 mL of 10% ethanol in tetrahydrofuran (containing no inhibitors). Polycyclic aromatic oxygm heterocycles (PAOH) and polycyclic aromatic sulfur heterocycles (PASH) are eluted with the polycyclic aromatic hydrocarbons (PAH) in fraction A-2. Because the PASH are in such low concentration in coal-derived materials, the alumina column procedure was scaled up (100 g of neutral alumina) 0 and 1 g of coal liquid (preadsorbed onto 50 g of neutral alumina) was used to obtain a neutral PAC fraction. The PASH were then separated from the PAH and PAOH by a procedure previously reported by Willey et al. (17). Briefly, the PASH were oxidized with HzOzto their corresponding sulfones, the sulfones were then separated from the other unoxidized neutral PAC on silica gel (fraction A-2a), reduced with LiAlH4 to the original PASH and repurified on silica gel (fraction A-2b). Several distinct c k of nitrogen heterocyclic compounds were present in the initial N-PAC, fraction A-3. Consequently, a method was developed to subfractionate the N-PAC into three additional classes of compounds using silicic acid adsorption chromatography. The alumina N-PAC fraction was adsorbed onto 0.5 g of silicic acid (Mallinckrodt No. 2847,100-mesh powder, used as received from supplier) and the solvent removed by stirring the mixture under dry nitrogen gas. A 22-mm i.d. column was packed with a hexane slurry of 2 g of silicic acid. The silicic acid with the adsorbed sample was then packed on top of the 2-g o column and eluted t give the following fractions and compound types: fraction S-1, secondary nitrogen polycyclic aromatic heterocycles (2"-PANH) with 50 mL of 1:l (vx) hexane:benzene;

fraction S-2, enriched amino polycyclic aromatic hydrocarbons (APAH) with 30 mL of benzene; fraction S-3, tertiary nitrogen polycyclic aromatic heterocycles (3"-PANH) with 50 mL of 1:l (v:v) benzene:anhydrous ethyl ether. Each of the alumina and silicic acid fractions was concentrated to approximately 2 mL on a rotary evaporator, taken to dryness under nitrogen gas, weighed, and dissolved in an appropriate volume of benzene for capillary column gas chromatographic analysis. A Hewlett-Packard 5880 gas chromatograph equipped with a fused-silica capillary column (20 m X 0.30 mm id.) coated fl with SE-52 methylphenyhilicone stationary phase ( i mthickness 0.25 Km) was used to obtain chromatograms of each fraction. The oven was held at 50 "C for 2 min during injection and then temperature programmed at 3 "C/min (alumina fractions) or 4 "C/min (PASH and silicic acid fractions) to a final temperature of 250 "C and held for 5 min. A flame ionization detector (FID) was generally used, but nitrogen and sulfur heterocycle fractions were also analyzed by using a nitrogen-selectivedetector (NPD) and a sulfur-selective detector (FPD). A Perkin-Elmer Sigma 2 gas chromatograph equipped with these detectors was used. Identification of specific compounds was accomplished by using gas chromatographic retention data and combined gas chromatography-mass spectrometry (GC-MS). Gas chromatographic conditions were similiar to those already described. A Hewlett-Packard 5982A quadrupole mass spectrometer was operated in the electron impact mode (70 eV electron energy). Spectra were acquired and processed with a Hewlett-Packard 5934A data system. Infrared spectrometry of the HPAH fraction was performed by evaporating the solvent from the sample and placing an aliquot between two salt plates. The sample was then analyzed by scanning from 4000 to 600 cm-' on a Beckman ACCULAB 2 infrared spectrometer.

