along with the samples, ensures that there is no loss of mercury from the sample to the surfaces of the irradiation container. The heavy duty polyethylene bags specially chosen and used to prepare the sample containers did not present problems of cross contamination similar to those reported by Bate ( I 7) Periodic blank determinations were made to determine the mercury levels of the polyethylene bag. None of the polyethylene sheets we have used so far showed any detectable amount of mercury. The alternate procedure to precipitate mercury as mercuric oxide suggested here works well only if adequate precautions are taken to ensure that the mercuric sulfide is free from other impurities prior to dissolving it in aqua regia. Other methods of precipitating mercury by a variety of reagents are suggested by Roesmer (5). The repeated counting of the gold foils was done to ensure that there was no uneven absorption of the low energy emissions from the deposited mercury because of preferential deposition on one side of the foil. We have observed two such foils during the counting of over 300 foils, and this was recognized as being due to the foil clinging to the side of the beaker during electrolysis. The analytical procedures described here were compared with other techniques used for the determination of mercury in biological and environmental samples. The results of two interlaboratory comparison studies using fish homogenates and soil samples are summarized in Tables VI1 and VIII. The fish tissues analyzed contained only natural forms of mercury and the results of their analyses reflect the problems involved in analyzing these samples. The accuracy of the analytical procedures detailed here was determined by radioactive mercury (HgZ+ form) tracers. (17) L. C. Bate, Rudioclirm. Ror/ioa/ra/. L e / / . ,6 (3), 139 (1971).
These tracer studies showed that the errors of this procedure were less than 15% a t 0.01 parts per million level and less than 5 at 2 parts per million level of mercury in biological tissues. The precision of the analysis as determined by repeat analysis of fish samples and sediment samples containing natural forms of mercury showed a standard deviation of less than 5 % a t 5-ppm levels, less than 7 x at 1.5-ppm levels, and less than 17% at 0.01-ppm ievels. The results of our analysis identified in Tables VI1 and VI11 have this precision and probably the same accuracy. Although the procedures described here can yield very reliable mercury values, it is recognized that the art of analysis still plays a significant role as is clearly evidenced from the distribution of results obtained using a particular kind of analytical procedure (Tables VI1 and VIII). With the recognized elusive nature of mercury, it may be necessary to take all the precautions mentioned here and probably more, to accurately determine the mercury content of environmental and biological samples. ACKNOWLEDGMENT
The authors wish to gratefully acknowledge the assistance of Dr. Wilbur L. Hartman, Bureau of Commercial Fisheries, U. S. Department of Interior, Sandusky, Ohio, and Dr. Robert A. Sweeney, Director, Great Lakes Laboratory, State University College at Buffalo, for their assistance in obtaining various samples used during this investigation.
RECEIVED for review March 18, 1971. Accepted May 11, 1971. Presented at the 161st National Meeting of the American Chemical Society, Los Angeles, Calif., March 1971. This investigation was supported by the Bureau of Sport Fisheries and Wildlife, U. S. Department of Interior.
Low-Resolution Mass Spectrometric Determination of Aromatics and Saturates in Petroleum Fractions C . J. Robinson Research and Decekpment Department, American Oil Company, Whiting, Ind. 46394 A new mass spectrometric procedure for determining up to 25 saturated and aromatic compound types in petroleum fractions eliminates the need for a variety of methods to cover wide ranges of boiling points and composition. A base-line technique to resolve the mass spectrum into saturates and aromatics spectra permits analysis of any sample boiling within the range of 200-1100 O F without need for a physical separation or a high-resolution mass spectrum. The entire composition i s accounted for in terms of 4 saturated hydrocarbon types, 12 aromatic hydrocarbon types, 3 thiopheno types, and 6 unidentified aromatic groups. The procedure gives reasonable and consistent results on a wide variety of samples. FOR SEVERAL YEARS low-resolution mass spectrometry has been used to determine hydrocarbon types in petroleum naphthas and middle distillates. The greater complexity of the mixtures boiling in the gas-oil range has ordinarily prevented direct determination of types; instead, the samples are usually separated into saturates and aromatics fractions, which are then examined mass spectrornetrically.
To realize the obvious savings in time and money by avoiding such physical separations, Gallegos e f a / . ( I ) have used the resolving power of high-resolution mass spectrometry to obtain separate spectra for saturates and aromatics. In addition, Ferguson and Howard (2) have described a lowresolution method to determine 17 hydrocarbon and sulfur types in 400-1000 O F gas oil without prior separation, but they have not published a detailed procedure. We have now developed a procedure to determine up to 4 saturated and 21 aromatic compound types in gas oils by using only the low-resolution mass spectrum and the number average molecular weight of the unseparated sample. It is an extension of our previously published determination of aromatics in petroleum fractions ( 3 ) and therefore includes ( I ) E. J. Gallegos, J . W. Green, L. P. Lindeman, R. L. LeTourneau, and R. M. Teeter, ANAL.CHEM., 39, 1833 (1967). (2) W. C. Ferguson and H. E. Howard, Thirteenth Annual Con-
ference on Mass Spectrometry and Allied Topics, St. Louis, Mo., May 16-21, 1965. (3) C. J. Robinson and G. L. Cook, ANAL.CHEM., 41, 1548 (1969).
ANALYTICAL CHEMISTRY, VOL. 43, NO. 11, SEPTEMBER 1971
1425
Table I. Low Mass Patterns of Saturates 1
L
Step 3
1 .oo 0.90 0.44 0.27
1 .oo 0.52 0.32
1.oo 0.63
‘Z
series +1
0
-1
-2
-3 -4
-5
Massb 430 57 71 85 99d 56 7OC 84 55c 69 83 97 111 125 68 82c 96 110 124 67 8lC 95 80 9dc 108 122 136 79 93c 107 121 135 149
1.oo 0.69 0.38 1 .oo 0.69 0.51 0.43 0.24 0.13 1.oo 1.40 1.10 0.66 0.43 1 .oo 1.25 1.17 1 .oo 1 .oo 0.90 0.75 0.63 1.oo 0.86 0.64 0.53 0.52
4
5
1 .oo 0.57
1.oo 0.79 0.68 0.36 0.19
1.oo 0.87 0.50 0.28
0.58 0.33
1.00 0.83 0.53 0.35
1 .oo 0.64 0.42
1 .oo 0.66
1 .oo 0.81 0.67 0.61
1 .oo 0.83 0.80
0.96
1.oo 0.79 0.75 0.73
0.90 0.90
1
.oo
1 .oo
0.54
1 .oo 0.94
1
1 .oo 1
.oo
.oo
1.oo 0.95
in CnHzn+* The z = - 5 peaks are monoisotopic; all others are polyisotopic. Lowest mass in each step is a reference peak. c Lowest mass in a series required for S ( M ) . S(99) = the smaller of: 0.206 X (HDI(85) - [H(85) S(85)I ), or HDI(59) where S(85) is the value obtained after completion of Step 3.
