which is many times greater than the vapor sampling rate of instruments A and C. For instrument B, the vapor sampling rate was approximately equal to the flow rate from the generator only for a few seconds at the start of the sampling cycle. This, however, was not considered to present a problem because of the relatively long sampling time of this instrument. It was concluded that this trace vapor generator can be used effectively to determine the explosives vapor concentration for which a detector alarm is triggered, or to determine the explosives vapor concentration for which the detector response is just recognizable. ACKNOWLEDGMENT The author thanks Edwin C. Kuehner and Stephen Cheder for their assistance, and Lorne Elias of the National Research Council, Ottawa, Canada, for his helpful suggestions. LITERATURE C I T E D (1) J. W. Harrison, "Comparative Evaluation of Trace Gas Technology", Vol. Il-Analysis and Evaluation of instrumental Methods, Contract DAAKO273-C-0128, Research Triangle Park, N.C. 27709 (1973). (2) W. A. Wall and H. M. Gage, Technical Memorandum 74-14, U.S. Army Land Warfare Laboratory, Aberdeen Proving Ground, Md. 21005 (AD-921744). (3) G. E. Spangler, Report 2089, USAMERDC, Attn: STSFB-XR, Fort Belvoir, Va. 22060. (4) W. A. Wall and H. M. Gage, Technical Memorandum 73-02, U.S. Army Land Warfare Laboratory Aberdeen Proving Ground, Md. 21005 (AD-921867).
(5)G. E. Spangler, Report 2083, USAMERDC, Fort Belvoir, Va. 22060. (6) W. A. Wall and H. M. Gage, Technical Report No. 74-13, U S . Army Land Warfare Laboratory, Aberdeen Proving Ground, Md. 21005 (AD-921743). (7) W. A. Wall, H. M. Gage, and H. T. Reiiiy, "Calibratlon of Effluvia Detectors, Techniques and Results," Technlcal Report 74-85, USALWL AD-922270L, Aberdeen Proving Ground, Md. 21005. (8) B. W. Liebel and R. M. Roberts, "Final Report of Trace Gas Acquisition System (TGAS)," Analytical Research Laboratories. Inc., Monrovia, Calif. 91016, Contra:t No. DAAK02-70-C-0644. (9) A. Dravnieks, Bomb Detection System Study," IlTRl Technical Report No. ADS-81 on FAA Contract No. FA 6 s WA-1200 (1966). (10) E. E. Hughes, W. D. Dorko, E. P. Scheide, L. C. Hall, A. L. Bellby, andJ. K. Taylor, Gas Generation Systems for the Evaluation of Gas Detecting Devices, National Bureau of Standards internal Report No. 73-292 (1973). (11) R. C. Weast and S. M. Selby, "Handbook of Chemistry and Physics", 48th ed., Chemical Rubber Co., Cleveland, Ohio, 1967-68. (12) W. C. McCrone and J. H. Andreen, Anal. Chem., 28, 1997 (1954). (13) J. D. Brandner, J. Ind. Eng. Chem. 30, 681 (1938). (14) B. Greifer, B. C. Cadoff. J. Wing, and J. K. Taylor, Development of Solid State Samplers for Work Atmospheres, National Bureau of Standards Internal Report No. 74-527 (1974). (15) To be published in J. Chem. Thermodynam.
RECEIVEDfor review May 6,1976. Accepted June 16,1976. Work supported by The National Institute of Law Enforcement, Department of Justice. Certain commercial equipment, instruments, or materials are identified in this paper in order to adequately specify the experimental procedure. In no case does such identification imply recommendation or endorsement by the National Bureau of Standards, nor does it imply that the material or equipment identified is necessarily the best available for the purpose.
Comparison of Various Structural Analyses for Pitch Fractions Yoshio Yamada," Takeshi Furuta, and Yuzo Sanada' National Research Institute for Pollution and Resources, Ka waguchi, Saitama, Japan 332
Comparison of structural parameters calculated by denslmetric, NMR, x-ray dlffraction, and computer methods has been made for solvent fractions from pitches. New structural parameters have been proposed by the combination of Brown-Ladner's NMR method and that of Diamond's x-ray diffraction. The results obtained by the combination method were in good agreement wlth those obtained by the computer method. The fact implies that the both structural analyses are available for the structural analysis of such carbonaceous materials as pitch and coal extracts. It was found, moreover, that the calculated parameters provide useful Information not only on the chemical structure but also on the solubillty of pitch.
