ANALYTICAL CHEMISTRY
1518 Table 11. d , A.
d , A.
1/11
52.
Kryptogenin acetate n 19 15.2 ". 0.04 14.0 9.07 0.09 0.17 7.90 7.03 0.20 6.21 1.00 0.13 5.87 1.00 5.52 0.13 4.94 0.01 4.71 0.13 4.38 0.83 4.17 0.20 3.92 0.27 3.72 0.27 3.50 0.09 3.34 0.09 3.25 0.13 3.08 0.09 2.97 0.04 2.85 0.13 2.77 0.02 2.68 0.01 2.56 0.09 2.46 0.20 2.37 0.09 2.29 0.04 2.23 0.02 2.17 0.01 2.11 0.09 2.07
1/11
52.
Kryptogenin acetate (contd.) 1.97 0.02 1.93 0.02 1.85 0.02 1.80 0.02 1.75 0.02
1"
IV. STEROID DERIVATIVES
K1. Methyl bisdehydroisynolic acid
X-Ray Diffraction Powder Data (Continued) d , .1.
1/11
K1. Methyl bisdehydroissnolic acid (contd.) 3.92 0.32 3.81 0.28 3.63 0.32 3.60 0.28 3.47 0.32 3.33 0.11 3.03 0.20 2.92 0.11 2.79 0.08 2.65 0.11 2.59 0.05 2.53 0.08 2.37 0.15 2.33 0.05 2.24 0.05 2.13 0.11 2.08 0.03 1.99 0.08
Ii2. Sec? 8/(4 deoxycholic acid 9.72 0.47 7.22 0.15 6.49 0.75 5.88 0.56
power pattern. At average room temperature, the particular form of crystal that appears generally depends upon the solvent from which the compound is recrystallized. In presenting the diffraction data of this paper, the recrystallizing solvents and also the resulting melting points for each steroid are specified. Before obtaining their x-ray diffractions, all of the steroids were recrystallized from various solvents. The method of recrystallization was as follows: A sample of the compound was dissolved in a minimum volume of solvent while being heated on a steam bath. Water was then added until the solution became cloudy. While this solution was maintained a t boiling temperature, a sufficient amount of solvent was added to redissolve the precipitate. The compound was allowed to crystallize a t room temperature. I n those cases where the above method failed to yield satisfactory results, ethyl ether was used to enhance the solubility
d , A.
1/11
d. A.
:{,!d"
$ ~ ~ x s c ~ ~ v.PSEE;DO ~ ~ (contd.) 5.57 0.75 5.31 0.08 5.10 1.00 4.90 0.38 0.38 4.47 0.47 4.20 0.15 3.88 3.73 0.38 3.56 0.28 0.28 3.41 0.11 3.23 0.33 3.14 2.97 0.15 0.05 2.88 2.80 0.05 0.15 2.70 2.65 0.05 2.54 0.05 0.15 2.50 0.08 2.44 0.11 2.34 0.15 2.28 0.15 2.17 0.03 2.12 0.05 2.07 0.05 2.03
1/11
STEROIDS
L1. Mes+3.4-bis(phydroxyphenyl). n-hexane
d . A.
1/11
L2. 3,4-Bis(p-hydroxyphenyl)3-hexene 9.41 0.15 8.23 0.11 0.27 7.47 0.55 5.85 5.26 0.02 1.00 4.85 4.45 0.73 4.21 0.20 3.95 0.27 3.74 0.11 3.55 0.34 3.42 0.41 3.16 0.04 3.06 0.04 0.24 2.91 2.56 0.11 2.51 0.15 2.40 0.05 2.31 0.05 2.25 0.02 2.19 0.15 0.02 2.04 0.04 1.95 1.88 0.05 1.84 0.08 1.76 0.08 0.02 1.68 1.65 Q.02 1.60 0.02 1.56 0.04
of the steroid. The substance was then allowed to crystallize from this solution, The solvents used were ethanol, methanol, acetone, and ether. The majority of the steroids were recrystallized from ethyl alcohol. The exceptions are noted in Table I. ACKNOWLEDGMENT
The technical assistance of Rachel Silas and Kathleen Matheson, who assisted in the preparation of the x-ray data, is gratefully acknowledged. LITERATURE CITED
(1) Beher, W. T., Parsons, J., Baker, G. D., ANAL.CHEM.27, 1569 (1955).