the same nomenclature for aromatics, and the same techniques for spectrometer operation, data acquisition, and computer storage and processing of the arrays of polyisotopic H ( M ) and monoisotopic H D I ( M ) peaks derived from the mass spectra. The average molecular weight is obtained from distillation data or by any other convenient method, and is stored in the computer for use in the computations. Low-resolution 70-volt mass spectra of petroleum fractions containing both saturates and aromatics generally have intense peaks due to saturates in the consecutive homologous series from C,HZ,+~through C,HZ,-, below about mass 127. Aromatics contribute heavily to most of the remaining peaks. However, saturates contributions in the aromatics region and aromatics contributions in the saturates region are significant. These cross-contributions are not constant, but depend on the average molecular weight of the fraction and the amounts and kinds of compound types present. An early step in our procedure divides the total sample spectrum into separate spectra for saturates, S(M), and aromatics, A ( M ) from which much of the interferences has been removed. Base-line techniques, similar in concept to those described for resolving overlapping aromatic types (3), are used in the aromatics regions of CnH2n+1through CnH2,-5 to resolve saturates and aromatics and to supply some of the peak height values needed for both S ( M ) and .4(M). Remaining 1426
values for S ( M ) are obtained from the saturates region after removing part of the aromatics contributions via a new pattern-fitting procedure. Remaining A ( M ) values are obtained from the observed peaks in the aromatics region of the total sample spectrum. Although neither S ( M ) nor A ( M ) are completely free of interference, a preliminary estimate of composition in terms of group types is made using our aromatics procedure (3) on A ( M ) and a new saturates procedure on S ( M ) . Once this information is obtained the remaining interferences are calculated and a final analysis is reported. The saturates spectrum containing only the peaks in C,H?,+, through C,H2,-5 provides insufficient detail to allow a complete resolution of the various cycloparaffin types found in petroleum. Hence, a procedure such as Hood and O’Neal’s ( 4 ) or Lumpkin’s ( 5 ) cannot be used. We have, instead, limited the saturates types to paraffins, noncondensed cycloparaffins, 2-ring condensed cycloparaffins, and 3-ring and greater condensed cycloparaffins. In addition, peak summations for certain cycloparaffins and/or olefins related to steranes and polycyclic triterpanes are determined so that their specific interferences in the aromatics spectrum can be estimated. These types, which we simply call steranes, are limited to compounds of CnH?n-G,CnH2n--8,and C,H?,-,o whose molecular weights fall in the range 340-428. The presence of steranes and polycyclic triterpanes in petroleum is well known. Steranes were observed by O’Neal and Hood (6); Meinschein (7) mentioned pentacyclic triterpanes and the possible existence of hexacyclic triterpanes. Hills and Whitehead (8) have confirmed the presence of the pentacyclic compounds, but there has been no confirmation for the hexacyclics. Nonetheless, we have included C n H ~ n - l ~ (possibly hexacyclic) compounds because there is often a small increase in parent peak heights in the C?,-Cal region. COMPUTATIONS
All components are initially expressed in terms of their contributions to the total ionization, 21, of the sample, where ZI is defined as the sum of all the polyisotopic peaks heights at whole masses beginning at mass 24 and extending to the end of the spectrum. The fraction of ZI obtained for aromatics accounts for the total aromatics in the sample. Saturates composition is then normalized to the nonaromatic fraction. For the computations, a given series of molecular ions is designated (M)+, and the series of ions at each of the next lower masses is expressed as (M-l)+. Polyisotopic peaks are used for the (M)+series; most, but not all of the ( M - 1)+ peaks are monoisotopic. Saturates Spectrum. The saturates spectrum, S ( M ) , consists of all peaks in the consecutive homologous series C , I H ~ n + l through C,H2,-5 starting with the low mass peaks indicated in Table I and extending to the high-mass end of each series. Peaks at the masses indicated in Table I, in the saturates region of the total spectrum, suffer from two types of aro(4) A. Hood and M. J. ONeal, “Advances in Mass Spe:trometry,” Vol. 1. J. D. Waldron. Ed.. Pergamon Press Ltd.. London. 1959, p p 179-192. (5) H. E. Lumokin. ANAL.CHEM.. 28, 1946 (1956). (6) M. J. O’Neal and A. Hood, “Abstracts-American Chemical Society, Division of Petroleum Chemistry,” Vol. 1, No. 4, Sept. 1956. (7) W. G. Meinschein, Bull. Amer. Ass. Petrol. Geol., 43, 925 (1959). (8) I. R. Hills and E. J. Whitehead, Nciturr, 209, 977 (1966).
ANALYTICAL CHEMISTRY, VOL. 43, NO. 11, SEPTEMBER 1971
matics interference: that due t o fragmentation processes yielding singly charged ions, and that due to doubly charged ions. The latter can be particularly troublesome in the analysis of cracked stocks. To allow analysis of such stocks without defining the sample source and the consequent use of dual or multiple equation coefficients, we force-fit these peaks to a n "average" saturates pattern of relative intensities in each series derived from the spectra of saturates fractions from several sources. This substantially reduces interference from doubly charged ions. Pattern-fitting is made on a peak-by-peak basis, one series at a time, using reference peaks in each series as starting points. No peak is allowed to exceed the value indicated by the average pattern, but any lower height is retained. After the series is fitted t o the given pattern, the reference peak is shifted one carbon number higher and the previously fitted peaks are compared to a new pattern. Again excess peak height is discarded. This stepwise shifting of reference peak and pattern comparison is continued until all peaks have been fitted. Data necessary t o perform the pattern fitting are given in Table I. All remaining peaks for S(M) are obtained by base-line techniques that effectively distinguish between contributions of saturates and aromatics in the aromatics region. The curves of peak heights us. 106/mass2for the CnH2n+Ithrough C,H,,-, series of saturates fractions from petroleum are essentially linear a t the higher masses. An example is given in Figure 1 for the CnH2n-1 series of a gas-oil saturates fraction. For clarity some of the peaks a t the crowded, highmass region are not shown, but the linear portion of' the curve above mass 251 includes most of the observed peaks in the series. A line drawn from the peak at the highest mass observed, through the linear portion of the curve, to the point X at mass 125 is the base line for this series. The ratio of the height of X at mass 125 to the observed peak height is nearly constant for gas-oil saturates and the ratios of base-line heights to observed heights are also nearly constant a t each mass above 125. Thus if the base line of a series is known, it is possible t o predict the saturates spectrum of the series in the aromatics region. The high-mass base point is the peak at the highest observed mass in a series. For samples containing aromatics, it is assumed to be due only to saturates since, at equal boiling points, aromatics peaks occur at lower masses than saturates. The second, low-mass point needed to establish the base line is chosen at a mass where aromatics contribution is small. The peak height at this mass is multiplied by the constant derived from saturates spectra that places the base line o n the linear portion of the saturates curve. Values along the base line corresponding to masses in the series are stored in S(M). For the nonlinear portion, base-line values are multiplied by fixed peak-by-peak correction factors and the products replace the original base-line values stored in S(M). Data for establishing base lines are given in Table I1 and the peak-by-peak correction factors are listed in Table 111. In using base lines and correction factors to derive peak heights, we never allow a calculated peak height to exceed the observed value. Thus, the final operation in preparing S ( M ) is to compare each calculated peak with the corresponding H or HDI peak and reset each excess value to the observed value. After these steps are finished, the S ( M ) ariay contains all the data needed for further processing. The remaining aromatics interference is removed later from peak summations rather than from individual peaks. Seven of the 14 homol-
-1
c
I
-w
(1
I Y
a w
40
&
20
c
0 n "
X-@OP
10
0
-X
0 1 0
20
I
I
I
I
I
30
40
50
60
70
IO~/MASS* L I I 447
1
I I I I I 251
321
209
223
195
I
167
I
I
I
153
181
125 139
MASS
Figure 1. Peak height cs. mass for the monoisotopic C,,H2,-, series in the mass spectrum of the saturates fraction from a virgin gas oil
-
X - - X = base points
Table 11. Data for Establishing Base Lines z
series +l
Mass 99
Low mass base point ___ Abscissa Ordinate" 102.03
0 -1
125
63.00
-2
124
65.03
-3
95 136 149
110.80
-4
-5
84
141.71
54.07
45.03
I .00 x 0.78 X 0.15 X 0.85 X 0.21 1.16 X 0.51 X
x
HDl(99) H(84) HDl(125) H(124)
HDf(95) H(136) HDf(149)
High mass base point Let M M A X equal the highest observed mass in a series. Thc coordinates for the high mass base point in that series will be: Abscissa = (lOOO/MMAX)Z Odd-mass ordinaten = 1.00 X H D I ( M M A X ) Even-mass ordinate" = 1.00 X H ( M M A X ) a Where H D I ( M ) is the monoisotopic and H ( M ) is the polyisotopic peak at mass M.