The structural analysis of coal was first studied by Van Krevelen et al. ( l )on , the basis of additivity of specific volume. Thereafter, a number of analyses have been carried out for coal extracts, petroleum fractions, and so on, by means of many experimental techniques (2-5). Recently, Hirsch and Altgelt ( 6 ) developed a method for estimation of many structural factors of petroleum heavy ends by using a computer. This procedure was modified by Katayama et al. (7) and successfully applied to tar-pitches with highly condensed aromatic components and to others containing naphthene rings or aromatic components alone. The computer method for obtaining structural parameters uses four kinds of experimental values, viz., elemental analysis, molecular weight, Present address, F a c u l t y o f Engineering, H o k k a i d o University, H o k k a i d o , Japan 060.
density, and NMR data. Such parameters as aromaticity can be calculated also by Van Krevelen's and Brown-Ladner's method ( I , 2 ) , but it is impossible to evaluate the number of structural units or aromatic rings per molecule without assumptions relating to the compactness of fused rings (810). The purpose of this investigation is to calculate the structural parameters by the computer method and to compare them with the results obtained by a combination of the NMR method and x-ray analysis termed Diamond's method (11, 12). EXPERIMENTAL F o u r k i n d s o f coal t a r - a n d petroleum-pitches were used in t h i s study. T h e samples were extracted w i t h benzene in a Soxhlet apparatus, a n d t h e n t h e extracts were separated i n t o t w o fractions o w i n g t o solubility in e t h y l ether; t h e soluble fraction is referred t o as A, a n d t h e insoluble fraction as B. T h e results o f elemental analysis for these fractions are indicated in Table I, a n d molecular weight, density, a n d hydrogen d i s t r i b u t i o n from NMR d a t a are summarized in T a b l e 11. T h e experimental procedure for these d a t a i s described in d e t a i l elsewhere ( 1 3 ) .
RESULTS AND DISCUSSION Definition of Terms. The following terms are used in the subsequent calculations. FC = fraction of aromatic carbon in fused rings per molecule. 4 = ring compactness factor described by Hirsch et al. (6). FB = ratio of attachments of aliphatic chains to peripheral ring carbons.
ANALYTICAL CHEMISTRY, VOL. 48, NO. 11, SEPTEMBER 1976
1637
~~
~
Table I. Analytical Data of Pitch Fractions Sample d.a.f., % Pitch No.
Ethylene tar-pitch
1 1
Coal tar-pitch
2 2 3 3 4 4
Petroleum pitch Petroleum pitch
Fraction
C
H
A B A B A B A B
93.70 93.13 92.24 92.07 94.55 93.79 94.67 94.42
6.15 5.12 4.91 3.97 5.28 4.39 5.26 4.31
Table 11. Results of Molecular Weight, Density, and Hydrogen Distribution
1 1 2 2 3 3 4 4
A B A B A B A B
570 2360 290 7 30 390 1350 540 1510
1.180 1.213 1.236 1.304 1.247 1.266 1.230 1.262
Hydrogen distribution %a HAR
HAL
HBL
HL
42.2 51.3
37.9 31.0 17.0 11.9 23.3 15.9 22.9 14.5
19.9 17.7 3.0 2.0
0.0 0.0 0.0 0.0 0.0 0.0
80.0
86.1 68.6 79.5 68.8 80.5
8.1
4.6 8.3 5.0
0.0 0.0
a HAR = hydrogen on aromatic rings. HAL = benzylic hydrogen. HBL = -CH and -CH, hydrogen on aliphatic car-
bon atom and -CH, hydrogen on aliphatic carbon atom p to an aromatic carbon. HL = -CH, hydrogen on aliphatic carbon atom y or further to an aromatic carbon.
