gas oils has been reported (1). Their concentration probably never exceeds 1 to 2% weight, and could be corrected for b y an independent determination of total oxygen (as in the case of nitrogen compounds). The F tandard deviation of this method is estimated at + l to 2% weight for e w h class reported, o n the basis of multiple determination for several of the sainples of Table IV. Linearity in the routine separation was checked for several samples; the recommended sample sizes are less than half the amount necessary to produce apprcviable dependence of the analysis results on sample size. The peculiar distribution of aromatic type LIS. boiling point in catalytically cracked samples has been noted by many workers in the past, and arises from the tendency of aromatic and olefinic components to he stripped of alkyl substituents down to a nucleus plus 1 to 4 alkyl carbon atoms. -1s a result, olefins concentrate in the C:: to CS range, monoaromatic’s predominate in the gasoline fraction, and the higher aromatic types are aistributed as in Figure 5 . The present method is not directly applicable to straight-run or related sample types, and may be inapplicable in principle to such samples. Absorptivity data for various diaromatic frac-
tions isolated from straightrrun sample types show considerable variability, (even after correction for alkyl sulfide content) and are generally substantially lower than values for corresponding cracked samples. This is believed to reflect the presence in straight-run samples of significant, but variable, amounts of the diphenylalkanes, conipounds that separate with the diaromatics and exhibit much lower absorptivities a t 230 mp. The diphenylalkanes would be expected to crack to gasolinerange benzenes, leaving little of this compound type in cracked gas oils. On the basis of preliminary data, the determination of monoaromatics in straight-run samples by the procedure described appears to give erroneously high results, for as yet undiscovered reasons. It is clear that a number of problems exist in the extension of this type of analysis to include straight-run sample types.
(3) Charlet, E. M., Lanneau, K. P., Johnson, F. B., Zbid., 26, 861 (1954). (4) Dietz, W. A., Dudenboetel, B. F.,
Priestly, W. Jr., presented before the Division of Petroleum Chemistry, 130th Meeting, Atlantic City, N. J., September 1956. ( 5 ) Flinn, R. A., Larson, 0. A., presented before the Division of Petroleum Chemistry, 138th Meeting, ACS, Sew York, N. Y., September 1960. ( 6 ) ,Friedel, R. A., Urchin, M., “Ultraviolet Spectra of Aromatic Compounds,” Wiley, Sew York, 1951. ( 7 ) Gordon, R. J., Nloore, R. J., Muller, c.E., ANAL.CHEM.30,1221 (1958). ( 8 ) Kearns, G. L., Maranowski, N. C., Crable, U. F., Zbid., 31, 1646 (1959). ( 9 ) Sixon, A. C., Thorpe, R. E., J .
( 1 ) Aczel, T., Bartz, K. W., Lumpkin,
Chem. Eng. Data 7 , 4 2 9 (1962). (10) Snyder, L. R.,ANAL. CHEM. 33, 1527 (1961). ( 1 1 ) Zbid.. I). 1535. ( l 2 j Ibid.: ‘p. 1538. (13) Ibid., 34, 771 (1962). (14) Snyder, L. R., J . Chromatog. - 6.. 22 (1961). (15) Zbid., 8,319 (1962). (16) Snyder, L. R., J . Phys. Chem. 67, 234 11963). (177 Zbid.: p. 240. (18) Ibid., p. 2344. (19) Snyder, L. R., Buell, B. E., AKAL. CHEM.36,767 (1964). (20) Snyder, L. R., Howard, H. E., Fergueson, W. C., Ibid., 35, 1676 (1963). (21) Snyder, L. R., Roth, W. F., Ibid., 36, 128 (1964).
34,1821 (1962). (2) Bartz, K. W., Aczel, T., Lumpkin, H. E., Stehling, F. C., Zbzd., p. 1814.
RECEIVEDfor review October 21, 1963. Accepted January 10, 1964.
ACKNOWLEDGMENT
The author thanks F. 0. Wood and -4.E. Youngman of this laboratory for assistance in the experimental work. LITERATURE CITED
H. E., Stehling, F. C., ANAL.CHEY.
Spectroplhotometric Titration of Primary Aliphatic Amides WILLIAM R. POST trnd CHARLES A. REYNOLDS Department of Chemistry, University of Kansas, Lawrence, Kan.
bA spectrophotometric titration method has been developed for the determination of primary aliphatic amides. The titration is carried out in an aqueous bromide solution, buffered a t pH 10 with a staidard solution of calcium hypochlorite. The end point i s determined from the ultraviolet absorbance of the excels hypobromite.