ogous series for S ( M ) , C,H?,+I through CnH?n--5,contain peak height values; the remaining series do not. Sterane Identification. For the detection of steranes, the base points given in Table IV are used to establish base lines on the H and HDI spectra for each of the three (M)+and three ( M - 1)' series on scales of linear height peak cs. linear mass. Each (M)+ series is first examined in the parent peak region for the presence of steranes, as indicated by H values larger than values represented by the base lines. If they are present, heights measured from the base lines are summed as indicated in Table IV. During the summation any H or H D f value that falls below its base line is ignored. Separatc sums representing the three steranes types are retained for correcting the aromatics spectrum. Aromatics Spectrum. The aromatics spectrum, A ( M ) , consists of the values H ( M ) minus S(M) for the even-mass series and HDZ(M) minus S ( M ) for the old-mass series.
ANALYTICAL CHEMISTRY, VOL. 43, NO. 11, SEPTEMBER 1971
1427
Table 111. Peak-by-Peak Factors to Correct S ( M ) for Deviation from Base Lines z
series +1
0
-1
-2
-3
Mass 113 127 141 155 169
Factor 1.000 0.946 0.896 0.889 0.944
98 112 126 140 154 168
1.200 1.230 1.170 0.930 0.850 0.900
139 153 167 181 195 209 223 237 25 1 265 279 293
3.200 2.370 2.130 2.000 2.000 1,570 1.420 1,420 1.330 1 ,250 1.140 1.070
138 152 166 180 194 208 109 123 137 151 165 179 193
0.932 0.862 0.854 0.900 0.940 0.960 3.850 3.110 2.540 2.290 2.080 1.900 1.900
z
Table V. Peaks Included in the Aromatics Spectrum, A ( M ) z
Mass 207 221 235 249 263 277 29 1
Factor 1.570 1.420 1.400 1 ,370 1.370 1.200 1.100
-4
150 164 178 192 206 220 234
0.954 0.875 0.854 0.865 0.870 0.900 0.950
-5
163 177 191 205 219 233 247 26 1 275 289 303
2,020 1.900 2.020 1.530 1.510 1,370 1.360 1.310 1.220 1.180 1.100
series -3
Table IV. Data for Sterane Summations Base pointsn Low mass High mass Stepb 330 442 1 329 441 2 204 246 2 203 245 2 246 274 2 245 273 2 CnHzn-s 328 440 1 327 439 2 244 272 2 243 27 1 2 1 326 438 CnHzn-io 325 437 2 242 270 2 24 1 269 2 Base points are peak heights at the masses indicated. Use polyisotopic heights for even masses and monoisotopic heights for odd masses. * Steps 1 are performed first by obtaining sums of heights above the base lines at the masses in each series between the base points. A positive sum confirms the presence of the sterane type and sums from Steps 2 are added to the Step 1 sum. A zero sum in Step 1 indicates absence of the type and Steps 2 are omitted. Type CnH~n-6
Only the peaks required for determining aromatics need be entered in A ( M ) . These consist of the homologous series that include as their lowest masses the first significant (M)+ and (M - 1)+ peaks. These first peaks and the values that go into A ( M ) are given in Table V. 1428
series
Lowest mass 78 91 104 117 130 129 128 141 154 167 166 179 178 191
-6
-7 -8 -9
- 10 - 11 - 12 - 13 - 14 - 15 - 16 - 17 - 18 - 19
Table VI. Factors for Determining Saturates Contributions to Aromatics Summations z-series sums - 10
-6 PIUS -7 -8 PIUS -9
PI~IS -11
Steranes 1.55 3.41 2.17 278 2 104 0.10 4.61 3.41 2129 0.02 0.30 4.61 Fractions of saturates total ions P CP CCP 278 0.0010 0.0023 Fi s 2 104 0.0005 Fm 2129 Fim where P = total ions from paraffins CP = total ions from noncondensed cycloparaffins CCP = total ions from condensed cycloparaffins = 0 . 9 X lo-' X MW + 0.0346 F78 0.8133 X F,04 = 0.1542 X lo-' X MW = 0.286 X lo-' X M W - 0.3726 X lo-' Fit1 CCP interference in any (M - I)+ series equals: CCP interference in the summation X ( M - l)+/[(M - I)+
+
+
( M)+1
P interference on 2 128O equals:
+
+
+
0.26 x [S(127) S(141) S(155) . , . toendl P interference in ( M - 1)+ of 2128 equals: P X (0.1675 X lo-' X M W - 0.1615 X 0. No base line is used for (M)+ of 2128. The paraffin interference is calculated from the C,H2,+, peak series of the S spectrum.
The resultant A ( M ) spectrum contains contributions from saturates, including steranes, that must be eliminated before the final analysis. However, these are removed later by correcting summations of peaks rather than individual peaks in the spectrum. First Aromatics Computation. With only slight modification, the previously described aromatics procedure (3) is applied t o the A ( M ) spectrum, and the same terms apply. The only new terms are related t o the saturates interferences that must be removed before the final aromatics results are obtained. Before the peak summations obtained for the class analysis are processed through the inverse matrix, interferences from steranes are removed from the summations representing Classes I, 11, and 111. The sums obtained for z = -6, -8, and - 10 steranes, in the formula CnH2n+P, are each multiplied by the appropriate factors given in Table VI to obtain their total contributions to 278, 2104, and 2129. Although these factors were derived from a published spectrum of cholestane
ANALYTICAL CHEMISTRY, VOL. 43, NO. 11, SEPTEMBER 1971
(9)and are subject to the errors that can occur when data from a single compound are extrapolated to a class of compounds, the results seem reasonable. The aromatics procedure ( 3 ) yields initial results as contributions of the compound types t o total ionization. F o r samples containing only aromatics, the sum of these contributions may not equal the sum of peak heights at whole masses-a discrepancy probably related to a particular mass spectrometer and its operating conditions. We have found, however, that the ratio of observed t o computed total ionization is essentially constant. Consequently, to obtain the agreement in total ionization required for the present method, all the inverse matrix coefficients for aromatics are multiplied by the above ratio, and the diagonal elements of the calibration matrix are divided by the ratio. (9) American Petroleum Institute, Research Project 44, “Selected Mass Spectral Data,” Thermodynamic Research Center, Texas A& M University, College Station, Texas, 1947-1968.