RESULTS AND DISCUSSION The separation of oxygen and nitrogen heterocyclic compounds in petroleum products by adsorption chromatography on alumina has been extensively studied by Snyder and Buell (18). Similarly, Schiller (13,19)used alumina chromatography to develop a rapid, simple separation method for heterocyclic compounds in synthetic fuels. The alumina step reported in this paper is a modified version of the method developed by Schiller. It was found in this study that column chromatography on neutral alumina minimized sample loss due to irreversible adsorption, eluted a wide range of solutes rapidly, and provided reproducible class separation according to the functionality of the heteroatom present in the polycyclic aromatic structure. Compounds which contain nitrogen heteroatoms of different functionality can be separated on silicic acid by varying the eluent polarity. Silicic acid has previously been used to separate indoles and carbazoles from other N-PAC in tobacco smoke (20) and synfuels (9). The silicic acid procedure reported in this paper was developed to complement the alumina method by providing an equally simple and rapid method for subsequent chemical class separation of the nitrogen-containing polycyclic aromatic compounds. Using a small quantity of adsorbent (2.5 g) reduces the contact time of the sample components with the silica structure and decreases degradation, artifact formation, and sample loss normally associated with chromatography on silica adsorbents. Numerous compounds are present in each chemical class, even after extensive prefractionation, and further high-resolution separation is required prior to the identification of specific components. The application of capillary column GC-MS to resolve and determine compounds in complex mixtures of PAC has been well established (21). This simple alumina-silica fractionation scheme combined with capillary column GC-MS provides a rapid, effective, and reproducible method for the detailed chemical characterization of synthetic fuels. Additionally, this two-step separation method can easily be scaled up to produce larger quantities of sample which is

1814

ANALYTICAL CHEMISTRY, VOL. 53, NO. 11, SEPTEMBER 1981

~l i'

TEMPERATUREI'CI TIME (mln)

sb
ib
do

do
d0

150

200

250
d0 $ 0

I 0

40

5b

Flgure 2. Capillary column gas chromatogram o aliphatic hydrocarbon fraction A-1. Peak numbers refer to compounds listed in Table 111. Conditions f were 20 m x 0.30 mm fused-silica capillary column coated with SE-52, temperature programmed from 40 O C to 250 O C at 3 OC/min.

TEMPERATUREI'CI TIME (mm.)

sb
0 I
lb

do
20

150

200
I

250

40

50

&O

70

Figure 3. Capillary column gas chromatogram o PAH fraction A-2. Peak numbers refer to compounds listed in Table 111. Condltfons are as f

in Figure 2.

sometimes required for the concentration of trace components and for bioassay to determine mutagenicity or carcinogenicity. The quantitative data obtained by using this procedure are presented in Tables I and 11. A total recovery of greater than 90% was obtained by this method for the separation of an SRC I1 heavy distillate into chemical classes. A chromatogram of each of the fractions is shown in Figures 2-9. Tables 1 1V 1list the compound types identified in each fraction. Aliphatic Hydrocarbons. Due to the nonpolar nature of the aliphatic hydrocarbons, this class of compounds elutes immediately from alumina with hexane. The gas chromatographic retention time of standard octadecane, n-C,,, was used to determine n-C18 in the aliphatic fraction. Mass spectral data supplemented the identification of n-C18and were used to determine the other constituents in the homologous series of straight-chain hydrocarbons. The nonlinearity of the base line in the n-Cls-n-Czoregion (see Figure 2) is caused by the numerous unresolved components (related olefins and branched isomers of the straight-chain parent hydrocarbons). The relatively low concentration of the aliphatic hydrocarbons in this coal liquid illustrates the highly aromatic nature of this product. The structural features of saturates in coal liquids are similar to those of some petroleums and both can be correlated to the distribution of similar compounds found in plant material (22).

Table I. Weight Percent of Neutral Alumina Fractions from an SRC I1 Heavy Distillate Coal Liquid fraction A- 1 A- 2 A-3 A-4 compound type aliphatic hydrocarbons neutral PAC N-PAC HPAH fraction" wt % 5.5 56.6 20.0 14.0 ___ total 96.1 a Average values for three different assays of SRC I1 heavy distillate. Table 11. Weight Percent of Silicic Acid Subfractions from an SRC I1 Heavy Distillate Coal Liquid wt % of heavy wt % fraction compound type of N-PAC distillate s-1 2"-PANH 59.6 11.9 5-2 enriched APAH 8.5 1.7 s-3 3"-PANH 14.9 __ 3.0 total 83.0 16.6

Neutral PAC. Naphthalene and triphenylene standards were used to ascertain the retention parameters of the neutral

ANALYTICAL CHEMISTRY, VOL. 53, NO. 11, SEPTEMBER 1981

1615

I
TEMPERATURE1 C)
TIME f mm )

,
r

50
T------

100

150

200
4 0
50

250

1 0

20

30

Flgure 4. Capillary column gas chromatogram of PASH fraction A-2b. Peak numbers refer to compounds listed in Table IV. Conditions were 20 m X 0.30 mm fused-silica capillary column coated with SE-52, temperature programmed from 40 O C to 250 O C at 4 OC/min.
1214 18 I 1 I2O

TEMPERATUREI'CI TlME(mln)

,
0

s J !