(2) Parsons, J., Beher, W. T., Ibid.,27, 514 (1955). RECEIVEDfor review September 24, 1955. Accepted June 4, 1956.
A System for the Rapid Reduction of Mass Spectrometric Data B. K. FRITTS
and
C . GORDON PEATTIE'
Exploration and Production Research Division, Shell Development Co., Houston, rex.
A system is described which represents a material improvement over the conventional method of reducing mass spectra by manual measurement of peak heights with a ruler. Spectra of compounds of high molecular weight are reduced about six times faster than previously, with no loss of accuracy and without operator fatigue. The data are presented in a form suitable for processing by electronic computers. Use of this system should encourage development of new high molecular weight methods by permitting reduction of spectra in a quantity which would be impracticable by the manual method.
T
HE rapid and accurate measurement of peak heights has always been a problem in mass spectrometry. Manual measurement has been most widely used; the peak height is read from a ruler laid on the spectrum or by use of grid lines provided on the record. Most workers use a ruler graduated in fiftieths or hundredths of an inch. The deflection is then recorded on the record or on a separate paper together with the galvanometer factor by which it is to be multiplied. Elaborate variations have only increased the time of measurement and have done little to reduce the high fatigue factor. While analyses were limited t o 'Present address. General Electrodynamics Corp., Garland, Tex.
V O L U M E 28, NO. 10, O C T O B E R 1 9 5 6 the maas range 1 to 100, and only a small. number of spectra were involved, this method of peak measurement could be tolerated. However, the inception of high molecular weight mass spectrometry with spectra ranging from mas8 40 to mas8 600 to 700 has made manual measurement of some 500 to 600 peaks impractical. Mistakes are frequent and production of results is low, tending to keep to a minimum the number of spectra measured. The Eystem described N&S designed originally for measurement of singletrace graphic records, hut was readily modified t o meet the speoifications needed for Eduction of mass spectrometric data. This modified system has been used for almost a year for the routine messurement of single and multitrace mass spectrometer records. With it, peaks can he measured quiokly and without fatigue, and the repeatahilityof any measurement is equal to that obtained by the manual method. The operator can go from one peak to the next with. out having to look away to record a peak height. Finally, i t presents a typed readout of the measurements and punches the same information into paper tape or cards which can be fed directly into an electronic computer. DATA REDUCTION SYSTEM
The data reduction system was made from four components obtained from TeleComPuting Cor 12838 Saticoy St., North Hollywood, Calif.: (1) Contact Tetieader, Ty e 35B,(2) Teleducer, Ty e 24A, (3) program nmt, Type 3 3 8 and (4) Flex+ *iter. &e Flexorniter was initially obtained from ~ ~ Control6 Corp., 1 Leighton Avo., Rochester 2, N. Y.,and adapted hy Telecomputing Corp. for use m t h this system. The Contact Telereader, shown in Figure 1 with a chart on its viewing screen, was modified for mas8 spectrometric use by adding five galvanometer sensitivity huttonsand five nero potentiometers. It contains t N 0 wires representing the z and y axes, which can be moved relative t o each other. The vertical wire is used simply as a visual aid in marking the peak to he measured. The horiOn the top of peak and thus sets up a eontal wire is voltage that is a fraction of a previously set maldmum This analog voltage is converted into decimal digits by the Teleducer. These digital present.ations of the Teleducer we then converted into sequential form by the program unit. The also permits introduction of supplementary informa. ~rogram this tlon, such as m m numbers m d gdvanometer.factors, information may he punched into paper tape by the Flexomiter. A typed readout is made a t the same time. ~
A system similar to this was described by Hoehgesang (a). The system discussed here and that described by Hochgesang were developed independently. PROCEDURE