Table VII. Peak Summations and Summations: 243 = 43 255 = 55 84 281 = 81 96 293 = 93 94
Total ionization contributions due to benzenes, naphthenebenzenes, and dinaphthenebenzenes (Classes IO, 110, and 1110) are summed to produce the total ionization from monoaromatics. The total ionization due to polyaromatics includes the sum of contributions from all the remaining classes. These two sums as well as the total ionization for all the aromatics are retained solely for calculating and removing contributions t o saturates. First Saturates Computation. Saturates composition is determined from four peak summations of the S ( M ) spectrum, which are processed by inverse matrix multiplication to yield total ionization contributions for four compound types. Thirty-three 4 X 4 inverse matrices, one for each carbon number from 8 through 40, are available. The average carbon number of the sample, defined as molecular weight f 14, determines which matrix is to be used. Specifications for peak summations and inverse coefficients are shown in Table VII.
Inverse Matrices for Saturates
+ 57 + 71 + . . . to end + 69 + 83 + . . to end + + 98 + 112 + . to end + 95 + 109 + to end + + 110 + 124 + . . to end + 107 + 121 + . to end + + 108 + 122 + to end , ,
,
,
I
.
I
Inverses: Paraffins Noncond. Cyclopar. 2-Ring Cond. Cyclopar. 3-Ring Cond. Cyclopar. Paraffins Noncond. Cyclopar. 2-Ring Cond. Cyclopar. 3-Ring Cond. Cyclopar. Paraffins Noncond. Cyclopar. 2-Ring Cond. Cyclopar. 3-Ring Cond. Cyclopar. Paraffins Noncond. Cyclopar. 2-Ring Cond. Cyclopar. 3-Ring Cond. Cyclopar. Paraffins Noncond. Cyclopar. 2-Ring Cond. Cyclopar. 3-Ring Cond. Cyclopar. Paraffins Noncond. Cyclopar. 2-Ring Cond. Cyclopar. 3-Ring Cond. Cyclopar. Paraffins Noncond. Cyclopar. 2-Ring Cond. Cyclopar. 3-Ring Cond. Cyclopar. Paraffins Noncond. Cyclopar. 2-Ring Cond. Cyclopar. 3-Ring Cond. Cyclopar. Paraffins Noncond. Cyclopar. 2-Ring Cond. Cyclopar. 3-Ring Cond. Cyclopar.
+
+
+
+
+
+
+
+ +
243 +2.0804 -0.2017 +0.0144
+O.oooO $2.0292 -0,3046 +O. 0238 -0.0009 $2.0439 - 0.4033 $0.0336 -0.0018
+2.0588 - 0.4812 +O. 0455 -0.0036 2.0767 -0.5653 $0.0631 -0.0067 +2.1005 -0,6474 +O. 0854 -0.0129 +2.1268 -0.7352 +O. 1142 -0.0211 +2. I478 -0,8169 +O. 1437 -0.0269 +2.1764 -0.9101 +0.1810 -0.0334
+
255
-0.2580 +2,2795 -0.1626 -0.0003 -0.3499 +2,3634 -0.1857 +O. 0008 -0.4568 +2.4505 -0.2319 +O. 0076 -0.5532 +2.5338 -0.2903 +0.0199 - 0,6491 +2.6356 - 0.3672 $0.0377 -0.7524 +2.7543 -0.4553 $0.0648 -0.8483 +2.8794 -0,5562 +o. 1001 - 0.9242 +2.9924 -0.6541 +O. 1213 - 1 ,0084 +3.1464 - 0.7643 +O. 1411
28 1 $0.0252 -0.2228 +3.6900 - 0.0153 +O. 0847 -0.7735 +3.0633 -0.0753 +O. 0885 -1.0453 +2.9440 -0.1825 +O ,0536 - 1 ,0732 +2.9971 -0.2982 +O. 0245 - 1.1368 +3.1410 -0.4378 +0.0011 - 1,1992 +3.2901 -0.5901 -0.0238 - 1.2758 +3.4558 -0.7444 -0.0536 - 1.3387 +3.5512 -0.7872 -0.0827 - 1 ,4208 +3.6475 - 0,8080
293 +0.0802 -0.7089 -2.9806 +4.1794 +0.0671 - 0.2875 - 2.4545 +4.2285 $0.1068 -0.0982 -2.3576 +4.3505 +0.1047 -0.0918 - 2.3803 $4.4420 +0.0355 -0.0407 -2.4846 +4.5708 -0,0216 -0.0429 -2.4474 +4.6736 -0.0858 -0.0320 - 2,4086 $4.7632 -0.1674 -0.0209 -2.3107 +4.7666 -0,2495 -0.0071 -2.2055 +4.7530
Cg
Cii
Cis
Cis
Cli
Ciq
CZI
C23
C?;
243 $2 0255 -0 2546 +O 0215 1 0 0002 +2 0356 -0 3570 +O 0290 -0 0014 +2 0510 -0 4418 $0 0400 -0 0 2 7 +2 0685 -0 5204 +O 0536 -0 0050 +2 0880 -0 6061 +O 0732 -0 0092 + 2 1144 -0 6905 +O 0986 -0 0167 $2 1424 -0 7814 +O 1303 -0 0242 +2 1602 -0 8620 +O 1620 -0 0299 +2 1922 -0 9522 $0 1999 -0 0366
255
281
293
-0.3055 +2.3236 -0.1961 -0.0021 -0.4006 2.4144 -0.2082 +O. 0043 -0.5066 +2.4898 -0.2609 +O. 01 38 - 0.6058 +2.5837 -0.3288 $0.0284 -0.7003 $2.6959 -0.4097 +O. 0509 - 0.8056 $2.8154 -0,5018 +0.0814 -0.8912 +2.9426 -0.6077 +0.1110 -0.9675 +3.0770 -0.7098 +0.1308 -- 1.0412 +3.2167 -0.8227
+O ,0687
+O. 0630
+
+ O . 1515
-0.5224 +3.3209 -0.0132 +O. 1057 - 1.0474 +2.9614 -0.1280 +O ,0694 -1,0582 +2.9737 - 0.2464 +O. 0384 -1.1001 +3.0622 -0.3681 +O. 0086 - 1.1667 +3,2183 -0.5159 -0.0089 - 1.2370 +3.3780 -0.6710 -0.0369 - 1,3090 +3.5042 -0.7684 -0.0709 - 1.3826 +3.5996 -0.7982 - 0.1034 - 1.4621 +3.7070 -0.8233
-0.4791
- 2.6629 +4.1784 +0.0700 - 0.0819 -2.3715 +4.2884 +O. 1424 -0.1030 - 2.3709 +4.4012 +O. 0726 - 0.0714 - 2.4212 $4.4970 +O . 0 109 -0.0437 - 2.4674 +4.6205 - 0.0485 -0 0393 - 2.4371 +4.7281 -0.1259 - 0.0243 -2.3651 +4 7637 -0.2027 - 0.0169 - 2.2607 +4.7655 -0.2915 $0.0046 -2.1721 +4,7657 I
(Continued next page)
ANALYTICAL CHEMISTRY, VOL. 43, NO. 11, SEPTEMBER 1971
1429
Table VII. Continued Summations:
+ + + + + + + + +
+ + + + + + +
+ + + + + + +
243 = 43 57 71 . . . to end 255 = 55 69 83 . . to end 84 98 112 , . to end 281 = 81 95 109 , . to end 96 110 124 , to end 893 = 93 + 107 121 , to end 94 108 122 . to end
Inverses.: Paraffins Noncond. Cyclopar. 2-Ring Cond. Cyclopar. 3-Ring Cond. Cyclopar. Para ffin s Noncond. Cyclopar. ' 2-Ring Cond. Cyclopar. 3-Ring Cond. Cyclopar. Paraffins Noncond. Cyclopar. 2-Ring Cond. Cyclopar. 3-Ring Cond. Cyclopar. Paraffins Noncond. Cyclopar. 2-Ring Cond. Cyclopar. 3-Ring Cond. Cyclopar. Paraffins Noncond. Cyclopar. 2-Ring Cond. Cyclopar. 3-Ring Cond. Cyclopar. Paraffins Noncond. Cyclopar. 2-Ring Cond. Cyclopar. 3-Ring Cond. Cyclopar. Paraffins Noncond. Cyclopar. 2-Ring Cond. Cyclopar. 3-Ring Cond. Cyclopar. Paraffins Noncond. Cyclopar. 2-Ring Cond. Cyclopar. 3-Ring Cond. Cyclopar.