; J a
Ib
10

150

200

2 b

:I0

40

2 0

6b

io

Flgure 5. Capillary column gas chromatogram of N-PAC fraction A-3. Peak numbers refer to compounds listed in Table 111. Conditions were as in Figure 2.

PAC fraction eluted from the alumina column. Benzene was chosen over toluene as tho eluent because it is a slightly more polar solvent and reduced the overlap of the polycyclic furans (PAOH) into the dhloroform fraction. The PAH (see Figure 3) are the chemical class alf highest Concentration in this SRC I1 heavy distillate and generally are major constituents in most coal-derived fuel products. The literature in this area is too extensive to cover here but has recently been reviewed (23, 24). The high de$ee of alkylation in this I'AH fraction is of particular interest. A rough correlation between degree of alkylation of PAH structures in energy-related materials and mutagenicity has been reported by Griest and co-workers (25). Extensive alkylation is observed in all fractions of this SRC I1 liquid, and in some instruces the alkylated species are more abundant than the parent PAC. Although the PASH, Figure 4, comprise! less than 1%of the total heavy distillate, the detailed characterization of the sulfur-containing compounds in synthetic fuel products is important in light of recent evidence that specific sulfur heterocycles have been shown to possess significant mutagenic activity (26,2T). Furthermore, the removal of sulfur from fuels is desirable to prevent the formation of noxious sulfur gases during combustion and to prevent catalyst poisoning in coal

conversion reactors. The identification of specific sulfurcontaining species in synthetic fuels can provide insights into better methods of sulfur removal. N-PAC.The elution of the N-PAC from neutral alumina with chloroform could be visually observed by the migration of a dark band down the column. Analysis by gas chromatography using a nitrogen-selective NPD verified that essentially all of the components in this fraction were nitrogencontaining compounds. Figure 5 is an FID chromatogram of the N-PAC fraction. Primary (APAH), secondary (2'-PANH), and tertiary (3O-PANH) nitrogen heterocycles (see Figures 6-8) occur collectively in the alumina N-PAC fraction. Additional chemical class separation was essential for three reasons. First, the difficulty of identifying individual constituents in this fraction was increased by the fact that more than one compound eluted from the capillary column at the same time. For example, tetrahydrocarbazole coeluted with two benzoquinoline isomers. This was also the case with many of the alkylated carbazoles, alkylated benzoquinolines, and three-ring APAH. Second, different concentration levels between chemical classes obscured groups of compounds which are important to the overall chemical behavior of the sample.

1016

ANALYTICAL CHEMISTRY, VOL. 53, NO. 11, SEPTEMBER 1981

___--

Table 111. Compounds Identified in the Alumina Chromatographic Fractions of an SRC T Heavy Distillate Coal Liquid I peak peak compound compound no. mol wt mol wt no. Aliphatic Hydrocarbon Fraction A-1 (Figure 2) 8 324 1 226 9 338 2 24 0 1 0 352 254 3 11 366 4 26 8 1 2 380 28 2 5 1 3 394 296 6 14 408 310 7
1

2 3 4 5 6 7 8 9
10 11

12 13
1

142 156 170 154 154 168 166 168 168 1841180 182 182 1801182 1981194 184 182 196 178
131

2 3 4 5 6 7
8

9 10
11

12
13

14
15 16

17 18
1

157 145 159 198 185 183 178 179 179 197 1711179 197 167 193 185 193 1811193 148 170 162 184 184 176 198 170 170 190 21 2 1841204 184 184 21 2 182