The spectrum to he reduced is plsoed upon the viewing screen under a glass plate. The aero traces st the end of the spectrum are matched against those a t the beginning to check the constancy of t h e galvanometer eeros during the run. The electrical zero is set in the Teleducer for each of the five galvanometer zero traces. Then the horizontal cross wire is set 9.50 inches above each base line, with a ruler as reference, and this deflection is set into the Teleducer. The cross wire is then set on the top of a peak. The proper galvanometer sensitivity button is de ressed and the readout initiated by tilting a toggle switch or pusging on a foot switch. The readout thus obtained contains the mess number, the galvanometer sensitivity factor, snd the gitlvanometer deflection. APPLICATIONS
Hydrocarbon Types in Gasolines. Gasoline fractions boiling in the three ranges, 85" to I14",114°to 156", and 156'to 180" C.,
1919
Figure 1. Contact Telereader
are analyzed for noncyclodklkanes, monacycloalkanes, dicycloalkanes, and monoaromatics by an unpublished method developed a t the Houston Manufacturing-Research laboratories of Shell Oil Co. Results obtained for two gasoline fractions hy use of the data reduction system agree well with those processed manually (Table I). The time required for manual measnrement of the peak heights was 6 minutes; with the data reduction system, 2 minutes. I n order to permit measurement of all the peaks without moving the chart, a mmll change in sweep rate of the mass 'pectrometer had to he made. If the chart must be moved during a set ~ of measurements, ~ more ~ than 2 minutes ~ will he~ needed to i complete the measurements. Hydrocarbon Types in Midrange Fractions. These fractions in two r6nges, IsOD ta 2500 and 2500 to 3250 C, They sse subjected to chromatography, and nonwcloakanes, noncondensed cyclaalkannes, condensed cyeloalkanes, monoaromatics, and dittromatics are determined in the saturate portion ( 1 ) . The spectrum to be measured with the data reduction system is scanned a t a rate such that the region to be measured lies entirely within the length of the viewing screen of the contact Telereader. Results obtained for this analysis with the manual method and with the data reduction system are in agreement (Table 11). The time required to measure the peaks used in the saturate inverse by the manual method was 10 minutes, not inoliiding the time to transfer these measurements to a data sheet; using the data reduotion system, i t was 3 minutes. No subsequent transfer of data WBR needed in this case. Hydrocarbon Types in Sediment Extracts. This analysis is based upon a parent peak method developed for the aialysis of petroleum waxes (3). It requires the measurement of peak heights from mass 100 to mass 500 to 600. The time required to measure 400 to 500 peaks mrtnuitlly is about 1.5 to 2 hours, d e pending upon the degree of fatigue of the operator; w t h the data reduction system, it is 20 to 27 minutes. A comparison of peak heights as measured by the two systems showed satisfaotory agreement. A frequently recurring difference of 2 between readings of the same peak by the two methods suggests B difference B significant in the choice of base hnes. This would not c a u ~ variation in analyses obtained with the two systems of measuring peak heights. Hydrocarbon Patterns. Pattern coefficients calculated from measurements on the same n-butane spectrum made manually and with the data reduction system showed satisfactory agreement. For pattern coefficients greater than LOO%, the agree-
ANALYTICAL CHEMISTRY
1520
ment was 99% or better. For coefficients of the order 0.107, the agreement was a t least 90%. The same satisfactory agreement was obtained for pattern coefficients of n-hexadecane. Pattern coefficients can also be determined directly with the data reduction system by setting the highest deflection in a given galvanometer trace into the Teleducer as 999 and then confining all subsequent peak height measurements to that particular galvanometer trace. The agreement of the pattern coefficients obtained in this way with those calculated from the manual measurements was poor. As might be expected, the greatcst deviations occurred in the case of low peak heights, which under normal conditions would be measured on a more sensitive galvanometer trace.
Table I.
Hydrocarbon Type Noncycloalkane Nonocycloalkane Dicycloalkane Monoaromatics CS
c7
0.5 0.0 0.0 0.0
CS CQ C1a a
Gasoline Analyses
85"-114° C. Fraction Manual DRSa, method, vol. % vol. % 70.2 70.4 28.6 28.4 0.0 0.0 1.3 1.3 0.8 0.8
114°-1560 C. Fraction Manual1-01. DRS, % method. vol. % 55.9 42.6 0.0 1.6
0.2 0.3 1.0
0.5 0.0 0.0 0.0
0.0
0.0
Table 11.