FD = ratio of benzylic carbon atoms to total aromatic carbons. C = total number of carbon atoms. H = total number of hydrogen atoms. M = number of structural units per molecule. CA = aromatic carbon atoms per molecule. CA, = aromatic carbon atoms per structural unit. VRS = molecular volume occupied by aliphatic chains. VRR = molecular volume occupied by fused rings. E = overlap volume correction. V = molecular volume adjusted from heteroatoms. R = total number of rings per molecule. RA = number of aromatic rings per molecule. RA, = number of aromatic rings per structural unit. R N = number of naphthene rings per molecule. fa = fractions of aromatic carbon per molecule. HAR = number of hydrogens on aromatic rings. C P T = total number of peripheral aromatic and naphthenic carbon atoms. CIT, = number of internal carbon atoms per structural unit. CAL = number of benzylic carbon atoms per molecule. L = number of attachments of aliphatic chain to rings. La = average size of aromatic ring systems. = number of carbon atoms existing in layer planes containing aromatic and benzylic carbon. Computer Analysis. Hirsch e t al. (6) have used independently two types of aliphatic hydrogen to serve as input data. One is the number of benzylic -CH2 hydrogens and the other is that of the benzylic -CHs hydrogens. Chemical shifts of the two types mentioned above are experimentally indistinguishable from the NMR spectrum. Hence, the key of our calculations is principally based upon the method modified by Katayama et al. (7). In order to reduce the assumptions, four floating parameters, FC, 4, FB, and FD, which vary
m
1638
Diff., 0
S
0 0 0
0.06 0.05 0.40 0.26 0.12 0.14 0.07
0.09 1.70 1.55 2.74 0.33 1.68
0
0.10
1.17
0 0
0.90 0.96
0
Table 111. Structural Parameters Calculated by t h e Computer Method
Sample
Av Pitch Frac- molecular Density d,*' N o . tion wt
N
Sample
FC
M
CA
RA
1A 1B 3A 3B 4A 4B
0.80 1.oo
1.3 6.0 1.5 3.9 1.2 4.0
34.1 159.0 26.9 99.3 37.6 112.0
8.2 44.8 6.4 27.8 8.8
0.88 1.00 0.88 1.00
31.1
RN
CAL
fa
2.7
5.5 15.9 2.7 4.0 3.0 4.5
0.77 0.87 0.88 0.94 0.89 0.95
0.0
1.6 0.0
1.7 0.0
within narrow ranges, are introduced into mathematical relations among structural parameters. One equation containing two variables, M and CA, is derived from the ring compactness relationship (6, 7); F1(CAyM) = (2 - CA) + M(2 - 0*4866' '$1 4(5.9948 M(S1 - CA) - 5.2593 M2}1'2 - CA/FC = 0
+
(1)
where SI is the saturation index: SI = 2 - C
+2 -H
(2)
On the basis of density-structure correlation, Katayama et al. ( 7 )gave a set of equations, by modifying the relations derived by Hirsch (14). A general formula summarizing these equations is F2(CA,M) = VRS
+ VRR - E - V = 0
(3)
The two simultaneous equations were first solved approximately by using the Newton-Raphson method (15), the floating parameters being altered a t regular intervals. At the same time, the number of naphthenic rings can be estimated from a rough approximation which was found by Katayama e t al. ( 7 ) , RA/R
HAR/(CPT - L)
(4)
In this manner, the numbers of naphthene rings (RN), benzylic carbon atoms (CAL), and aromatic rings (RA) per molecule were determined from Equation 4 and the values of M and CA. A set of solutions from experimental data satisfying the structural restrictions of the program are listed in Table 111. No results were found by this method for samples 2A and 2B, respectively. Combination Analysis of t h e N M R Method a n d Diamond's X-ray Diffraction Method. According to BrownLadner's method ( 2 ) , CA* and CAL* are calculated by assuming that the atomic ratio of hydrogen to carbon for the a-carbon groups and that for the other carbon groups are 2, respectively. Assuming the model of molecular structure for pitch fractions as a polynuclear-polycondensed aromatic ring system, the number of ring systems (that is to say, the number of structural units per molecule) is one of the important parameters. M* was evaluated as follows: La is calculated by analyzing
ANALYTICAL CHEMISTRY, VOL. 48, NO. 11, SEPTEMBER 1976
Table IV. Structural Parameters Calculated by Combination of Brown-Ladner's and Diamond's Methods CA * RA* CAL* fa*a fa**a Sample M* 0.77 0.78 8.8 6.6 34.4 1A 2.0 154.0 43.0 18.6 0.84 0.88 1B 7.3 1.2 0.94 0.88 5.3 2A 1.1 20.8 54.0 14.7 1.7 0.96 0.94 2B 1.5 2.4 0.90 0.87 27.5 6.9 3A 1.5 27.7 4.7 0.94 0.93 100.0 3B 3.8 9.9 3.3 0.90 0.85 38.1 4A 1.9 0.95 113.0 31.3 4.7 0.95 4B 3.4 a fa* obtained by Brown-Ladner's NMR method. fa** obtained by Van Krevelen's densimetric method.
