LITERATURE CITED
Table 1.
Titrations of Bases Using Acetous Fluorosulfonic Acid with Crystal Violet
Base Pyridine Quinoline a-Picoline Dimethylaniline Diethylaniline Potassium acetate
Table II.
Gram Equivalent Titrated x lo4 Experi- Theomental retical 8.21 8.23 5.91 5.95 10.62 10.54 6.81 6.82 26.71 27.00 12.77 12.74
Error,
yo
-0.24 $0.67 -0.75 -0.14 -1.07 -0.24
Color of Indicator a t End Point Green Green Green Green Green Green
Titrations of Bases Using Acetous Fluorosulfonic Acid with Malachite Green
Base Pyridine Quinoline a-Picoline Dimethylaniline Diethylaniline Potassium acetate ~~~~
Titrant Concentration, Gram Equivalent Weight of per Liter Base, G r a m 0.18288 0,06492 0.13756 0.07679 0.31128 0.09594 0.17068 0.08250 0.21676 0.39800 0.21508 0.12498
Titrant Concentration, Gram Equivalent Weight of per Liter Base, Grams 0.06492 0.18288 0.13756 0.07674 0.31248 0.09594 0.17068 0.08250 0.21676 0.39800 0.12498 0.21508
For both indicators, the color changes at the end point were very sharp, and no precipitate formed in either case. Fluorosulfonic acid has been successfully used to estimate alkaloids and amino acids. The details regarding the estimation of acetates, bases, alkaloids, and amino acids are being published separately. I n addition to being a stronger acid,
Gram Equivalent Titrated X lo4 Experi- The+ mental retical 8.22 8.21 5.97 5.95 10.46 10.54 ~. 6.81 6.82 26.71 26.80 12.77 12.74
Error,
76
-0.12 -0.33
+, o- . x
. -
-0.14 -0.34 -0.24
Color of Indicator at End Point Yellow Yellow
ypiiow . ~ ~ ~
Yellow Yellow Yellow
fluorosulfonic acid has the advantage that it is available in a pure form and its solutions can be completely anhydrous. With perchloric acid, however, this is not true. Perchloric acid solutions have to be prepared indirectly, or water has to be removed with acetic anhydride and other such reagents. Pure perchloric acid is unstable, and only a 70% solution is available.
(1) Barr, J., Gillespie, R. J., Robinson, E. A., Can. J . Chem. 39, 1266 (1961). (2) Conant, J. B., Hall, N. F., J . Am. Chem. SOC.49, 3047 (1927); 49, 3062 (1927). (3) Dimorth, K., hleyer-Brunot, H. G., Biochem. 2.323, 338 (1952). (4) Emeleus, H. J., Haszeldine, R. N., Paul. R. C.. J . Chem. SOC.1955. 563. (5) F1 (E
pharm. franc.
(7) Herd, R. I Sa'. Ed. 31, 9 (8) Kolthoff, I J . Am. C (9) Koltho 56, 1007 (1Y34). (10) Levi, L., Oestreicher, P . M ., Framilo, C. G.. Bull. Narcotics. U . N. Devt. Social 'Agairs 5, 15 (1953). (11) hlarkunas, P. C., Riddick, J. A., ANAL.CHEM.23, 337 (1951). (12) Meyer, J., Schramm, G., Z. anorg. Chem. 24, 206 (1932). (13) Paul, R. C., Sandhu, S. S., Singh, J. S., Singh, G. S., J . Indian Chem. SOC. 35, 877 (1958). .(14) Seaman, W., Allen, E., Ibid., 23, 592 (1951). (15) Shkodin, A. W., Izmailov, N. A., J . Gen. Chem. U.S.S.R. 20,39 (1950). (16) Siggia, S., Hanna, J. G., Kervenski. I. R., ANAL.CHEX 22, 1295 (1950). (17) Splenger, C. H., Kaelin, H. A., Pharm. Acta Helv. 18, 542 (1943). (18) Tomicek, O., Collection Czechostou. Chem. Commun. 13, 116 (1948). (19) Tuthill, S. M., Xolling, 0. W.,Roberts, K. H., ANAL.CHERT.32, 1679 f 1960).