2
3

4 5 6 7 8
9

10
11

12 13 14 15 16

PAC Fraction A-2 (Figure 3 ) C,-naph thalene 14 196 C,-naph thalene 1 5 204 C,-naphthalene 16 192 biphenyl 17 204 acenaph thene 18 206 dibenzofuran 1 9 202 fluorene 20 202 C,-biphenyl/C,-acenaph thene 21 C, -biphenyl/C,-acenaphthene 22 216 C,-naphthalene/C,-fluorene 23 216 C, -dibenzofuran 24 216 C,-biphenyl 25 230 C,-fluorene/C,-biphenyl 26 228 C,-naphthalene/C,-fluorene 27 228 di benzothiophene 28 242 tetrahydrophenanthrene 29 252 C,-biphenyl 30 252 phenanthrene N-PAC Fraction A-3 (Figure 5) C,-indole 1 9 193/201 C,-quinoline 20 181 C,-indole 21 193 C,-indole 22 207 C,-diphenyl ether 23 195 C,-quinoline 24 201 tetrahydrobenzoquinoline 25 195 phenanthrene 26 195 benzoquinoline 27 209 benzoquinoline 28 203 C,-tetrahydro benzoquinoline 29 203 tetrahydrocarbazolelbenzoquinoline 30 223 C,-tetrahydrobenzoquinoline 31 2171237 carbazole 32 231 C,-benzoquinoline 33 221 C,-tetrahydrocarbazole 34 229 C,-benzoquinoline 35 217 6,-carbazole/C,-benzoquinoline 36 231 HPAH Fraction A-4 (Figure 9) C,-indanol 17 198 hydroxy biphenyl 18 182 C,-indanol 1 9 182 20 198 C, -hy droxy biphenyl C, -hy droxy biphenyl 21 182 22 198 C, -indan ol C,-hydroxy biphenyl 23 2261196 24 212 hydroxy biphenyl hydroxybiphenyl 25 2121196 C,-indaaol 26 220 21 226l210 C,-hy droxy biphenyl C,-hydroxy biphenyl/C,-indanol 28 234 29 2401224 C, -hydroxy biphenyl 30 220 C, -hydroxy biphenyl 31 220 C,-hy droxy biphenyl hydroxy fluorene 32 234

C,-biphenyl/C,-dibenzothiophene

1-phenylnaphthalene C ,-phenanthrene 2-phenylnaph thalene C,-phenanthrene fluoranthene pyrene not identified benzo[a]fluorene benzol b ]fluorene C,-pyrene/C,-fluoranthene
C,pyrene/C,-fluoranthene

benz[a]anthracene chrysene C,-benz [ a ] anthraceneIC, -chrysene benzofluoranthene benzopyrene

C,-carbazole C,-benzoquinoline C,-benzoquinoline C,-car bazole C,-benzoquinoline C,-carbazole C,-carbazole C,-carbazole azafluoran thene azapyrene C,-carbazole C,-azapyreneIC,-azafluoranthene 16,-carbazole C,-azap yrene/C,-azafluoranthene
tetrahydrobenzocarbazole

naph thoquinoline benzocarbazole C,-benzucarbazole C,-hydroxybiphenyl hydroxyfluorene hydroxy fluorene C,-hydroxybiphenyl hydroxy fluorene 6,-hydroxyhiphenyl C,-h ydroxybiphenyl1C ,-hydroxyfluorene C,-h y droxy biphenyl C,-hydroxy biphenyl/C,-hydroxyfluorene naphthylphenol C,-hy droxybiphenyl/C ,-hydroxyfluorene C, -naphthylphenol
C,hydroxybiphenyl/C,-hydroxyfluorene

n aph th ylph enol. naph thylphenol C,-napht,hylphenol ~.

-___-

Although the APAH were lower in concentration than either the secondary or tertiary PANH (see Table II), it has been suggested that the APAH are largely responsible for the biological activity of synfuel products (9,28). Third, detailed assignment of structure was complicated by the presence of isomeric compounds which had different molecular structures and functionalities but produced nearly identical mass spectra. For example, the methylbenzoquinolines and the amino-

anthracenes have molecular weights of 193 amu and all fragment similarly by mass spectrometry making exact structural identification nearly impossible. Therefore, it was necessary to separate the nitrogen-containing compounds into subfractions. The N-PAC constitute a major fraction of not only this coal liquid but also other coal liquefaction products, shale oils, and coal tars (3). Extensive work has been reported on t,he sep-

ANALYTICAL CHEMISTRY, VOL. 53, NO. 11, SEPTEMBER 1981

1617

r-

13

TEMPERATUREI'CI
TIME(m4n)

I 50

io0

150

ZOO

250

io

20

30

4 0

50

Figure 6. Capillary column gas chromatogrami of 2O-PANH fraction S-1. Peak numbers refer to compounds listed in Table V. Conditions were as in Figure 4.
L

TEMPERATUREICI
TlME(mkr 1 )

,
I

i0

100

i50
20

ZOO

250

10

3b

4 0

50

Flgure 7. Capillary column gas chromatogram of enrlched APAH fraction S-2. Peak numbers refer to compounds listed in Table V. Conditlons were as in Figure 4.