C26
+
C28
+
c 3 D
+
C3?
+
c 3 4
+
C36
+
Can
+
+
CtO
z43 $2.2126 -0.9998 $0.2240 -0.0433 $2.2519 - 1.1024 $0.2845 -0.0588 $2.2935 -1.1964 $0.3358 -0.0743 +2.3300 -1.2801 $0 ,3984 -0.0956 $2.3724 - 1.3796 $0. 4785 -0.1280 $2.4166 -1.4912 $0.5711 -0.1687 $2.4756 -1.6263 $0.6929 -0.2294 $2.5460 -1.7851 $0.8438 -0.3191
255
-1.0875 $3.3017 -0.8942 +O. 1709 - 1.1724 +3.4776 - 1.0666 $0.2196 -1.2657 $3.6525 - 1.2134 $0. 2692 - 1.3570 $3.8393 -1.4039 $0.3389 - 1.4631 $4.0670 - 1.6396 $0.4381 - 1,5734 +4.3096 - 1.9029 $0. 5634 - 1.7083 $4.5961 - 2.2339 $0.7426 -1.8812 $4.9471 -2.6428 $1.0049
28 1 293 -0.1168 -0.3554 - 1.5021 $0.0140 $3.7740 -2.1879 $4.9164 -0.8631 -0.1416 -0.4672 - 1 ,5959 $0.0387 -2.3576 $3.9492 -0.9668 $5.3190 -0.1372 -0.6135 - 1 ,7029 +O .OS86 $4.1277 -2.5608 -1.0901 $5.8093 -0.1249 -0.8019 $0.0916 -1.8301 +4.3493 -2.8043 - 1.2438 $6.3808 -0.0799 -1.0034 -1.9840 $0.0838 +4.6054 - 3.0855 - 1.4454 $7.1168 - 0.0186 -1.2688 -2.1574 $0.1244 -3.4453 $4.9001 -1.6935 $7.9904 +0.0621 -1.5332 -2.3510 +O. 1017 -3.8725 $5.2550 - 2.0224 +9.0974 $0.1913 - 1.8720 - 2.5895 $0.0539 +5 ,6978 -4.4441 - 2.4861 f 10.6097
Before matrix multiplication, it is necessary t o correct the summations for aromatic contributions. Because the lowmass peaks include aromatics ions, the original base lines used in obtaining S ( M ) and A ( M ) are in error by amounts that depend on the total quantity of aromatics in the sample. Additional aromatics contributions are also present on those low-mass peaks that were previously pattern-fitted. Corrections for base-line error are made by first calculating the fraction of aromatics on each low-mass base peak, assuming that this fraction is constant for all peaks determined by the base line, and, finally, by removing the total contribution found from the saturates summation and adding it to the aromatics summation. The base-line technique assumes that the highest mass peak in a series, S(MMAX), is aromatic-free. Therefore, all calculations for base-line error are made on values of S ( M ) - S(MMAX). Data to determine aromatics on the low-mass base peaks, given in Table VIII-A, were obtained from the spectra of a number of aromatic fractions. The product of the factor, F, times aromatics total ions is the contribution from aromatics at a low-mass base peak. Aromatics contributions to the pattern-fitted peaks in the saturates summations depend on both the average molecular 1430
c 2 7
Cm
C3l
Caa
C3j
C37
Cai
243 $2.2320 - 1,0492 $0,2493 -0.0495 +2.2729 -1.1444 +O .3026 -0.0643 i-2.3106 - 1.2370 $0.3657 -0.0840 $2.3489 - 1.3269 $0.4369 -0.1108 +2.3941 - 1,4368 +O. 5238 -0.1472 $2.4448 -1.5536 +O. 6268 -0.1966 +2.5038 - 1 ,6930 $0. 7589 -0.2686
255
- 1.1249 +3.3783 -0.9633 $0.1899 - 1.2222 +3.5560 - 1.1192 +O. 2380 - 1.3083 +3.7423 - 1.3038 $0. 3008 - 1.4102 $3.9498 - 1. 5 196 +O. 3843 -1.5146 +4.1819 - 1.7643 +0.4961 - 1.6404 1-4.4490 - 2.0587 +O ,6474 - 1.7836 t4.7519 -2.4204 t o . 8605
28 1 -0.1354 - 1,5430 $3.8536 -0,9120 -0,1379 - 1.6452 +4.0278 -1,0254 -0.1330 - 1.7630 +4.2319 - 1.1631 -0,0950 -1.9116 $4.4752 - 1.3366 -0,0560 -2.0627 1-4.7420 -1.5580 +0.0215 -2.2572 +5.0738 -1.8515 +0.1160 -2.4584 t5.4585 -2.2332
29 3 -0.4098 +O ,0350 -2.2792 $5.1179 -0.5384 +O. 0449 - 2,4554 $5.5673 -0.7039 +O. 0754 -2.6844 +6.0919 -0.9185 $0,1112 -2.9412 +6.7199 - 1.1365 +0.1061 - 3.2428 +7.5122 -1.4023 f0.1358 -3.6723 1-8.5329 -1.6941 1-0.0745 -4.1241 +9.7905
weight of the sample and the amounts of mono- and polyaromatics. The fractions of total ions for the two aromatics groups are calculated from the equations given in Table VIII-B. These equations are relatively simple and, admittedly, can only approximate the aromatics contributions to saturates, but they fit most of our spectra of aromatics fractions. After the correction steps have been completed, the saturates summations are processed by inverse matrix multiplication, which yields approximate values of total ions for the four saturates types. These are used to correct the various aromatics summations. For such corrections, total ions for the two condensed-ring cycloparaffin types are combined and used as a single type. Final Aromatics Computation. Contributions from saturates to 278, 2104, 2129, and 2128 are arbitrarily divided into two categories: those that must be removed from the (M)+plus (M - 1)+ sums, and those that interfere specifically with the (M - I)+ sums used to resolve overlapping aromatic types. Data to remove contributions from the (M)+ plus (M - 1)+ sums are given in Table VI. The corrected sums are next multiplied by the 7 X 7 inverse matrix coefficients to obtain final values of total ions in each of the seven classes..