(20) Wagner, C. D., Brown, R. H., Peters, E. D., J . -4m. Chem. SOC.69, 2609 (1947). (21) Woolf, A. A., J . Chem. SOC.1954, 2840. RECEIVEDfor review August 23, 1961. Accepted February 27, 1962.
infrared Determination of Aldehydes An Improved Group Type Analysis E. L. SAIER, L. R. COUSINS, and M. R. BASILA Gulf Research & Development Co., Pittsburgh 30, Pa.
b
An improved infrared procedure is described for the group type analyses of aldehydes. These analyses are based upon the integrated absorptivity of the aldehydic C-H stretching vibration. The application to both aromatic and aliphatic aldehydes i s discussed.
I
an effort to extend and improve previous work (7) pertaining to a group type analysis for oxygenated materials, a study was made of the aldehydic C-H stretching vibration in N
824
ANALYTICAL CHEMISTRY
a series of aromatic and aliphatic aldehydes over the region from 2600 em.-' to 2900 cm.-l Generally, two bands characteristic of aldehydes are observed in this region. The assignment of these bands has been discussed by several authors (1, 2, 4-6). The most recent work supports the hypothesis that the characteristic doublet occurring in this region results from a Fermi resonance interaction between the aldehydic C-H stretching fundamental and the first overtone of the aldehydic C-H bending vibration (8). This paper describes group type anal-
yses for both aromatic and aliphatic aldehydes. The application to aromatic aldehydes is new and the application to aliphatic aldehydes results in considerable improvement over our original method ( 7 ) . This improvement is due largely to the use of integrated absorptivities rather than peak values. EXPERIMENTAL
All data were obtained using a Perkin-Elmer Model 21 spectrophotometer equipped with a LiF prism. The wavelength scale was expanded to 50 cm. per micron and the scanning speed
* I-
*
I-
3000
2900 2800
2700
2600
FREQUENCY, crn:'
Figure 1. A. B.
Infrared spectra Benzaldehyde Heptanal
was 5l/2 minutes per micron. The aldehydes used were of the highest purity available and in several cases were redistilled to obtain higher purity, The solvent was Fisher certified reagent carbon tetrachloride. The cell was 14 mm. in length and was fitted with sodium chloride windows. The concentrations ranged from 2 X to mole of aldehyde per liter. 18 X Integrated absorptivities were calculated for all of the aldehydes. The solvent formed the base line for the aromatic aldehydes and the integral extended over both C-H bands from 2632 cm.-' to 2890 cm.-' The only exception was in the case of methylsubstituted aromatic aldehydes wherein it was necessary to draw the base line over this area. For the aliphatic aldehydes, only the lower frequency band was integrated (2632 cm.-l to 2778 cm.-l) and the base line was again drawn on the spectra. The base line technique was necessary because of interfering methyl and methylene absorption. RESULTS AND DISCUSSION
In this work all aldehydes studied were in such concentrations that the possibility of intermolecular hydrogen bonding was eliminated. Each aldehyde was measured a t several concentrations and no concentration-dependent frequency shift was observed. A total of 14 aromatic aldehydes and five aliphatic aldehydes was studied. All of the aromatic aldehydes showed two intense bands in the region from 2600 em.-' to 2900 cm.-l which in the majority of cases were accompanied by much weaker overtone and combination bands. The benzaldehyde spectrum shown in Figure 1,A , is a typicalexample. The aliphatic aldehydes also showed two bands, but the higher frequency band
was not as discrete as in the aromatic because of an overlap with the methyl and methylene absorption which is characteristic of the aliphatic-type compound. This overlap is illustrated in the spectrum of heptanal given in Figure 1, B. Since the evidence to date favors the interpretation that the characteristic doublet is caused by a Fermi resonance interaction (8),and if it is assumed that the intensity of the unperturbed bending overtone is negligible with respect to the unperturbed stretching fundamental, i t follows that the sum of the integrated absorptivities of the perturbed bands of the doublet is essentially equal to the integrated absorptivity of the unperturbed stretching fundamental (3). Thus, in settingup a group typeanalysis, both bands of the doublet should be integrated. This is possible when working with aromatic aldehydes which have no interfering methyl or methylene stretching absorptions. However, with the aliphatic aldehydes, the higher frequency doublet band is partially obscured by the methyl and methylene absorption, and the integration is limited to the lower frequency band only. These techniques are justified by the results. In Table I the calculated absorptivities for both the aromatic (integrated over both bands) and the aliphatic (integrated over the low frequency band) aldehydes are given. A group coefficient can be calculated for each type of aldehyde. The standard deviation calculated for the absorptivity of the aliphatic aldehyde group
Table 1.
