Application of Infrared Internal Reflection Spectrometry to Studies of the Electrochemical Reduction of Carbonyl Compounds David R . Tallant and Dennis H. Evans Department of Chemistry, Unicersity of Wisconsin, Madison, Wisc. 53706
MOSTof the studies of electrochemical reactions b y means of spectral observations using optically transparent electrodes have been confined to the ultraviolet and visible regions of the spectrum. Only three examples of extensions into the potentially more useful infrared region (1-3) have been reported. In one of these ( I ) an optically transparent thinlayer electrode (OTTLE) was constructed using sodium chloride plates with several infrared-transparent gold minigrids sandwiched between them. The electrochemical reduction of ninhydrin in acetonitrile was studied. The sensitivity of the technique was excellent and the infrared spectra of the products of the ninhydrin reduction proved valuable in interpreting the nature of the reaction. In earlier work ( 2 ) the internal reflection spectrometric (or attenuated total reflection, ATR) method was applied in the infrared region using n-type semiconducting germanium plate-electrodes. Qualitative changes in infrared spectra were obtained during electrolyses of several species a t the germanium plate, but n o indication of the sensitivity of the method was provided and only qualitative correlations between the observed spectra and the electrode potential were obtained. In another investigation using internal reflection spectrometry with germanium electrodes (3) the infrared absorptions of the ions of supporting electrolytes in aqueous solution were monitored as a function of potential and were rationalized in terms of the expected ionic concentrations in the double layer. In connection with studies of the electrochemical reduction of aldehydes and ketones (4, 5 ) a further investigation of the internal reflection spectrometric method for observing infrared spectra during electrolysis was initiated. The present report summarizes the results of this investigation. EXPERIMENTAL
Dimethylsulfoxide (DMSO) was obtained from J. T. Baker Chemical Co. ( 8 T change. M, moderate: 5-8x T change. W, weak: < 5 % T change with respect to transmittance before electrolysis. Concentration of electroactive species: 0.2M. Only observed at control potentials >0.6 V cathodic of Ell2. 4
centrations of quinone and radical anion would be established very rapidly, and the rate of growth of the band at 1500 cm-1 should be limited only by the instrument response time, a few seconds. Instead, almost a minute is required before a constant absorbance is obtained. This reflects the time required for a steady state electrode potential distribution to be attained. Electrolysis begins at the edges of the plate and moves toward the center. This slow response prohibits quantitative comparison between observed and predicted absorbances as has been accomplished in the visible region for various electrochemical methods (9-12). Further evidence that the band at 1500 cm-l is caused by the radical anion is afforded by the cyclic potential scan experiment depicted in Figure 4. This experiment was performed with 10 m M quinone, about the lowest concentration which may be studied with the present instrumentation. The currents were still as large as 2.5 mA so resistance effects were appreciable. The absorbance increases as the potential is made more negative. (The small absorbance maximum in the region of -0.3 to -0.8 V was mentioned earlier. In this case a scan with only supporting electrolyte was performed prior to obtaining the data in Figure 3.) The absorbance reaches a maximum at about -1.8 V and then decreases as more of the plate reaches potentials on the second wave where the surface concentration of the radical anion is low. The entire plate may not be at potentials on the second wave even at a command potential of -2.5 V, yet the absorbance has decreased to less than one half its peak value. i f the product at the second wave absorbs at 1500 cm-I, its absorbance must be less than that of the radical anion. On the return scan, the absorbance again increases as the potential of most of the plate returns to
(9) W. N. Hansen, T. Kuwana, and R. A. Osteryoung, ANAL. CHEM., 38, 1810 (1966). (10) B. S. Pons, J. S. Mattson, L. 0. Winstrom, and H. B. Mark, Jr., ibid., 39, 685 (1967). (11) W. N. Hansen, R. A. Osteryoung, and T. Kuwana, J. Amer. Chem. Soc., 88,1062 (1966). (12) A. Prostak, H. B. Mark, Jr., and W. N. Hansen, J. Phys. Chem., 72, 2576 (1968).