B
of the nonreactive nature of the amide functional group, most of the available analytical procedures for amides are diflicult and timeconsuming. For example, saponification procedures for amides require at a minimum 30 minutes reaction time ( 6 ) , and procedures based upon conversion to hydroxamic acids require a 1- to 6-hour reaction lime (1, 7 ) . The lithium aluminum hydride method (8), although relatively specific, requires a steam distillation step after a 15minute reaction period and the proECAUSE
cedure cannot be used for dilute aqueous solutions of amides. h’onaqueous acidbase methods (3, $) are not specific for amides, since many other weakly basic substances interfere. Probably the most generally applicable method for primary amides is that developed b y Mitchell and Ashby ( 5 ) which utilizes 3,5-dinitrobenzoyl chloride as a reagent, but even this procedure requires a reaction time of from 30 minutes to 1 hour and also requires a blank titration for every sample. I n the present work the first reaction step in the familiar Hofmann rearrangement of a n amide to an amine has adapted to a spectrophotometric titration procedure for primary aliphatic amides. A sample of amide is dissolved in an aqueous solution which is 0.1M i n potassium bromide and which is strongly buffered at p H 10 with borax. The solution is titrated with a standard solution of calcium hypochlorite, and the hypobromite ion produced in situ
immediately reacts with amide to form the N-bromoamide. The titration is followed by observing the absorbance of hypobromite a t 350 mF. A complete titration can be finished in 5 to 10 minutes. EXPERIMENTAL
Reagent and Apparatus. An approximately O.1.V solution of calcium hypochlorite was prepared b y dissolving reagent grade calcium hypochlorite in water, followed b y filtration t o remove solid calcium carbonate. T h e solution was standardized iodometrically. T h e titer of this standard solution decreased approximately 0.2y0per week. A borax butrer solution was prepared by adding a concentrated, carbonatefree, sodium hydroxide solution t o a saturated solution of sodium tetraborate until a p H of 10.0 was reached. Most of the amides used were of the highest purity obtainable from commercial sources and were used without VOL. 36, NO. 4, APRIL 1964
781
Table I. Photometric Titration of Primary Aliphatic Amides
So. of
Av.
purity, Std. Amide" trials yo dev.* Formamide 4 98.9 0.4 Acetamide 16 99.6 -0.4% Propionamide 4 100.6 0.6% n-Butyramide 4 99.4 1.0 n-Valeramide 4 99.7 0.9 a-Chloroacetamide 4 100.7 0.4 Trichloracet0.0 amide 3 103.0 0.3 Adipamide 4 96.5 0.6 Acrylamide 3 98.6 Furamide 6 100.5 1.3 0.1 Nicotinamide 3 101.0 Surrinimide 3 98.7 0.6 a Sample 0.15 mmole, except for adipamide and furamide, where 0.08 mniole was taken. b In the case of acetamide and propionamide, relative error is reported.
further purification. I n the case of acetamide and propionamide, several recrystallizations from benzene-ethyl acetate mixtures were performed, but no change in titration results were noted. The spectrophotometric titration apparatus has been described before (4). Procedure. A 10-ml. aliquot containing 0.15 mmole of amide was pipetted into a quartz titration cell, followed b v 5 ml. of 1M potassium bromide solution and 30 ml. of the borax buffer. The solution was then titrated with the standard hypochlorite solution. ,4bsorhance readings at 350 mp were taken when equilibrium had been reached after the addition of each increment of the titrant. Blank titrations were performed in the same manner except that 10 ml. of water were added instead of the amide aliquot. I n some of the re.;ult. reported here, actual absorbance reading. were taken only after an escess of hypochlorite had been added. KO ahsorbance measurements were made before the equivalence poiril. Instead, a standard first segment for every titration curve was prepared by measuring the absorbance of a freshly prepared solution of N-bromopropionamide at 350 mg calculating a molar absorptivity for thiq compound and then calculating the absorbance of the titration solution before the equivalence point of the titration assuming a stoichiometric yield of Nbromopropionamide. The intersection of this standard first segment with the best straight line drawn through the actual titration absorbance values was taken as the end point of the titration. RESULTS AND DISCUSSION
T h e results of the titration of a number of representative aliphatic amides are shown in Table I. Although most of the results reported in Table I were obtained with a 0.15 mmole sample, variation in the sample size from 0.08 to 0.34 mmole did not influence either the average result or the
782 *
ANALYTICAL CHEMISTRY
- , 0
rnl. of 0.1 N OCI-
Figure 1 . Spectrophotometric titration of propionamide
Figure 2. Absorption curves of species involved in propionamide titration . .