Aromatic Aldehydes Benzaldehyde p-Nitrobenzaldehyde p-Bromobenzaldehyde p-Chlorobenzaldehyde m-Chlorobenzaldehyde o-Chlorobenzaldehyde p-Tolualdehyde m-Tolualdehyde o-Tolualdehyde p-H ydroxybenzaldehyde o-Hydroxybenzaldehyde p-Methoxybenzaldehyde
(I
2900 2800 2700 FREQUENCY cm?
2600
I
Figure 2. Infrared spectrum of typical oxygenated material
is of the order of &lo% and indicates the probable error in the analysis of their mixtures. Previously, this was =t25% ('7). We are also now able to analyze for aromatic aldehydes in the same manner with a probable error of =t127& again based on the standard deviation for the absorptivity for this group. Prior to this, no attempt had been made to analyze for the aromatic type. Beers' law behavior was confirmed by varying concentrations in the range from 2 X to 18 X 1 0 - ~mole per liter. We have recently published values for the relative intensities of the doublet bands for the para-substituted aromatic aldehydes (8). These relative intensities vary over a considerable range according to the degree of the Fermi resonance interaction. The substituent apparently influences the unperturbed frequencies to some extent
Summary of Aldehyde Data
Inteerated AbsoGtivity,
p-Dimethylaminobenzaldehyde 4-Hydroxy-3-methoxybenzaldehyde
Aliphatic Aldehydes Heptanal Octanal 2-Ethylhexanal Nonanal Decanal
3000
3.46-3.80 17.6 18.2 17.6 20.8 18.1 13.8 23.5 17.9 16.9 17.4 18.8 21.8 19.9 18.8
p
Mole/Liter Band I, X low3 Cm.-l 4.95 4.06 5.69 7.61 5.92 7.80 4.25 4.24 4.38 2.65 4.78 4.13 2.16 3.61
Av. 1 8 . 6 f 2 . 3 Absorptivity, Mole/Liter 3.6-3.8~ 7.0 7.3 8.0 8.1 8.9 7 . 9 & 0.8
x
10-3 3.72 2.98 3.25 3.20 2.65
Band 11,
2805. 2817" 2828" 2825" 2828" 2862" 2817" 2806" 2841 2804" 2839 2840" 2817" 2812"
Cm.-1 2726" 2722" 2729 2723 2722 2746 2729" 2719 27290 2729 2747 2732 2732" 2717
Band I, Cm.-l
Band 11, Cm. - l
2809 2825 2801 2810 2810
2710 2710 2693 2709 2710
Av. Band has shoulders which arise from overtone and combination bands.
VOL 34, NO. 7, JUNE 1962
825
causing variations in the degree of interaction. However, attempts to correlate the relative intensities with the Hammet u substituent constants failed, indicating that the influence is complex and not merely due to the relative electron donating or withdrawing properties of the substituent. The constancy of the integrated absorptivities of the aromatic aldehydes which extend over both bands of the doublet is in accord with our previous discussion. The constancy of the integrated absorptivities of the low frequency doublet band of the aliphatic aldehydes is somewhat surprising in the light of the relative intensity results obtained for the aromatic aldehydes. The probable explanation lies in the very simiIar structural environment of the aldehydic group
in the series of aliphatic aldehydes. It is possible that substituents on the alpha or beta carbon atom can strongly influence these results. This possibility has not been investigated. In applying group type absorptivities to the determination of aldehydes in the presence of other organic compounds, a base line procedure is always used and contributes greatly to the improved accuracy. The method of drawing this base line is illustrated in Figure 2. The sample was an oxygenated material and on analysis was found to contain 43.0 mole % aldehyde, 10.6 mole % ’ alcohol, and 0.4 mole % ’ acid. The need for a base line technique arises from general background absorption. Acids, in particular, contribute to this background absorption because
of their intense hydrogen bonded hydroxyl absorption in this region. LITERATURE CITED
(1) Eggers, D. F., Lindgren, W. C., ANAL. CHEM.28,1328 (1956). (2) Evans, J. C., Bernstein, H. J., Can. J . Chem. 34, 1083 (1956). (3) Herzberg, G., “Infrared and Raman Spectra of Polyatomic Molecules,” p. 266, Van Nostrand, New York, 1945. (4) Pinchas, S., ANAL. CHEM. 27, 2 119551. ( 5 j - 1 6 2 , 29,334 (1957). (6) Pozefsky, A., Coggeshall, N. D., Ibad., 23, 1611 (1951). (7) Saier, E. L.. Hughes, R. H.. Ibid.. 30,513(1958).’ ’ (8) Saier, E. L., Cousins, L. R., Basila, M. R., J . Phys. Chem. 66,235 (1962). ’
-
RECEIVEDfor review January 5, 1962. Accepted April 4, 1962.
