A Sealed Polarographic Cell for Prolonged Studies of Reactions at the Dropping Mercury Electrode Paul Arthur,’ David S. Rulison,* and K. Darrell Berlin3 Department of Chemistry, Oklahoma State University, Stillwater, O k l a . 74074
THATTHE PRODUCTS of electrolysis occurring at a dropping mercury electrode on occasion can be different from those produced at a larger electrode, such as a mercury pool, is a generally accepted fact. Various reports support this (1-3). Ordinarily, if information concerning the mechanism of oxidation or reduction of materials at the dropping mercury electrode is desired, coulometric data are obtained using a small volume of solution of the material of interest, such as approximately 0.2 to 0.3 ml. (4-6). If sufficient quantities of product are needed for analysis after electrolysis, a larger electrode, such as a mercury pool, is used (7,8). However, in some studies it may be desirable to obtain not only coulometric data using a dropping mercury electrode but also to produce sufficient quantities of materials from reaction at this electrode for product analysis by gas chromatography (GLC) and other methods. In view of the required time increase for electrolysis at a small electrode, solvent evaporation must be controlled in order to correlate concentration and wave height. Previous studies in these laboratories have shown that the situation is not greatly improved by presaturating the purging gas with the solvent being used (9). The apparatus described herein is comprised of a polarographic cell suitable for work with nonaqueous organic solutions. The cell may be hermetically sealed after flushing with a suitable gas. Interference by atmospheric oxygen is eliminated as is the need for a continuous stream of gas over the solution during electrolysis. Using a DME, successful electrolyses over a period of ten days have been performed with this apparatus. Coulometric data may be gathered during the period of the run, and sufficient amounts of products are formed for analysis by methods such as GLC. Standard materials have been used to check the accuracy of the coulometric determinations, and the apparatus has been found useful in experiments to determine the mechanisms of reductions at the DME. Although errors have occasionally been reported in connection with millicoulometry using very small solution volumes (IO), the present apparatus has yielded accurate nvalues for all known materials studied. Deceased.
* Du Pont Teaching Fellow, 1967-68. a
To whom inquiries should be addressed.
(1) Robert I. Gelb and Louis Meites, J . Phys. Chem., 68, 2599
(1964). (2) H. A. Laitinen and T. J. Kneip, J . Am. Chem. SOC.,78, 736
(1956). (3) J. J. Lingane, Anal. Chim. Acta, 2, 584 (1948). (4) . , Robert D. Weaver and Gerald C. Whitnack. Anal. Chim. Acta, 18, 51 (1958). (5) G. A. Gilbert and E. K. Rideal, Trans. Faraday SOC.,47, 396 (1951). (6) M. Vajda and F. Ruff, “Polarography 1964,” Vol. 2, Graham J. Hills, Ed., Macmillan, London, 1966, p 759. (7) S. Wawzonek and R. C. Duty, J . Electrochem. SOC., 108, 1135 (1961). (8) David L. Smith and Philip J. Elving, J. Am. Chem. SOC.,84,
/c;\
SEPTUM FITTINGI
24/40
4W50
-PLATINUM HARNESS
Figure 1. Polarographic cell and associated equipment
EXPERIMENTAL
Apparatus. Part A in Figure 1 is a diagram of the main cell, of the three-electrode type and using the acetone calomel electrodes (ACE) originally reported by Arthur and Lyons (11). The lower part of the cell is constructed so that mercury from the dropping mercury electrode maintains a constant level. Waste mercury drains into the tube labelled Part B, as shown in the side view of the cell. The expanded portion ( E ) in the capillary tubing at the bottom of the cell was found to prevent mercury from trapping small pockets of the solution in the capillary tubing and carrying them along into the waste mercury collector. The main cell body has similarities to a cell originally used by Arthur and Vander Kam (12) except that the bridging tubes connecting the DME compartment to the reference electrodes have high-resistance asbestos fibers at each end and filler tubes so that a variety of electrolytes may be employed. The present cell has 100,000 ohm asbestos fibers at the DME compartment ends of the bridging tubes and 40,000 ohm asbestos fibers at the reference cell ends. The resistances of these fibers were measured with a conductance bridge using 0.1 M aqueous potassium chloride solution.