precision. The time to complete a titration ranged from 5 to 10 minutes. A typical titration curve is shown in Figure 1, where curve A represents a blank titration and curve B is a n actual titration curve for propionamide. I n all cases the slope of the rising leg of an actual titration curve was identical with that of a blank. The titration reaction is the first step in the Hofmann rearrangement reaction-Le., the formation of the N bromoamide from amide and hypobromite in aqueous solution:
H
H
Typical ultraviolet absorption spectra of the species involved in the titration reaction are illustrated in Figure 2 for the propionamide system. Curve A represents the absorption spectrum of propionamide itself, curve B the absorption spectrum of hypobromite at p H 10, and curve C that of N-bromopropionamide. The concentrations of all three substances were identical. It can be seen from Figure 2 that a t 350 mp hypobromite is the major absorbing species, while the N-bromoamide absorbs only slightly. I n the case of propionamide and the higher primary aliphatic amides the subsequent rearrangement of the N bromoamide occurs rapidly enough to cause errors in the titration procedure if too long a time is taken for the titration. The error is caused not by the absorbance of the rearrangement products at 350 mp, but by the decrease in absorbance of hypobromite due to the oxidation of the rearrangement products by hypobromite. To prevent this error, the hypochlorite is added quickly from the buret until enough has been added to overun the end point; absorbance readings due to the excess hypobromite
A = Propionamide 6 = Hypobromite at p H = 10 C = N-Bromopropionamide
are then taken at three or four increments of titrant past this point. The best straight line drawn through these three or four points constitutes the second leg of the titration curve (Figure 1).
The first leg of the titration curve in the case of these amides was constructed by measuring the absorbance of a freshly prepared solution of N bromopropionamide at 350 mp, calculating a molar absorptivity for this compound and then calculating the absorbance of the titration solution assuming a stoichiometric yield of N-bromopropionamide. This standard first leg of the titration curve was used in all results reported on propionamide, n-butyramide, n-valeramide, and adipamide. Because of the very small slope of this initial leg, changes in concentration of the amide being titrated caused almost no change in the final percentage purity calculated from the results of the titration. I n the case of all the other amides, except acrylamide, the first leg of the titration curve was experimentally measured rach time, since the subsequent rearrangement of the N-bromoamide occurred slowly enough that no appreciable oxidation of rearrangement products by hypobromite took place within at least 20 minutes. Acrylamide continues to react with hypobromite past the theoretical formation of N-bromoacrylamide, so that a direct titration of this amide is impossible. However, the absorbance of the titration solution a t 380 mp, after enough hypobromite has been added to overrun the N-bromoamide equivalence point, is a linear function of time over a period of about 6 to 10 minutes. Therefore, this absorbance-time plot can be used to yield a zero-time absorbance by extrapolation. A straight line with the slope of a blank titration drawn through this zero-time point is used as the second leg of the titration curve; the standard first segment of the N-bromopropionamide curve was used as the first leg of the titration curve. The results for acrylamide using this tech-
nique are as reproducible as the results for the other amides LLS shown in Table
I. A number of amides could not be determined by this procedure for a variety of reasons. Nonamide and oxamide are not soluble enough in water to produce a solution capable of titration. The formation of the A'bromo derivative of isobutyramide proceeds so slowly compared to the rearrangement of N-bromoisobutyramide t h a t no titration was feasible. I n the case of the malonaniide, consistently high results were dmays obtained, indicating that partial bromination of the methylene carbon atom u a s also taking place (a). The possible inter'erence of other
functional groups was investigated by titrating equimolar mixtures of acetamide and representive compounds containing these functional groups. Ester.;, alcohols, nitriles, carboxylic acids, and ethers did not interfere, but amines, aldehydes, and methyl ketones were oxidized by hypobromite under the conditions of the titration, and high results were obtained. However, preoxidation of the sample solution with bromine in neutral solution removes the interference of all three of these functional groups. Most aromatic amides and N-methyl formamide interfere with this procedure, but the presence of higher iY-a!kyl amides and the di-N-alkylamides do not cause any interference.
LITERATURE CITED
(1) Bergman, F., ANAL. CHEM.24, 1367 (19.52). (2) Freund, Be?. 17, 780 (1884). ( 3 ) Higuchi, T., Barnstein, C. H., \ - - - - ,
Ghassemi. H.. Perez. W. E.. ANAL. CHEM.34; 400 (1962).' ( 4 ) McKinney, R. K., Reynolds, C. A.,
Talanta 1, 46 (1958). ( 5 ) Mitchell, J., Jr., Ashby, C. E., J . A m . Chem. SOC.67,161 (1945). ( 6 ) Olsen, S., Dze Chemae 56,202 (1943). ( 7 ) Polya, J. B., Tardew, P. L., ANAL. CHEM.23, 1036 (1951). (8) Siggia, S., Stahl, C. R., Ibid., 27, 550-2 (1955). (9) Rimer, D. C., Ibid., 30,77,453 (1958).