Mass Spectrometric Determination of the Ratio of Branched to Normal Hydrocarbons Up to C,, in Fischer-Tropsch Product A. G. SHARKEY, Jr., J. L. SHULTZ, and R. A. FRIEDEL Pittsburgh Coal Research Center, Bureau of Mines, U . S. Department o f the Interior, Pittsburgh, Pa,
b A method is described for the determination of carbon number distribution and ratio of branched to normal paraffins in mixtures consisting primarily of saturated hydrocarbons. Separation of the normal and branched paraffins, CI1 to C18, is carried out using a Molecular Sieve technique similar to the elution method described by O’Connor and Norris. Mass spectra obtained before and after Molecular Sieve separation are compared using an internal standard, thus eliminating the need for recovering the normal paraffins from the sieve material. Calibration data for the branched-chain paraffins are obtained from fractions of Fischer-Tropsch synthesis product. Carbon number distribution data, CI1 to CB, and the ratio of branched to normal components, CII to C18, were obtained for the hydrocarbons in a Fischer-Tropsch synthesis product following hydrogenation. Agreement between determined and calculated values for normal paraffins, C, to C18, indicates the validity of Molecular Sieve methods in a range where the desired pure compounds are not available for synthetic blends.
T
Fischer-Tropsch synthesis is being investigated by the Bureau of Mines as a means of converting coal to hydrocarbons. Gaseous and liquid HE
826
m
ANALYTICAL CHEMISTRY
hydrocarbons from the Fischer-Tropsch synthesis have been studied by several investigators including Anderson, Friedel, and Storch (@, Weitkamp, et d . (14), Bruner (4), and Gall and Kipping (8). While carbon number distribution data for the hydrocarbons have in certain instances been obtained up to C ~ Oinformation , concerning the ratio of branched to normal hydrocarbons in Fischer-Tropsch product is not available above Ca. The desired information can be obtained for the paraffis plus olefins in the product if hydrogenated material is analyzed. The purpose of this investigation was to derive methods for determining carbon number distribution data, and the ratio of branched to normal paraffins, in hydrogenated fractions of Fischer-Tropsch synthesis product. Theoretical chain branching schemes for the hydrocarbons in Fischer-Tropsch product have been proposed by Anderson, Friedel, and Storch, and others (2, 6, 15, 16). The carbon number distribution for the hydrocarbons is also predicted by these schemes. Previous investigations have shown that these chain branching schemes adequately describe the synthesis product up to Cs ( 2 ) . For these comparisons, the individual isomers for the C3 t o CShydrocarbons were determined. Several mass spectrometric methods for determining the ratio of branched
to normal paraffins have been described, In all of these methods, certain assumptions have been made concerning fragmentation patterns and sensitivity factors for the branched paraffis. In the first investigation of this type, O’Neal and Wier assumed that the pure compound, 3-ethyl tetracosane, was representative of the isoalkanes in petroleum waxes ( l a ) . Brown et al. “felt that an average isoparaffin would dissociate so as to lose 43 mass units,” and therefore, used peaks corresponding t o (molecular weight -43 mass units) for isoparaffin determinations (3). Ferguson and Howard described a method for determining the is0 t o normal paraffin ratio in gasoline range petroleum ( 5 ) . The above authors used thermodynamic equilibrium data of Prosen, Pitzer, and Rossini to weight individual cs to c8 isoparaffins and obtain representative sensitivity factors a t each of these carbon numbers ( I S ) . Using Molecular Sieve techniques, average isoparaffin sensitivities were also obtained for the CS to Cl1 paraffins in representative gasolines. In this instance, Ferguson and Howard had to assume that a representative fraction of isoparaffis was obtained by the Molecular Sieve technique, as very few pure isoparafis above Clo are available for synthetic blend determinations. O’Connor and Korris described a complete analytical scheme for char-