Table I V . Compounds Identified in the PASH Subfraction A-2b (Figure 4)of an SRC I1 Heavy Distillate Coal Liquid peak no. mol wt compound peak no. mol wt compound
1
2 166

180
188 184 178

3 4 5 6 7 8

178
198 21 2

fluorene C, -fluorene tetrahy drodibenzothiophene dibenzothiophene phenanthrene anthracene C,-dibenzothiophene C,-dibenzothiophene

9 10

11
12 13

14 15 16

20 2 226 208 20 2 222 234 24 8 26 2

fluoran thene C,-dibenzothiophene phenanthro [4,5-bcd ]thiophene pyrene C,-phenantho [ 4 , 5 - b c d]thiophene benzonaphthothiophene
C,-benzonaphthothiophene C,-benzonaphthothiophene

aration and identificatio'n of several pyridines and anilines; a number of two- and three-ring N-PAC have also been identified in synthetic fuel products (7, 13,19,29-33). Relatively fewer studies havia been reported on the isolation and determination of compounds containing a larger number of rings. The one- and two-ring N-PAC systems are of general

interest because they are present in higher concentration levels, are more volatile, some are commercially useful, and a few are biologically active. However, it is believed that isomers of increasing ring number generally display a higher degree of mutagenicity and carcinogenicity. For example, carbazole is noncarcinogenic, benzo[a]carbazole is moderately

1618

ANALYTICAL CHEMISTRY, VOL. 53, NO. 11, SEPTEMBER 1981

13
r-

TEMPERATURE1 CI TIME (mln.1

,
,
0

50

xx)

lb
20

260

7 250

1 0

30

40

50

Flgure 8. Capillary column gas chromatogram o 3O-PANH fraction S-3. Peak numbers refer to compounds llsted in Table V. Conditions were f as In Figure 4.

Table V. N-PAC Compounds Identified in the Silicic Acid Subfraction of an SRC I1 Heavy Distillate Coal Liquid peak peak no. mol wt no. mol wt compound compound 145 2 159 3 184 4 173 5 198 6 178 7 171 8 212 9 167 10 167
1 1

C,-indole C,-indole C,-diphenyl ether C,-indole C,-diphenyl ether phenanthrene te trahy dr ocarbazole C,-diphenyl ether carbazole carbazole isomer

2"-PANH Fraction S-1(Figure 6 ) 11 167 12 181 1 3 195 14 209 1 5 223 16 237 1 7 217 18 217 1 9 231 20 245

carbazole isomer C,-carbazole C,-carbazole C,-carbazole C,-carbazole C,-carbazole benzocarbazole benzocarbazole C,-benzocarbazole C,-benzocarbazole

2 3 4 5 6 7
8

143 169 183 197

179 167 193 193 207

Enriched APAH Fraction 11 naphthylamine 12 aminobiphenyl 13 C, -aminobiphenyl 14 C,-amino biphenyl aminofluorene 15 16 benzo [ h ]quinoline ' 17 carbazole

9
10

5-2 (Figure 7) 221/207 C,-benzoquinoline/C,-aminophenanthrene 235 C,-benzoquinoline 233 tetrahydronaphthoquinoline 217 aminopyrene/aminofluoranthene 233 tetrahy dronaph thoquinoline 229 naphthoquinoline (possibly benz[c]acridine') 231 C,-aminopyrene/C,-aminofluoranthene 18 229 naphthoquinoline (possibly C,-benzoquinoline benz [ulacridine") benzanthrone aminophenanthrene/aminoanthracene 20 244 230 C,-benzanthrone C,- benzoquinoline 21 phthalate ester

3"-PANH Fraction 5-2 (Figure 8) 8 197 C,-tetrahydroquinoline C,-quinoline 157 9 193 C,-benzoquinoline C,-quinoline 2 171 1 0 201/203 C,-benzquinoline/azafluoranthene 3 185/158/183 C,-quinoline/unidentified peaks 11 221 C,-benzoquinoline C,-quinoline 4 199 tetrahy droquinoline 1 2 203 azapyrene 5 183 1 3 217 C,-azapyrene/C,-azafluoranthene acridine" 6 179 14 229 naph thoquinoline phenanthridine'/benzo[f]quinoline' 7 179 " Possible structures assigned by retention data comparisons with standard compounds. 1

carcinogenic, and dibenzo[cg]carbazole is a potent carinogen (34). Therefore, it is important to isolate and identify the larger compounds in synfuel products. Several two-, three-, and four-ring N-PAC have been identified in the silicic acid subfractions of the alumina N-PAC fraction. The concentration envelope for this SRC I1 distillate cut is comprised

mostly of two-, t,hree-,and faur-ring N-PAC. However, work has been done with standard N-PAC of up to five rings (dibenzacridines and dibenzocarbazoles) and other synfuel products which shows the applicability of this separation method to the isolation of compounds containing more than four rings.