ANALYTICAL CHEMISTRY, VOL. 43, NO. 11, SEPTEMBER 1971
I
Table VIII. Fractions of Aromatics Total Ions Appearing on the Low-Mass Base Peaks and on Low-Mass Peaks in the Saturates Summations
+ + + + + +
Condensed-ring cycloparaffins contribute significantly to the ( M - l)+ series of Classes I, 11, and 111, and paraffins contribute to the ( M - 1)+ series of Class IV. It is necessary first to calculate the total interference to a series (Table VI), and then to distribute this quantity among the nominal, first, and second overlaps. The spectra of a number of saturates fractions have shown that about 0.0317 of the total ions from condensed-ring cycloparaffins is o n the ( M - I)+ peaks for the nominal, IO, type of Class I. This value is subtracted from the total condensed-ring interference in ( M - 1)+ of Class I t o yield the interference to the overlap sums. Half of such interference is subtracted from the first overlap and half from the second overlap. Negative results, if found, are set to zero. Amounts of interference in Classes I10 and 1110 are equal t o the amount in Class IO multiplied by the ratios F,O~/F,~ and F I ~ ~ I Frespectively ~S, (Table VI). Again, the remaining interferences are divided equally between the first and second overlaps, and negative results are set t o zero before continuing. The entire ( M I)+ contribution from parafins to Class IV is removed from the second overlap. For Class I the small contribution from paraffins is subtracted from the nominal type. Similar contributions from noncondensed cycloparaffins t o Classes I and I1 are deducted from the respective nominal types. Any negative results are set to zero. When steranes are present, corrections for their contribul)+ series of 878, 8104, and 8129 are tions t o the ( M obtained by multiplying the total interferences as calculated from the data in Table VI by ( M - l)+/[(M - 1)+ (M)+]. The results are then assigned 17% to the nominal type and 83 to the second overlap type. The total corrections to the ( M - 1)+ series also must include those for aromatics (3). Once all these corrections have been made, total ion contributions of the aromatic types
-
-
+
I
I
I
I
I
1 1
PARAFFINS
1
0.6
A. On low-mass
base peaks: Base peak Equation" 99 F = 0.159 X 0.892 x 10-8 X MW F = 0.948 X l W a 84 125 F = 0.326 X 0.178 X X MW 124 F = 0.153 x 10-3 0.100 x 10-6 x MW 95 F = -0.0013 0.1717 X lo-' X MW F = 0.147 X 136 0.124 x 10-5 x M W 149 F = 0.356 X 0.187 X lW5X M W B. On low-mass peaks in saturates summations: Summation Equationn 243 FM = 0.415 X X M W - 0.0611 FP = 0.282 X X MW - 0.0538 255 FM = 0.354 X X MW - 0.0548 FP = 0.129 X X MW - 0 0197 F M = 0.056 X X M W - 0.0061 281 FP = 0.019 X lW3 X MW - 0.0030 FM = 0.046 X X MW - 0.0027 293 FP = 0 014 X X MW - 0.0004 a MW is average molecular weight of sample. F is fraction of total ions for all aromatics. FM is fraction of total ions for all monoaromatics FP is fraction of total ions for polyaromatics.
I
w
w 0.2
[
0.1 3-RING
+
C O N D . CP
0.0
II
1
I
I
I
I
I
N O N C O N D . CP
0,4t
w
I
I
I
'1
0.31
\
Ln _.
fi
0.21
2.RING COND. C P
3-RING
0.0' 8
I 12
I 16
i
-I
1
C O N D . CP
I
I
I
20
24
28
I 32
I I 36
AVERAGE C A R B O N N U M B E R
Figure 2. Z43/ZI and ZSSjZI us. average carbon number for saturated hydrocarbon types are calculated. Since the total ionization is approximately proportional to the liquid volume of sample charged to the mass spectrometer (IO), the ratio of total ionization for each aromatic type t o the total ionization of the sample is its fractional liquid volume concentration in the sample. Final Saturates Computation. The entire procedure for the first saturates computation is repeated with the new values obtained in the final aromatics computation. The results are in terms of total ionization (and, consequently, relative liquid volumes) for each of the four saturates types. These values are normalized to the total saturates value represented by sample minus total aromatics. RESULTS AND DISCUSSION Saturates. The procedure to determine saturates uses most of the peaks available in the saturates spectrum. In
any scheme using summations the choice of peaks to be summed for a group type is dictated not only by the spectral features of the compounds included in the group, but also by the features of all other compounds likely to be present in the sample. The objective is to maintain high and constant sensitivity for the compounds in the group, but low and constant sensitivities for compounds in other groups. Inclusion of all ( M - 1)+ peaks, and ( M ) + peaks for cycloparaffins, results in higher sensitivities than are obtained with only a few peaks in each sum. Inclusion of masses 43 and 57 for paraffins and 55 for noncondensed cycloparaffins keeps sensitivities high at low carbon numbers. Inverse matrices for saturates are derived from calibrations containing interpolated and extrapolated data principally from published spectra (9). These data, in the form Zx/ZI, where Z x is a summation and ZI is the sum of all peak heights at whole masses, are shown as functions of average carbon number in Figures 2 and 3. We have included matrix co(10) A. Hood, ANAL.CHEM., 30, 1218 (1958).
ANALYTICAL CHEMISTRY, VOL. 43, NO. 11, SEPTEMBER 1971
1431
0.4
4 I
2 - R I N G COND. CP
0.3
>
I
I
PARAFFINS
' 0 . 0
>
I
I
0.1
3.RING + COND. CP
1
0
.
8
1
2 - R I N G C O N D . CP
0 12
' 16
d
I
N O N COND. CP
,
I
I
1
20
24
28
32
-
36
AVERAGE CARBON NUMBER
Figure 3. ZSl/ZI and Z93/ZI us. average carbon number for saturated hydrocarbon types
0.28
Table X. Total Aromatics in Gas Oils and in Catalytically Reformed Naphtha Aromatics-liquid vol. MW This method Silica gel Gas Oil A 25 3 21.9 19.6 Gas Oil B 340 38.9 39.8 Gas Oil C 297 46.2 45.1 Gas Oil D 339 17.6 19.6 278 29.6 29.9 Gas Oil E 298 38.5 39.2 Gas Oil F 129 32.5 32.0" Naphtha a Mass spectrometric naphtha type analysis.
0.24
0.20
-w
0.16
\
m d
w
0.12
0 BENZENES 0 T M A ESTERS
/
00
0.041
I
0'
h.
