Polarographic Behavior of Organic Compounds Efect of Ionic Strength, B u f e r Nature and Concentration, and pH PHILIP J. ELVING, JOSEPH C. KOMYATHY, ROBERT E. VAN ATTA, CHING-SUNG TANG, AND ISADORE ROSENTHAL The Pennsylcunia State College, State College, P a . In the development of analytical procedures and in general polarographic investigation involving organic compounds, the importance of calibration and study at the same level of ionic strength and buffer characteristics is indicated as a result of a survey of the factors affecting polarographic behavior. In order to evahtate the effects of the medium on polarographic behavior, a study has been made of a-bromo-n-butyric acid over the pI1 range of 1 to 12 and the normal working range (0.1 to 3 M) of ionic strength, using most of the commonly encountered buffer systems. The change in E0.6 was found to be dependent upon the nature of the buffer components, the concentration of the buffer components, the pH of the solution, and the ionic strength of the solution. Regions of ionic strength exist in which the measurement of Ea.&will yield values which are more valid when used to compare the results of various investigators than Eo.5 values nieasured in other regions. In order to obtain consistent results which will lend themselves to duplication and to comparison upon theoretical bases, the test solutions should be made up of definite ionic strength, the value of which should he stated in the same manner as the routine statement of capillary constants. The exact natiire and concentration of buffer coniponents sh011lrf be specified.
LTHOUGH the importance of the effwt of pH on the polarographic reducibility of organic compounds has been demonstrated in recent years, there has been no systematic study of the effect of ionic strength and its variat,ion due to the use of variws buffers in the polarographic invest,igation of organic compounds. The u d practice in buffering solut,ionr for organic polarographic study is to u,w a standard buffer. or set of buffers at the concentrations listed in t'he literature--e.g., the handbook. However, any one buffer, even if used over a pH range of only 2 units, c m vary tremendously in iouic strength. For. example, in the cme of a phosphate or citrate buffer uscd in the region of pH equal to the pK, of the third ionization, t,he ionic strength CFQ change by a factor of 9 / 4 over the recommended pH interval of 2 ( p K f 1). Such changes in ionic strength with one buffer over a pH range or between two buffers at any one pH are probably responsible for much of t,he confusion arising when attempts u e made to duplicate reported value. In addition, even if two different buffers at the same ionic strength are used, polarographic results may vary considerably, depending on the nature of the buffer components. These variations can lead to distortion that will completely mask the pH-dependence of the process. This type of confusion has led to the attempt made in this paper a t evaluating the phenomena of variation of polarographic behavior with ionic strengt,h and buffer composition. One of the earliest observations of the effect of specific buffer components on the half-wave potential, E0.6, was made by Furman and Stone ( 5 )in the course of the investigation of the polarographic behavior of several anthraquinones in various commonly used buffering media; shifting of the Eo.5 to more negative values was taken as evidence of complex formation with borate and phosphate buffers. On further investigation, Stone (9) found that buffer constituents may play a vital role in the polarographic reduction, if a relatively stable species is formed between the buffer anion and the reducible material or one of its reduction products. DeFord and Andersen ( 2 ) have studied the variation of the Eo.sof cadmium as a function of ionic strength in various supporting electrolytes of ionic strengths extending up to 12 M ; flo.' a t first became more negative as t'he ionic strength increased
and then shifted to lays nrgativc values with continued increase of ionic strength, the magnitude of the shift depending on the nature and total ionic strength of the supporting electrolyte. Elving, Rosenthal, and Kramer (3),in a polarographic investigation of iodoacetic acid and the bromoacetic acids, obtained significant variations in the Eo1 in rome cases where two or more buffers at the same pH and ionic strength were used. The variation of the buffer component concentration has been found to affevt the Eo of reducible organic compounds (I,$, 4 , 8 ) . -4s such phenomena arc of great importance to the interpretation of the data obtained in the polarography of organic compounds, tJie polarographic behavior of a-bromo-n-butyric acid has been studied over the pII range of 1.0 to 12.4and the normal norking range (0.1 t o 3 M ) of ionic strength, using m w t of the commonly encountered buffer systems. Ionic strength as varied in two ways: by altering the buffer component concentration, and by adding potassium chloride. The concentration of the buffer components was varied in order to determine the magnitude of this effect on the Eo.