ANALYTICAL CHEMISTRY, VOL. 53, NO. 11, SEPTEMBER 1981

1810

TEMPERbTUREICI TIME (min.)

sb
rb
20

4 0

150

200

250

d o 720 d0 10 f Flgure 9, Capillary column gas chromatogram o HPAH fraction A-4. Peak numbers refer to compounds listed in Table 111. Conditions were as In Flgure 2.
0

The enriched APAH fraction, Figure 7, is of particular interest. Wilson et al. (28) initially suggested the presence of two-, three-, and four-ring primary aromatic amines in a similar SRC I1 heavy distillate. The resulta presented in this paper confirm the presence of these APAH and also report the occurrence of several alkylated species of the parent APAH. The APAH werle not completely separated from the 3 O - P A . Preliminary data suggest that the tertiary nitrogen heterocycles overlapping into the APAH fraction are either alkylated or angular-structured species where the lone pair of electrons on the nitrogen heteroatom is shielded. Therefore, some of the 3O-PANH and the APAH have similar retention characteristics. Many of the 3O-PANH that eluted in this fraction were more than twice as concentrated as the APAH and frequently obscured their presence. Thus, in addition to the retention and GC/MS data used, a derivatization method has been used (:?5) to ensure positive identification of the APAH listed in Table V. Although there was overlap between these two classes of N-PAC compounds in this fraction, the APAH were sufficiently enriched to verify their presence. HPAH. The hydroxyl PAH (see Figure 9) were eluted last from neutral alumina with 10% ethanol in tetrahydrofuran. Many HPAH have been eihown to be present in synthetic fuel products and several methods have been developed for their isolation and determination (8,29,30,36). The major acidic compounds found in synfuels are phenol and the alkylphenols. Also, HPAH such as hydroxybiphenyls, indanols, naphthols, acenaphthenols, and hydroxyfluorenes have been reported to occur in these matrices (29). Infrared spectrometry of the HPAH frriction showed the strong presence of hydroxyl functional groups and indicated that there was little if any carbonyl oxygen. This observation agrees well with others who have reported the absence of carbonyl groups in coal liquefaction produck (29). The major components identified by GC-MS in this fraction were indanols, hydroxybiphenyla, hydroxyfluorenes, and naphthylphenols. Low levels of nalphthols were also detected. As with the other fractions of this coal liquid, a high alkyl content w s a observed. The projected increase in use of synthetic fuels produced from coal, shale, and tar sands to meet growing energy demands has created a need jfor faster and more reliable methods of analysis of products and byproducts from these processes. The methodology presented in this paper and applied to the analysis of a representative solvent-refined coal liquid has been shown to be a rapid, effective, and reproducible means to a complete chemical analysis of these materials. Although not discussed, this separation method has been applied to the

analysis of other SRC I1 products, coal tar, SRC recycle solvents, and SRC I process materials. In each case, the compounds in these materials were successfully separated according to chemical functionality as described for the SRC I1 heavy distillate.