the [ll]band of an x-ray diffraction pattern. The correlation between La and N is shown as e a = 2.5
6
(5)
CA,* is given by CA,* = N X CA*/(CA*
+ CAL*)
(6)
CA*/CA,* equals the number of structural units per molecule. CA,* generally relates to CIT,* by: CIT,* = CA,*(l - Hau/CA*)
(7)
where Hau/CA* is the atomic hydrogen to carbon ratio of the hypothetical unsubstituted aromatic molecule and evaluated by Brown-Ladner's method ( 2 ) .Since the following relation holds between CIT,* and RA,*: RA,* = (CIT,*
+ 2)/2
(8)
RA* can be obtained by multiplying RA,* with M*. T h e calculated values are summarized in Table IV. Comparison of the Structural Analyses. The fractions of aromatic carbon (fa*, fa**)obtained by Van Krevelen's and Brown-Ladner's methods ( I , 2) are comparable to the fa values in Table 111. It may seem reasonable that the values of fa, CA, and CAL are very close to those of the corresponding fa*, CA*, and CAL*, because input data employed are the same. It should be noted, however, that the resultant values of M* obtained with the combined NMR and x-ray diffraction methods are also very close to those of M. T h e result means that both the computer- and the combination-method are available for the structural analysis of pitches. When the input data are determined experimentally in advance, the computer method gives the various structural
parameters in a short time, say, less than 2-3 min. It should be noted, as pointed out by Hirsch et al. ( 6 ) , that approximation using the way of Newton-Raphson may yield a set of meaningless values, depending on the initial values. Relationship between S t r u c t u r a l P a r a m e t e r s a n d Physical P r o p e r t i e s of Pitches. The number of structural units is one of the important factors governing not only the chemical structure but the characteristics of pitch. It was shown in the previous paper (23)that the number of structural units relates closely to the flatness and mobility of pitch molecules in the liquid state and influences upon the texture of the mesophase formed when the sample is carbonized a t about 430 OC. Fraction B is composed of more aromatic carbons, rings, and structural units than fraction A, as can be seen from Tables I11 and IV. This can be easily interpreted in terms of their different solubility in ethyl ether. In addition, the more definite evidence appears in the floating parameters, FC and RN. In conclusion, it was found that the structural parameters calculated by the different methods give useful information about the chemical structure and make possible a comprehensive understanding of certain characteristics of pitch fractions. ACKNOWLEDGMENT The authors are grateful to U. Katayama for his valuable discussion on the computer method. LITERATURE CITED (1)
D. W. Van Kreveien and H. A. G. Chermin, Fuel (London), 33, 79
(1954). (2) J. K. Brown and W. R. Ladner, Fuel(London), 39, 87 (1960). (3) R. B.Williams and N. F. Chamberlain, Proc. 6th World Petroleum Congress, Frankfurt, 1963, V. No. 17, p 217. (4) J. G. Speight, Fuel(London), 49, 76 (1970). (5) J. P. Dickie and F. T. Yen, Anal. Cbem., 39, 1847 (1967). (6) E. Hirsch and K . H. Altgelt, Anal. Cbem., 42, 1330 (1970). (7) U. Katayama, T. Hosoi, and G. Takeya, Nippon Kagaku Kaishi (J. Cbem. Soc. Jpn, Chem. lnd. Cbem.), 1975, 127. ( 8 ) G. A. Haley, Anal. Chem., 43, 371 (1971). (9) L. H. Ali, Fuel(London), 50, 298 (1971). (IO) S. Yokoyama, N. Ounisi, and G. Takeya, J. Fuel Soc. Jpn, 52, 906 (1973). (11) R . Diamond, Acta Crystallogr., 11, 129 (1958). (12) M. Shiraishi and K. Kobayashi, Nippon Kagaku Kaisbi(J. Cbem. SOC.Jpn, Cbem. Ind. Cbem.), 1972, 1135. (13) T. Furuta, M. Shiraishi, and Y. Sanada, Seklyu Gakkaisbi(J. Jpn Pet. Inst.), 18, 1086 (1975). (14) E. Hirsch, Anal. Cbem., 42, 1326 (1970). (15) H. Margenau and G. M. Murphy, "The Mathematics of Physics and Chemistry", Van Nostrand, New York.1943, pp 475 ff.
RECEIVEDfor review April 1, 1976. Accepted June 4,1976.
ANALYTICAL CHEMISTRY, VOL. 48, NO. 11, SEPTEMBER 1976
1639