0.0080.1 0 4
I
I
I
I
I
I
I
8
12
16
20
24
28
32
~~~
'
36
CARBON A T O M S I N N-ALKYL GROUP
Figure 4 . 243/2I US. n-alkyl chain length for n-alkylbenzenes and n-alkyl trimellitic anhydride esters efficients for 3-ring+ condensed cycloparaffins a t CS; because low average carbon number samples often contain these compounds as heavy ends, it would be wrong to exclude them. The curves representing interferences at 243 are of particular interest. At the higher carbon numbers, long n-alkyl side chains apparently cause interfering contributions that increase with increasing chain length. As shown in Figure 4, to obtain better data, we plotted Z43/ZI for n-alkylbenzenes ( 9 ) and for 4-n-alkyl esters of trimellitic anhydride (11) US. chain length. The data for the two compound types are in good agreement and a curve representing the T M A esters is smooth enough to extrapolate with some confidence. When (1 1) Unpublished
1432
Table IX. Analysis of Aromatic-Free and Saturate-Free Samples Liquid volume Acid-treated Aromatics white oil fraction (389 MW) (353 MW) Paraffins 19.8 0.0 Noncondensed cycloparaffins 44.9 2.8 2-Ring condensed cycloparaffins 19.2 0.4 3-Ring+ condensed cyclopdraffins 14.8 2.7 Total saturates 98.7 5.9 Benzenes 0.0 7.8 Naphthenebenzenes 0.0 10.4 Dinaphthenebenzenes 0.0 13.6 Naphthalenes 0.0 4.4 Acenaphthenes, dibenzofurans 0.2 9.8 Fluorenes 0.1 10.9 Phenanthrenes 0.1 6.5 Naphthenephenanthrenes 0.0 5.1 Pyrenes 0.0 4.7 Chrysenes 0.0 2.5 Perylenes 0.1 1 .o Dibenzanthracenes 0.3 0.6 Benzothiophenes 0.0 3.4 Dibenzothiophenes 0.0 3.1 Naphthobenzothiophenes 0.2 0.9 Unidentified 0.3 9.4 Total aromatics 1.3 94.1 Calculated total ionization, % 97.5 100.5
spectra from this laboratory.
applied to the cycloparaffins, values from the curve supplement data from published spectra and permit extension of 243121 to high carbon numbers. Sulfur Compounds. Of the several sulfur compound types often found in petroleum, our procedure specifically considers only the thiopheno types. In some oils thiaalkanes and dithiaalkanes may be present in significant amounts. Since the mass spectra of these compounds (9) show that many of their large peaks are at low masses not included in S(M) or A(M), the calculated total ionization should be less than 21. However, since the saturates are normalized to sample minus aromatics, the result is that much of the sulfur is distributed among the saturated types. Olefins and Cracked Stocks. In the development of this procedure, we did not consider the possible presence of olefins. We believe that the presence of nonaromatic olefins would lead to errors in the saturates values while the presence of aromatic olefins would lead to errors in the aromatics values. Most of the olefins produced by catalytic cracking of petroleum appear in the gas and naphtha fractions. Products boiling above gasoline do not usually contain large amounts of olefins and we have had no difficulty in using the procedure on such materials. Typical Analyses. Our analyses of several typical petroleum samples are summarized in Tables IX through XIII.
ANALYTICAL CHEMISTRY, VOL. 43, NO. 11, SEPTEMBER 1971
Table XI. Comparison of Aromatics Compositions Liquid vol. Virgin gas oil Coker gas oil MW
=
Whole oil
278
M b
MW = 297
Arom. frac.m
Whole oil
Arom. frac.a
4.9 4.8 5.2 Benzenes 4.6 4.8 Naphthenebenzenes 5.8 4.6 4.6 6.0 Dinaphthenebenzenes 4.3 4.2 3.3 Naphthalenes Acenaphthenes, 2.8 2.9 4.3 dibenzofurans 2.4 2.4 3.9 Fluorenes 1.7 1.7 2.3 Phenanthrenes 0.3 0.0 1.5 Naphthenephenanthrenes a Normalized to the total aromatics found for the whole
5.3 5.5 5.6 3.3 4.8 4.4 2.8
Liquid vol. Virgin gas oil Coker pas oil
Pyrenes Chrysenes Per ylenes Dibenzant hracenes Benzothiophenes Dibenzothiophenes Naphthobenzothiophenes Unidentified
278
MW
Whole oil
Arom. frac.5
Whole oil
0.7 0.1 0.0
0.4 0.1 0.1 0.0 2.2 1.4
2.0 0.6 0.3 0.1 5.3 4.0 1.1 0.6
0.0
1.9 1.3 0.0
0.0
0.0
0.0
2
297 Arom.
frac.a 2.0 0.6 0.4 0.1 6.1 4.3 1.1 0.0
0.0
oil.
Table XII.
Composition of a Wide Boiling West Texas Cas Oil and Its Distillate Fractions in Liquid Volume Per Cent Calcuc u t 1 cut 2 cut 3 cut 4 Cut 6 cut 7 cut 5 Cut 8 lated Analyzed Volume ”/, of whole feed 13.5 12.2 13.2 12.5 10.4 13.2 12.2 12.8 100.0 100.0 Boiling range, ‘F 727778345562665830868909408Average molecular weight Composition : Paraffins Noncondensed cycloparaffins 2-Ring condensed cycloparaffins 3-Ring condensed cycloparaffins Total saturates Benzenes Naphthenebenzenes Dinaphthenebenzenes Naphthalenes Acenaphthenes, dibenzofurans Fluorenes Phenan t lire ties Naphtheiiephenanthrenes Pyrenes Chryenes Perylenes Dibenzanthracenes Benzothiophenes Di benzothiophenes Naphthobenzothiophenes Unidentified Total aromatics Calculated total ionization,
+
632 21 1
684 249
7 30 28 1
782 307
847 344
89 3 379
32.7 25.7 11.0 6.0 75.4 7.3 4.5 3.3 5.5 1.1 0.7 0.2 0.0 0.1 0.0 0.0
29.2 21.7 9.4 6.1 66.4 5.2 4.0 5.5 4.7 4.3 2.3 1.3 0.4 0.7 0.1 0.0
21.8 19.2 9.0 5.9 55.9 5.4 4.4 5.4 3.4
20.3 17.7 9.2 5.6 52.8 4.8 4.4
18.0 18.8 8.8 7.8 53.4 4.0 3.3 5.0 2.0 4.9
15.7 19.8 9.4
0.0
0.0
1.9
3.3 1.7 0.1
0.0 0.0 0.0
24.6 95.4
0.0
33.6 99.4
5.0
4.0 3.3 1.2 1.7 0.8 0.2 0.0 3.3 5.9 0.1 0.0 44.1 101.4
For example, Table IX shows results for a highly acid treated (aromatics free) white oil and a n aromatics (saturates free) fraction. Despite such extreme ranges in concentrations, the results are reasonable and good balances are obtained for the total ionization of both samples. As shown in Table X, total aromatics determined on several gas oils agree with data obtained for fractions separated on silica gel, and the value obtained for a catalytically reformed naphtha agrees with that determined by a naphtha type analysis. Likewise, as shown in Table XI, the aromatics determined in whole distillates agree well with the results obtained by our analysis for aromatics only (3). The results in Table XI1 illustrate that the method clearly distinguishes changes in the concentrations of various compound types with boiling range. These samples were obtained from a wide-boiling virgin gas oil that had been distilled to give eight cuts of approximately equal volume. The increase in noncondensed cycloparaffins with boiling range above cut 4 may be due to an increase in isolated multiring structures (4). Also, we believe that the sharp increase in 3-ring+ condensed cycloparaffins that occurs above cut 4 is’
5.5
2.6 5.1 4.5 3.8 2.4 2.8 1.2 0.3 0.1 3.4 4.5 0.2 1.6 47.2 102.6
5.1 3.1 1.7 2.9 1.5
1.o
0.6 3.3 3.0 1.8 3.4 46.6 98.6
11.0
55.9 3.9 2.5 4.6 2.3 4.2 4.5 2.5 0.5 2.3 1.6 1.2 0.7 2.9 2.5 2.5 5.4 44.1 95.2
944
414 10.6 18.5 7.9 10.2 47.2 3.8 3.9 5.4 1.2 4.0 4.7 2.4 3.7 -2.5 1.8 1.4 0.5 2.9 2.4 1.7 10.5 52.8 100.8
1180 468
6.5 20.2 6.8 11.4 44.9 3.7 4.0 6.0 0.5
3.0 4.7 1.8 3.7 1.8 1.1 1.6 0.8 2.6 2.3 1.5 16.0 55.1 100.5
1086 35 1 19.6 20.3 9.0 8.0 56.9 4.8 3.9 5.1 2.8 3.9 3.8 2.3 1.6 1.8 1.o 0.7 0.3 2.9 2.8 1.o 4.5 43.2
...