1412 (1962).
(9) Paul Arthur and R. A. Hilbig, Oklahoma State University, Stillwater, Okla., unpublished results, 1962. (10) This has been pointed out by others. See A. M. Wilson, J . Eiectroanal. Chem., 10,332 (1965).
(11) Paul Arthur and Harold Lyons, ANAL. CHEM.,24, 1422 (1952). (12) Paul Arthur and R. H. Vander Kam, ibid., 33,765 (1961). VOL. 40, NO. 8, JULY 1968
1389
The cell top is equipped with a stopcock for use as a gas vent and with a hook of 16-gauge platinum wire so that a small thermometer slung in a harness of similar wire may be suspended inside the cell. The design of the mercury column inlet avoids the usually leaky ports in the tops of many common cells through which rubber tubing containing mercury is passed. The tube entering the side of the cell and passing down inside is designed to avoid splashing solution onto the walls of the DME compartment. The length of the tube is adjusted so that the end is barely above the top of the solution, the volume of which is approximately 10 ml. Part C is a deoxygenator in which dry nitrogen bubbles through the solution in order to remove dissolved oxygen. Two gas streams are used, one rapid to flush out the entire system and a slower one to bubble through the solution. A portion of the solution is measured quantitatively in the calibrated side bulb; the solution is then allowed to flow into the main body of the cell. Instead of a glass frit, a bulb with 18 to 20 small holes is used to disperse the gas as it has been found that certain glass frits retain traces of water after being washed unless heated for a lengthy period. Part D is a relatively simple device to contain solutions of materials which, after electrolysis, are susceptible to air oxidation or hydrolysis due to water vapor in the atmosphere. Operation. The entire apparatus is assembled using liberal amounts of silicone stopcock lubricant. By means of a small squeeze bottle, triple-distilled mercury is forced into the cell to the desired level by way of the disconnected waste mercury tube. Part B is then clamped in place. Dry nitrogen, which is presaturated with the desired solvent, is admitted through the inlet tubes at the top of the deoxygenator. A fast stream enters the left tube and passes through this into the main cell body. A slower stream enters the compartment in the deoxygenator which will later contain the solution. Both of these streams vent through the stopcock in the top of the cell. Ordinarily a simple device such as a gas dispersion tube in a test tube containing a relatively nonvolatile solvent is used to estimate the rate of flow of gas. Rubber septums are placed in the mouths of the bridging tubes on the main cell. Nitrogen is admitted to these through 4-inch, 16-gauge hypodermic needles fastened to the ends of pieces of rubber tubing connected to the gas source. Smaller 22-gauge needles are used to vent the nitrogen. The gas streams are allowed to flush the entire apparatus in this manner for approximately 2 hours. After this time, 3 ml of a solution containing an appropriate electrolyte are injected into each of the bridging tubes by means of a hypodermic syringe. Solutions (0.1M) of lithium chloride, lithium perchlorate, and tetra-N-butylammonium perchlorate in 2-propanol, 2-butanone, and acetonitrile, respectively, have all been used successfully in this work. A large volume (perhaps 25 to 30 ml) of the solution of the material for electrolysis is then injected through a previously inserted rubber septum in the top of the deoxygenator. Nitrogen is bubbled through these solutions for approximately 30 minutes, The inlet and venting needles are then removed from the septums in the tops of the bridging tubes. The stopcocks in the deoxygenator are maneuvered in such a manner as to force some of the deoxygenated solution into the calibrated side bulb. This solution is then allowed to flow out of the deoxygenator and into the DME compartment in the cell. This procedure of deoxygenation is the reverse of that originally recommended by Arthur and Lyons ( I I ) , but the end result is identical. In succession, Stopcocks 1-3 are then closed and the nitrogen supply is stopped. Stopcock 4 is opened to allow excess mercury to drain, and the reference electrodes are placed in the appropriate compartments. The rack holding the entire apparatus is then lowered so that the lower part of the cell is immersed in a paraffin oil bath maintained at 25.0 f 0.1 “C. After allowing sufficient time for thermal equilibration, the 1390
0
ANALYTICAL CHEMISTRY