I
I
0 0 0 0 0 0
e t
0
g
0
O
eo
0
a .010
0 0 0 0
0
0 0 0
..oo
0
a Q
0
0
.om4
0
0
0 0 00
I
1
00-
I
t
I
I
values on the first wave. Finally, the absorbance decreases as the potential of the center of the plate (and later the edges) becomes more anodic than the first El,*and the surface concentration of the radical anion is again zero. The slow scan rate was chosen both to minimize the current and to accommodate the slow instrument response resulting from the necessary noise suppression adjustments when 10 X scare expansion is used. The combination of iR effects and the slow movement of the electrolysis “front” from the edge to the center of the electrode causes the absorbance changes on the cathodic scan to occur at more negative potentials than predicted from the half-wave potentials. Nevertheless Figure 4 indicates that the 1500 cm-’ absorption is caused by the product of the first wave, the radical anion, and the average surface concentration of the radical is reduced as the potential of most of the plate reaches the second wave. For benzil, similar experiments indicate that the product absorption band at 1375 cm-’ is the major band of the benzil VOL. 41, NO. 6, MAY 1969
837
anion. The radical anions of benzophenone and acetophenone are more reactive and it has not been possible to identify which (if any) bands are caused by the radical anion and which are caused by other products. Further support for the assignment of the 1375 cm-l band to benzil radical anion is provided by the transmission infrared spectrum of this species. Benzil radical anion was generated by conventional controlled potential electrolysis using 0.1M TBAP in DMSO in an all-glass cell with mercury pool working electrode. The average n-value for electrolysis at -1.3 V us. PQRE was 1.04. The electrolysis solution had an anodic polarographic wave at the first E,/, of benzil and a cathodic wave at the second E,,z of benzil. Controlled potential coulometric oxidation of 5 m M radical anion 30 minutes after its formation indicated that 70 remained. In order to obtain a spectrum of the benzil radical anion, a 10 m M solution of benzil was electrolyzed to 95 completion (8.5 minutes), a sample of the dark blue electrolysis solution was transferred to a transmission infrared cell (0.1 mm) under nitrogen and the spectrum recorded within another 8 minutes. The spectrum exhibited no benzil absorptions, but a strong, broad band was found at 1375 cm-l. Thus the band observed at this frequency in the internal reflectionelectrolysis spectrum of benzil must be that of the radical anion. The bands at 1500 and 1375 cm-l for quinone and benzil radical anions, respectively, contain a strong carbonyl component which has been shifted to lower frequencies because the C-0 bond order is decreased in the radical anion. Very few spectra of species with C-0 bond order near 1.5 have been reported. All of the cyclic oxocarbon anions exhibit a strong absorption near that of the quinone radical
x x
anion [squarate, C40aZ-,1530 cm-l (13), croconate, C5062-, 1570 cm-l (13), and rhodizonate, CBOBz-,1500 cm-l (14)] and all have C-0 bond orders near 1.5. Considerably smaller shifts have been reported for the radical anion of ninhydrin (1) for which absorptions are found at 1710, 1680, and 1660 cm-l compared to 1750 and 1725 cm-l in unreduced ninhydrin. In the ninhydrin study the 1300-1600 cm-l region was obscured by intense solvent absorption. The structure of the 1-electron reduction product is not known with certainty ( I ) , so further comment on the small shift of the carbonyl bands would be premature. These studies indicate clearly that the internal reflection technique possesses sufficient sensitivity to be applied in the infrared region and that it is possible to identify those absorption bands caused by a given product or intermediate by observing the infrared spectra as a function of electrode potential. In order to improve the accuracy of potential control and to permit the acquisition of meaningful data at short electrolysis times it will be necessary to devise internal reflection plate-electrodes of much lower electrical resistance. Thin evaporated metal films (12) on germanium plates are now being investigated as a possible solution to this problem. RECEIVED for review January 6,1969. Accepted February 26, 1969. Research supported by the Wisconsin Alumni Research Foundation and the National Science Foundation, Grant GP-8350.
(13) M. Ito and R. West, J. Amer. Chem. Soc., 85, 2580 (1963). (14) R. West, H.-Y. Niu, D. L. Powell, and M. V. Evans, ibid., 82,6204 (1960).
Determination of Amino Hydrogen in Water Application to Residual Chlorine Analysis Jack L. Lambert and Jorge Olguin Department of Chemistry, Kansas State Uniuersity, Manhattan, Kan. 66502
IN A PREVIOUS PAPER, Zitomer and Lambert (1) described a method for the determination of trace amounts of ammonia by chlorination with hypochlorite in solution buffered at pH 5.2, followed by destruction of excess hypochlorite with nitrite ion, and spectrophotometric determination of the chloramine produced by use of the cadmium iodide-modified linear starch reagent. In a subsequent paper ( 2 ) , the inhibiting effect of bromide ion on the chlorination of ammonia to form trichloramine was used as the basis for the determination of bromide ion. This work has been extended to the differential determination of amino hydrogen in primary and secondary aliphatic amines by making use of the inhibiting effect of bromide ion on the chlorination of ammonia and primary amines. This differs from the method of Dahlgren ( 3 ) , in which chlorination of ethylamine, diethylamine, and triethylamine at three different pH values did not result in
complete differentiation. Application of the differential amino hydrogen determination to residual chlorine analysis is described.
(1) F. Zitomer and J . L. Lambert, ANAL.CHEM., 34,1738 (1962). (3F. Zitomer and J. L. Lambert, ibid., 35, 1731 (1963). (3) G. Dahlgren, ibid., 36,596(1964).
(4) “Standard Methods for the Examination of Water and Wastewater,” 12th ed., Am. Public Health Assoc., Inc., New York, N. Y.,1965.
838
0
ANALYTICAL CHEMISTRY
AMINOHYDROGEN
Solutions are prepared with ammonia-free water made by ion exchange (4). Absorbances are measured at 598 mp with 1-cm cuvettes. AMINESOLUTIONS FOR PREPARING CALIBRATION CURVES. Ten ppm stock solutions are prepared by weighing the appropriate amounts of ammonium chloride or amine hydrochloride and diluting to proper volume. Fifty ppb working solutions are prepared by further dilution. BUFFERSOLUTION.A solution of 136.0 grams of sodium acetate trihydrate and approximately 21.2 ml of glacial acetic acid per liter is adjusted to exactly pH 5.2 with a pH meter.