ACKNOWLEDGMENT The authors thank Robert M. Campbell for his technical

LITERATURE CITED
Rubln, I. B.; Guerln, M. R.; Hardlgree, A. A.; Epler, J. L. fwlron. Res. 1976. 12. 358-365. Sexton, d. J. Arch. Enwiron. Health 1980, 1 , 208-231. Dong, M.; Schmeltz, I.; LaVole. E.; Hoffmann. D. In Carclnogenesls, vol. 3: Polynuclear Aromatic Hydrocarbons; Jones, P. W., Fruedenthal, R. I., Eds.; Raven Press: New York, 1978; pp 98-108. Koppenhal, D. W.; Manahan, S. E. Enwlron. Sci. Techno/. 1978, 10, 1104-1107. Drushel, H. V.; Sommers, A. L. Anal. Chem. 1988, 3 , 19-28. Hartung, 0. K.;Jewell, D. M. Anel. Chlm. Acta 1982, 26, 514-547. White, C. M.; Schwelghardt, F. K.; Schuitz, J. L. Fuel Process. Techn ~ l 1978, 1 , 209-215. . Hertz, H. S.; Brown, J. M.; Chesler, S. N.; Guenther, F. R.; Hllpert, L. W.; May, W. E.; Parris, R. M.; Wise, S. A. Anal. Chem. 1980, 52, 1650-1657. Ho, C.-H.; Clark, B. R.; Guerln, M. R.; Ma, C. Y.; Rao, T. K. Prepr. Pap., Am. Chem. Soc., Diw. Fuel Chem. 1979, 24 (I), 281-291. Jones, A. R.; Guerin, M. R.; Clark, B. R. Anal. Chem. 1977, 4S, 1766-1 771. Wllson, 8. W.; Peterson, M. R.; Pelroy, R. A,; Cresto, J. T. Fuel 1981, 60, 269-294. Farcaslu, M. Fuel 1977, 56 (I), 9-14. Schiller, J. E.; Mathlason, D. R. Anal. Chem. 1977, 49, 1225-1228. Jewel, D. M.; Weber, J. H.; Bunger, J. W.; Plancher, H.; Latham, D. R. Anal. Chem. 1972, 44, 1391-1395. Clark, B. R.; Ho, C. H.; Jones, A. R. A&. Ctmm. Ser. 1978, No. 170, 282-294. Bartle, K. D.; Lee, M. L.; Wise, S. A. Chem. Soc. Rev. 1981, 35, 113-158. Wllley, C.; Iwao, M.; Castle, R. N.; Lee, M.L. Anal. Chem. 1981, 53, 400-407. Snyder, L. R.; Buell, B. E. Anal. Chem. 1988, 40, 1295-1302. Schlller, J. E. Anal. Chem. 1977, 49, 2292-2294. Snook, M. E.; Arrendale, R. F.; Higman, H. C.; Chortyk, 0. T. Anal. Chem. 1978, 50, 88-90. Lee, M. L.; Wright, B. W. J . Chromtogr. Scl. 1980, 18, 345-357. White, C. M.; Shultz, J. L.; Sharkey, A. G. Nature(London)1977, 268, 620-623. Guerln. M. R. In Polycycllc Hydrocarbons and Cancer; Gelboln, H. V., TsO, P. 0. P., Eds.; Academic Press: New York, 1978; Part I. White, C. M. In Handbook of Polycycllc Aromatlc Hydrocarbons; Bjarseth, A., Ed.; Marcel Dekker: New York, In press. Grlest, W. H.; Tomklns, B. A.; Epler, J. L.; Rao, T. K. In Poiynuclear Aromatic Hydrocarbons; Jones, P. W., Leber, P., Eds.; Ann Arbor Sclence Publishers: Ann Arbor, MI, 1979; pp 395-409. Karcher, W.; Nelen R.; Depaus, R.; van Eljk, J. In Proceedlngs of 5th International Symposium on Polynuclear Aromatic Hydrocarbons; Battelb Columbus Laboratories: Columbus, OH, 1980. Pelrov. R. A.. Pacific Northwest Laboratorv. Richland. WA.. 1981. un-. . publiihed work. Wllson, B. W.; Pelroy, R. A.; Cresto, J. T. Mutat. Res. 1980, 79, 193-202.

1620

Anal. Chem. 1981, 53, 1620-1627

(29) Schwelghardt, F. K.; White, C. M.; Friedman, S.; Schukz, J. L. Organic Chemistry of Coal; American Chemical Society: Washington, DC, 1978 pp 240-257. (30) Uden, P. C.; Carpenter, A. P.;Hackett, H. M.; Henderson, D. E.; Siggla, S. Anal. Chem. 1970. 51. 38-43. (31) Tesarlk, K.; Ghyczy, S. J.Chmmatogr. 1874, 91, 723-731. (32) Paudler, W. W.; Cheplen, M. Fuel 1979, 58, 775-778. (33) Brown, D.; Earnshaw, D. G.; McDonald. F. R.; Jensen, H. E. Anal. Chem. 1970, 42, 146-151. (34) Lee, M. L.; Novotny, M. V.; Bartle, K. D. AnalyticalChemistry of Polycyclic Aromatic Compounds; Academic Press: New York, 1981; A p (35) Later, D. W.; Lee, M. L., Brigham Young University, Chemistry Departaendlx 5 -.