RECEIVED for review October 17, 1963. Accepted December 20, 1963. This work was supported by the Directorate of Chemical Sciences, Air Force Office of Scientific Research.
Spectrometer Cells for Single and Multiple Internal Reflection Studies in Ultraviolet, Visible, Near Infrared, a n d Infrared Spectral Regions WILFORD N. HANSEN and JAMES A. HORTON North American Aviation Science Center, Canoga Park, Calif.
b Several types of spectrometer cells based on single and multiple internal reflections in the UV-VIS-NIR and IR spectral regions are described. The cells are designed for use in standard spectrometers. All arc? relatively simple, the simplest comprising a single Dove prism with means for holding the sample. These cells permit a well defined variable angle and polarization. Performance data are given.
T
reflection technique (2, 4, 7) is particilarly useful in obtaining spectra of liquids and solids with intense optical absorption in that troublesome interference effects are avoided and one need not work with thin samples of known thickness. I n addition, more informa .ion can be obtained from internal reflection spectra t h a n from transmission spectra. By taking two spectra of the same material at different angles or polarizations, spectra of the optical constants can be obtained (3, 5 ) t o describe completely t h e optical properties of the sample for the wavelength range involved. The optical constants cannot be obtained from transmission spectra alone. ?Vhen the angle of incidence is larger than the critical angle, the reflevtion has been referred to as attenuated total reflection (bTR). I n our studies of the in1 ernal reflection technique we have developed various HE INTERNAL
cells for use in standard spectrometers. Some of these cells are simple to make and use and should prove useful for routine work. A number of these cells will be described in this paper. EXPERIMENTAL
Single Reflection Cells. T h e simplest cell is a Dove prism as shown in Figure 1. (Various glass prisms may be obtained at low cost from Edmunds Scientific Co.) T h e sample is placed in optical contact with t h e internally reflecting face, and t h e cell is placed in the sample compartment of a standa r d spectrometer such t h a t the beam passes through as indicated. The light beam is inverted by a single internal reflection, but the beam path outside the prism is otherwise unchanged. While some light may be lost because of nonnormal incidence on the inlet and exit faces, there is little change in reflectivity for angles of incidence up to about 45'. Reflectivities a t the first face for normal incidence for prisms of indices of refraction 1.5, 2.0, and 2.4 are 0.04, 0.11, and 0.170, respectively. At 45' the reflectivities are 0.05, 0.12, and 0.179, respectively, an insignificant increase. I n actual practice, as indicated below, i t is optically more efficient to operate at angles other than normal. I n any case the angle must be greater than 60" to give significant power loss. Placing a reflection cell in the spectrometer beam will, in general, cause some defocusing. The transmittance of the cell itself is a direct measure of the
light lost from this and other causes. For a 45' Dove prism of KRS-5, 3 cm. long, placed in a 13eckman IR-4 XaC1 prism spectrophotometer, the transmittance was greater than 70% from 2 microns to 16 microns. For most of this range it was about 76%. ['sing a well adjusted instrument, these results have been obtained repeatedly. It might be thought that these results are impossibly high, in light of the fact that 67yc is the theoretical maximum for natural light. It should be noted, however, that the spectrometer accepts most effi.ciently light polarized parallel t o the plane of incidence of its prisms. This is just the polarization that is least rejected by the Dove prism. For the parallel component of the light, the theoretical maximum transmittance for a 45' KRS-5 Dove prism is 85%. The above facts show that the efficiency of the Dove prism is close to theoretical maximum, and t h a t the effects of defocusing are negligible for the case given. Using a Perkin-Elmer 421 grating spectrophotometer, the efficiency of the KRS-5 Dove ranged between 55 and 75%. Here the angular spread of the beam is much greater than above, yet the efficiency is still acceptable. I n the ultraviolet-visible-near infrared spectral region, using a Cary 14R spectrophotometer, the observed transmittances of glass Dove prisms were close to the theoretical maximum (ca. 90%). This was true when the monochromator was either before the sample or after the sample. It was true for Dove prisms varying in length from 4.5 inches to 2 cm. VOL. 36, NO. 4, APRIL 1964
a
783