CONCLUSIOhS
The suggested system for reduction of mass spectrometric data has several noteworthy advantages over the manual method. It is considerably faster, with no loss of accuracy. It has almost eliminated operator fatigue, and has eliminated transfer of data from the chart to calculation forms. Its relative ease tends to encourage development of new high molecular weight methods by permitting reduction of spectra in a quantity that would be impracticable by the manual method. The system is remarkably trouble-free and requires only an hour for the training of an operator.
0
0 2
0 1 0 0
2 2 0 0
Data reduction system.
Analyses of a 180" to 250" C. Fraction
REPEATABILITY
An idea of the repeatability with which one peak can be measured was obtained when three different operators measured the same peak using both methods of measurement. The galvanometer zero was redetermined before each measurement. Because of the subjectivity inherent in repeated measurements of the same peak by the manual method, only one manual measurement was made by each operator. The repeatability obtained with the data reduction system was as good as that given by the manual procedure.
5.5 9
42 (3 0 0 1 6
Hydrocarbon Type Noncyoloalkanes Noncondensed cycloalkanes Condensed cycloalkanes Monoaromatics Diaromatics Data reduction system.
DRSa, Vol. %
Manual Method, VOI. %
ACKNOWLEDGMENT
The authors wish to acknowledge the material assistance of W. 0. Lease, who installed the equipment and made many modifications, and of H. W. Schutz and H. M. Hicks, who made the measurements. LITER4TURE CITED (1) Clerc, R. J . , Hood, A., O ' N e a l , AI. J., Jr., d s . 4 ~ .C H m f . 27, 568 (1955). (2) Hochgesang, F., Socony->lobi1 Research a n d D e v e l o p m e n t D e p a r t m e n t , Paulsboro, N. J., ASTAI E-14 meeting, San Francisco, Calif., 1955. (3) O ' N e a l , 11.J.,Jr., Wier, T. P . , Jr., - 4 ~ 4CHEM. ~ . 23,530 (1951). RECEIVEDfor review M a y 3, 1956. 84, Shell Development Co.
Accepted June 28, 1956.
Publication
Spectrophotometric Determination of Aluminum in Ferrous and Nonferrous Alloys Application of 8-Hydroxyquinaldine ROBERT J. HYNEK and LEWIS J. WRANGELL Research Laboratories, Allis-Chalmers Manufacturing Co., Milwaukee
The new analytical reagent, 8-hydroxyquinaldine, does not react with aluminum but does react with many elements that ordinarily interfere in the determination of aluminum. This property is used in a widely applicable method for the elimination of some interferences prior to the spectrophotometric determination of aluminum-8-hydroxyquinolinate in chloroform at 389 mp, A mercury cathode partially removes metallic ions. Interferences are ultimately eliminated by adjustment of the pH to 9.2, and use of 8-hydroxyquinaldine, chloroform, and hydrogen peroxide. For certain alloys, it is not necessary to use the mercury cathode. Gross separations of interferences can be accomplished by the direct application of S-hydroxyquinaldine and chloroform to the dissolved sample. The method has been applied to a wide variety of standard and commercial alloys containing up to 10% aluminum.
P
1, Wir.
RIOR to the application of 8-hydroxyquinaldine to the determination of aluminum. the Allis-Chalmers Research Laboratories has used three methods of analysis, all of which have disadvantages. The first method is a colorimetric procedure for a variety of ferrous and nonferrous alloys. After an initial separation of interferences by sodium hydroxide, the aluminum is precipitated as aluminum phosphate. The blue phosphomolybdate complex is developed from the phosphorus contained in the precipitate, and the light absorption is measured in a filter photometer. The aluminum equivalent of the complex is obtained from a calibration curve. This method is limited in precision, and it is difficult to eliminate interferences completely. Furthermore, if the aluminum content is less than 0.0107; in steel and iron, a lengthy ether extraction is necessary to remove iron from the larger samples required to provide sufficient aluminum for a satisfactory determination. The second method, used for ferrous alloys, involves a partial separation of iron from alumin\im I)>- the use of sodium bicarbon-