19.1 20.0 9.2 8.2 56.5 4.9 4.5 5.8 3.5 4.4 3.5 2.1 2.1 1.9 0.9 0.3 0.2 3.2 2.7 0.8 2.7 43.5 98.3
due, at least partly, to steranes. The appearance of polynuclear aromatic structures in what seems to be the proper boiling range order is gratifying. An apparent anomalous increase in naphthenephenanthrenes in the last two cuts is attributed to unidentified Class I aromatics. However, the good agreement between the calculated and analyzed composition of the whole oil indicates that the method is consistent over various boiling ranges. Table XI11 shows the composition of the SAE-10 distillate fraction and all the other products obtained during the solvent extraction and dewaxing of a Louisiana crude. These data substantiate prior general knowledge of such materials. For example, the extract contains 76% of the aromatics and only 1 5 % of the saturates, while the wax contains 1 8 % of the saturates (including 50% of the paraffins) and slightly more than 2 % of the aromatics. Table XI11 also includes two compositions calculated for the distillate from the data for extract plus waxy raffinate and for extract plus wax plus base stock. The close agreement with the composition of the original distillate confirms the consistency of the method over a range of compositions.
ANALYTICAL CHEMISTRY, VOL. 43, NO. 11, SEPTEMBER 1971
1433
Table XIII. Compositions of Products from Processing SAE-10Lousiana Distillate for Lubricating Oil Base Stock Liquid volume,
Distillate, % Paraffins Noncondensed cycloparaffins 2-Ring condensed cycloparaffins 3-Ring condensed cycloparaffins Total saturates Benzenes Naphthenebenzenes Dinaphthenebenzenes Naphthalenes Acenaphthenes, dibenzofurans Fluorenes Phenanthrenes Naphthenephenanthrenes Pyrenes Chrysenes Perylenes Dibenzanthracenes Benzothiophenes Dibenzothiophenes Naphthobenzothiophenes Unidentified Total aromatics Calculated total ionization, Z
+
Distillate
Waxy raffinate
100.0 21.3 27.1 14.8 15.4 78.6 2.6 1.6 3.0 2.1 1.6 2.1 1.6 0.4 0.9 0.4 0.5 0.7 0.6 0.5 0.4 2.4 21.4 99.0
72.4 26.1 32.5 16.4 17.3 92.3 0.7 0.0 0.8 1.4 0.8 0.7 0.0 0.0 0.0 0.0 0.1 0.6 0.1 0.0 0.3 2.2 7.7 97.3
CONCLUSION An important feature of our procedure is its applicability to most, if not all, samples encountered in petroleum processing. Thus, it is now possible to use a single method to determine compound types in fractions ranging in volatility from heavy naphthas to heavy gas oils and in composition from all saturates t o all aromatics. This is not only an advantage for the analyst, but, of more significance, it avoids the discontinuities in data that can occur when separate methods are used. In our method, agreement between calculated and observed total ionization is about i S % for samples boiling in the range of 400-1000°F and increases only slightly if the range extends to 200-1100°F. When compared with the method of
Extract
Wax
Base stock
27.6 7.1 13.5 8.1 12.2 40.9 5.7 4.4 7.4 4.7 5.3 6.7 4.9 3.0 4.4 2.1 0.9 0.9 1.5
14.4 73.2 12.9 5.0 5.2 96.3 0.2 0.1 0.3 0.4 0.7 0.0 0.1 0.0 0.0 0.0 0.0 0.2 0.0 0.4 0.0 1.3 3.7 97.8
58.0 15.5 36.7 17.6 21.1 90.9 1.1 0.0 0.8 1.7 1. o 1.2 0.0 0.0 0.0 0.1 0.2 0.6 0.2 0.0 0.3 1.9 9.1 97.0
1.5
1.1 4.6 59.1 101 . o
Extract + waxy raffinate
Extract + wax + base stock
100.0 20.9 27.2 14.1 15.9 78.1 2.1 1.2 2.6 2.3 2.0 2.4 1.4 0.8 1.2 0.6 0.3 0.7 0.5 0.4 0.5 2.9 21.9
100.0 21.6 26.9 13.1 16.3 77.9 2.2 1.2 2.6 2.3 2.1 2.6 1.4
...
0.8 1.2 0.6 0.4 0.6 0.5 0.5 0.5 2.6 22.1 ...
Hood and O’Neal for saturates (4, in the applicable carbon number range, it gives about equal results on paraffins, but somewhat higher noncondensed and 2-ring condensed cycloparaffins at the expense of 3-ring+ compounds. We have found no need for separate inverses to accomodate predominantly n-paraffinic and isoparaffinic samples.
COMPUTER PROGRAM The computer program in use in this laboratory, written in Fortran IV, will be available to prospective users of’our procedure. Details may be obtained from the author.
RECEIVED for review January 28, 1971. Accepted May 13, 1971.
SpectrofIuorimetric Determination of Orthophosphate as Rhodamine B Molybdophosphate G . F. Kirkbright, R. Narayanaswamy, and T. S. West Chemistry Department, imperial College, London, S . W.7., U.K. Between 0.04 and 0.6 pg of phosphorus is determined as orthophosphate via formation of the ion-association complex of molybdophosphate with the basic dyestuff Rhodamine B. After extraction of excess dye reagent into chloroform, the Rhodamine B molybdophosphate is extracted into ch1oroform:butanol (4:l v/v) and the intensity of the fluorescence in this solvent measured at 575 nm with excitation at 350 nm. Optimal conditions for the determination have been established, and the effects of 100-fold excesses of each of 37 foreign ions have been examined. The determination is highly selective; large amounts of silicate do not interfere, and As(lll) and V(V) are tolerable at 25- and 50-fold weight excesses, respectively. The structure of the ion-association complex has been examined.
CQMMONLY EMPLOYED METHODS for the determination of orthophosphate are based on the absorption spectropho1434
tometry in solution of molybdophosphoric or molybdovanadophosphoric acids, or of their reduction products in aqueous or organic media. These methods are neither highly sensitive nor selective, although the latter may be attained by solvent extraction of the heteropoly acid or the use of masking agents (1-4). Sensitive and selective procedures have been described for the determination of orthophosphate by indirect amplification, in which the twelve molybdate ions associated with each phosphate ion in the heteropoly acid are determined by molecular or atomic absorption spectrophotometry after (1) C. Wadelin and M. G. Mellon, ANAL. CHEM., 25, 1668 (1953). (2) J. Paul, Ana/. Chirn. Acra, 23, 178 (1960). (3) J. Paul, Mikrochim. A c f a , 1965, 830. (4) R. A. Chalmers and A. G. Sinclair, Anal. Chirn. Acta, 34, 412 (1966).
ANALYTICAL CHEMISTRY, VOL. 43, NO. 11, SEPTEMBER 1971