(36) Schabron, J. F.; Hurtubise, R. J.; Silver, H. F. Anal. Chem, 1978, 51,
1426- 1433.

RECEIVED review April 1, 1981. Accepted June 1, 1981. for This work was supported by the Department of Energy, Contract No. B-B4843AV, with Pacific Northwest Laboratory, and the Department of Energy, Division of Biomedical and Environmental Research, Contract No. DE-AC02-79EV10237, with Brigham Young University.

Distribution Processes of Inorganic Solutes in Gel Chromatography

Department of Chemistry, Faculty of Science, Tokyo Metropolitan University, Fukaza wa, Setagaya-ku, Tokyo 158, Japan

Rokuro Kuroda
LBboratory for Analytical Chemistry, Faculty of Engineering, University of Chlba, Yayoi-cho, Chlba 260, Japan

A novel concept regarding the separation mechanism has been presented In Inorganic gel chromatography In aqueous media. The distribution of Ionic species Is taken to be a combined effect of size exciuslon and partltlon, the latter Involving ail of the operative factors other than the exclusion effect. The partition depends strongly on the type of the counterion and the cdon in the eluant. Theoretical resuns are In good agreement wlth experimental data, and R , values of ionic species can be estimated semlempirlcally on the basis of the theoretical approach presented.

Gel chromatography is a form of liquid chromatography for separating solute compounds primarily according to their size differences, and this technique has been applied extensively in biochemistry and organic polymer chemistry, In the past decade, gel chromatography of inorganic compounds has also been investigated (1). There are many reports on the gel chromatographic behavior of metal ions and simple anions in aqueous media mainly on highly cross-linked gels with small pore sizes. Most investigators have attempted to explain the gel chromatographic behavior of these small ions in terms of steric exclusion although other secondary effects, such as adsorption, are assumed to be involved in the separation process. The distribution coefficients and ionic sizes have been found well correlated in some cases, in agreement with predicted values based on the size exclusion effect (2-4). However, the gel chromatographic behavior of ionic substances has not been found to be explained exclusively with the exclusion mechanism. There are few papers, however, which succeeded in explaining the separation mechanism in a consistent manner in inorganic gel chromatography. Ogata et al. (5) have shown that the elution volume of magnesium ion on Sephadex G-15 column depends strongly on the type of counteranion in the eluant, the elution volume increasing in the order sulfate < chloride < nitrate < perPresent address: St. Marianna University School of Medicine, Sugao, Takatsu-ku, Kawasaki 213, Japan.
0003-2700/61/0353-1620$01.25/0

chlorate. They also reported (5,6)that this order corresponds to the order of the elution volumes of these counteranions when eluted with 0.1 M sodium chloride solution. Tarutani et al. (7) have observed a similar anion effect on the chromatographic behavior of some divalent metal ions. On the other hand, the effect of the type of coion, Le., the ion with the same charge as the sample ion, has so far scarcely been investigated. We felt that such counterion and coion effecta should reflect the separation mechanism inherent in gel chromatography of ionic substances and attempted to solve the separation mechanism in this paper by assuming the distribution of the solute to be effected as a consequence of combined size exclusion and partition which involves all factors other than exclusion effect. The counterion and coion effects in gel chromatography of ionic solutes will be understood on the basis of the mechanism proposed.

THEORETICAL SECTION
If a solute compound is introduced into a system consisting of liquid and gel phases, and then the system is allowed to come to equilibrium, the distribution coefficient can be expressed by eq 1, provided only the size exclusion effect is

K, = V,,/Vi
responsible for the distribution, where K. is the distribution coefficient relating to the size exclusion mechanism, Vi is the volume of the internal solvent of the gel phase, and V, is the nonexcluded volume in gel phase. In such a case, the system can be regarded as a volume controlled partitioning with concentration in the available part of the gel phase identical with that in the external solution phase (8). In the exclusion mechanism, the gel acts as an inert matrix, holding the solvent in its pores. However, it seems likely that the properties of the solvent in gel differ from those of bulk of the solvent by the possible interaction between the solvent molecule and the gel matrix. For instance, water molecules i i ~ macromolecular gels generally exhibit physical properties distinct from those of ordinary free water by the interaction with hydrophilic groups or hydro@ 1981 American Chemical Soclety