electrodes are connected to the polarograph and an initial polarogram is taken. A Sargent Model XXI Polarograph was used in these studies, along with the iR Compensator originally developed by Arthur and Vander Kam ( I t ) and commercially available from E. H. Sargent Co. The cell resistance is well within the range of the iR Compensator. The appropriate constant voltage to be applied during electrolysis in each case is determined by examination of the initial polarogram and is obtained from a commonly available stable external battery. The value of this voltage is periodically measured by means of a student potentiometer. Using an accurate electrical timer, the time is measured throughout the electrolysis and periodically the wave height is measured. When sufficient electrolysis has occurred, as determined by examination of the decrease in the polarographic wave height, Stopcock 2 is opened and nitrogen pressure is applied to the cell in order to force the mercury out of the constant level tube into the reservoir. When all the mercury has been so removed, Stopcock 4 is closed and Part B is removed. Part D,which has been thoroughly flushed with dry nitrogen, is then clamped in place. Stopcocks 5 and 6 on Part D are then opened, as is Stopcock 4 on the main cell. Nitrogen pressure forces the solution out of the cell and into Part D. Stopcock 6 is then closed and Part D is rotated on the spherical joint. Stopcock 6 is then opened very slightly and any mercury may be expelled which was inadvertently transferred along with the solution. Using this procedure, only one or two drops of solution are lost along with the mercury. Both stopcocks on Part D are now closed and the electrolyzed sample may be stored for future use, such as for product analysis using various analytical techniques. The septum fitting on Part D permits convenient removal of electrolyzed sample by means of hypodermic syringe. The coulometric data obtained during the electrolysis may be employed in calculations to obtain the number of electrons transferred per molecule or ion. In this work, mathematical integration of an equation identical in form to that discussed by Meites (13) is usually performed. RESULTS AND DISCUSSION
This apparatus has been used successfully during long POlarographic runs to determine n-values for the reduction of various substances, both organic and inorganic. Cadmium chloride in various concentrations in anhydrous 2-propanol with 0.1M lithium chloride yielded values of 1.95 to 2.04 electrons, with an average error of 1.8 %. The range of concentrations studied was from 5 x lO-4M to 3 x lO+M. The solvent was dried using a modification of themethod of Arthur, Haynes, and Varga (14). Studies have been made in acetonitrile dried in a manner similar to that used for 2-propanol. Lithium perchlorate in a concentration of 5 x lO-8M with 0.1M tetra-n-butylammonium perchlorate as carrier electrolyte yielded a value of 1.003 electrons for the reduction of lithium over a 15-hour period. The carrier electrolyte was prepared using a modification of the procedure reported by Cokal and Wise (15). Magnesium perchlorate under like conditions yielded a value of 2.012 electrons. Organic materials have been studied in acetonitrile. With 0.2M tetra-n-butylammonium perchlorate as carrier electrolyte and 0.04M benzoic acid to serve as proton source, (13) Louis Meites, “Polarographic Techniques,” Interscience, New York, 1965, p 532. (14) Paul Arthur, W. M. Haynes, and L. P. Varga, ANAL.CHEM., - 38, 1630 (1966). (15) E. J. Cokal and E. N. Wise, J. Electround. Chern., 11,406-15
(1966).
benzil yields two well defined waves of approximately equal height. The half-wave potentials are 0.772 and approximately 1.57 V us. ACE. Under identical conditions, benzoin yields a wave identical in shape to that of the second benzil wave and with identical half-wave potential, indicating that benzoin is the product of reduction for the first wave. When benzil is reduced at a potential on top of the first wave, concentrations of up to 0.02M yielded a n-value of 2.04 electrons (16). It is interesting to note that the “depletion effect” discussed by Reynolds and Shalgosky (17) evidently did not interfere with the determinations of n-values as in each case the value obtained was within a few per cent of an integral number. Using conditions identical to those with benzil, solutions of diethyl benzoylphosphonate (18) have been electrolyzed for lengthy periods. A value of 2.03 electrons was found for the (16) A similar reduction in an aqueous system at low pH values has been reported. See Robert H. Philp, Jr., Robert L. Flurry, and R. A. Day, Jr., J . Electrochem. SOC.,111,328 (1964). (17) G.F. Reynolds and H. I. Shalgosky, Anal. Chim. Acra, 10, 386 (1954). (18) K. D.Berlin and H. A. Taylor, J . Am. Chem. SOC.,86, 3862 (1964).
reduction, and sufficient amounts of products were formed to permit analysis by GLC. The reduction was found to yield diethyl a-hydroxybenzylphosphonate as the only major product. In the absence of a proton source, fractional electron values are obtained, and benzoin is found to be the product. Further work on esters of this nature will be discussed in a forthcoming report. CONCLUSIONS
The apparatus described herein has been found useful for determining accurate coulometric values for various metal ions and organic materials. Sufficient reductions of two organic compounds at the DME have been accomplished so that product analysis was performed with accuracy by GLC. Various organic oxidations or reductions could be determined in a like manner. In addition, inorganic complexes could be electrolyzed and the mechanisms examined by coulometric determinations and product analysis. RECEIVED for review February 1, 1968. Accepted March 8, 1968. Work partially supported by the Public Health Service, G M 10367-06.
Determination of Complex Constants by Gas Chromatography H. Schneckol Organisch-chemisches Institut, Universitat Mainz, Germany
DISTRIBUTION MEASUREMENTS to determine complex formation constants were first used 30 years ago. In 1938, Winstein and Lucas determined equilibrium constants for olefin-Ag+ complexes by distribution between two liquid phases (CCla and aqueous salt solutions of constant ionic strength) (I). Recently, Muhs and Weiss applied gas chromatography (GC) for similar investigations by using AgNO3 loaded liquid phases (2). This very elegant and fast method suffers from the disadvantage that the plots of partition coefficient H us. AgN03 concentration are not linear because of a salting-out effect at high AgN03-concentration; an empirical approach had to be developed in order to evaluate slope and intersection with the abscissa required for calculation of the equilibrium constant K1, according to the Equation (2)
where KL is the partition coefficient of the uncomplexed solute in 4 M LiN03 solution. Interest in both procedures has focussed on the determination of olefin-AgN03 complexes ( I , 2). In the investigation of the polymerizability of unsaturated nitriles by AgN08 (3), we tried to apply the latter method to find out whether complex formation between salt and monomer took place. It was felt, however, that by maintaining constant ionic strength of the liquid phase, the salting-out effect might be eliminated, 1 Present address, Dunlop Research Centre, Sheridan Park, Ontario, Canada.
Table I. Equilibrium Complex Constants Kl and Partition Coefficients of Some Unsaturated Compounds with Ag+ at 61 “C Compound KL KI(I/mole) Acrylonitrile 85 0.703 a-Chloro acrylonitrile 51 0.283 Methacrylonitrile 65 0.774 Crotononitrile 165 0.940 Methyl methacrylate 51 0.0
thus improving the technique. LiN03 was used to maintain = 4. It is evident from Figure 1 that both partition coefficient and complex constant can be determined directly by the improved technique from the straight lines obtained. The values calculated are seen in Table I. The linear shape of all lines indicates that a) there is no interaction between LiN03 and the solute, b) the salting-out effect of both nitrates (Ag and Li) is at least similar. The complex constants were also determined by the older method ( I ) in aqueous lMAgNOa/NaNOs-benzene at 25 “C. In all cases, the equilibrium constants were larger by a factor of 5 compared to those of Table I. This increase might be due to the different ionic environment and to the lower temperature. The formation of weak complexes between AgNO3 and benzene ( 4 ) might, however, complicate the interI.(
(1) S. Winstein and H. J. Lucas, J. Amer. Chem. SOC.,60, 836
(1938). (2) M. A. Muhs and F. T. Weiss, ibid., 84,4697 (1962). (3) H.Schnecko, Makromolekulare Chem. 111, 146 (1968).
(4) L. J. Andrews and R. M. Keefer, J. Amer. Chem. SOC.,71,
3644 (1949). VOL. 40, NO. 8, JULY 1968
0
1391