&;it was hoped to obtain some indication as to whether interaction between the electroreducible substance and the buffer components occurred. Any variation of Eo6 under identical conditions including similar ionic strength, except that of buffer component concentration, could he interpreted as being due to the formation of complex species. It was felt that from the data obtained in this study, the optimum operating conditions for each buffer could be determined and Borne insight be gained into the nature of the reduction pr0c-s and the factors influencing it. The selection of a-bromo-n-butyric acid as the substmce to be used in this investigation was based on several factors. The Eo is within the readily measurable range of -0.3 to -1.4 volts for the pII range of 1 to 13; this potential range permit8 the use of all the common buffers, some of which give decomposition waves beginning at relatively low potentials. The compound itself is stable over the time required. In addition, the compound, containing a four-carbon chain, is not the introductory member of the series. It was felt best to avoid the introductory members of series, as they usually behave in an anomalous manner. The reduction of the a-bromo-n-butyric acid involves the fission
V O L U M E 23, NO. 9, S E P T E M B E R 1 9 5 1 of tlw carl)on-bromine bond and the form:ttion of ri-but \ ri( acid. The clwtrodc reaction is irreversible, as indicated by the slope of tlic w a w , the IlkoviE equation indicates thc tiio-electron reducbtion procws which results in current flow. The latter is a diffusion-controlled process, as indicated by the effects of variation of concentration of reducible \pecies, of head of mercury, and of temperature variation upon the current flow. While not of absolutely symmetrical charact cr, as iz the wave obtained in the case of a reversible speries, the wave is suficiently well defined to leave little question as to the determination of the polarographic values from it. R liile this particular compound yields the behavior reported, all other reducible organic species do not necessarily show the same type of behavior. However, the results reported indicate a behavior which may be found with other substances. I n particular, similar effects have been observed in the polarography of other halogenated compounds where the electrode reaction involves the fission of carbon-halogen bonds. Studies similar to the one described should be made of other irreversible organic systems as ne11 BS of reversible organic Bystems. It would be of considerable interest to investigate irreversible and reversible redox systems in which ionization does not occur, in order to determine whether the sigmoid curve and iso-Ep point phenomena subsequently discussed are associated with pHdependence. Such studies are nom in progress in the authors' laboraton
1219 I esistance
of the solutions was measured by means of a General Radio Co. Type 650-A impedance bridge. Preparation of Buffer Solutions. The buffers used in this investigation are described in Table I. Buffer A was prepared by f m t diluting the calculated amounts of hydrochloric acid and of potassium chloride to approximately 250 to 400 ml. The p H of the solution was then measured continuously during dilution t o approximately 475 ml. ; standard solution of hydrochloric acid or of sodium hydroxide was added to adjust any variation in p€I due to dilution. The solution was then transferred to a t500-ml. volumetric flask and diluted to mark. Buffers B, F, €I, and .J were similarly pi epared. Buffer C was prepared as needed from stock solutions of the tuc) components (0.22 and 0.67 .If citric acid and 0.11 and 0.33 'Id disodium hydrogen phosphate). The test solutions were prcpared by mixing stork solutions in the required proportions, diluting with an equal volume of a solution of a-bromo-n-butyric acid, adding the calculated amount of potassium chloride to adjust the ionic strength, adding small amounts of the required
~ _ _ _ _ _ _ _ _ _ ~ _ _ Table I.
Coniposition of Buffers
(Potassium chloride added as ionic strength component DesigBuffer nation 1" Ionization Conatants (8) HC1-KCl 1.0 1.0, 2.0. 3.0, 4.5 Ha citrate: 8 . 4 X IO-' Nar citrate-€IC1 1.8 x 10-6 4 0 x lo-; 2.2, 3.6. 5.7, 7.9 C HaPOd: 1.1 X 10HI citrste-XazliP0, 7 . 5 x 10-6 4 . 8 x 10-1: I) 3.5, 3.6, 3.7, 4.3, HOAc: 1.86 X lo-* HOAc-NaOAc 5.0, 5.2, .5.3 E 6.0, 6.1 F HaBOa: 6 . 4 X 10-10 7.8 ?;HIOH: 1.8 X 10-8 8.2, 8.5, 8.8 G H 9.9 10.4. 11 5 , 12.4 J W 3.5
Materials. All buffer components were analytical reagent grade chemicals. The a-bromo+-butyric acid (Eastman Iiodak white label) was recrystallized several times from pentane. The values of the melting point and the density agreed with those reported in the literature. The compound was polarographically pure. h solutions of a-bromo-n-butyric acid stood, slow decomposition of the ac!d Table 11. Polarographic 1alues and Conbtants of a-Hromo-n-butyric Acid occurred-e.g., a decrease of 25% in in Yarious Buffer Media the diffusion current of 0.5 millimolar a-bromo-n-butyric acid in buffer C a t (Potassium chloride added a$ ionic strength component) pH 5.7 nnd 7.9 was observed over :i Concentration Averaged Values of period of one month. On titration of Buffer the solution, the total acidity was found First Second a-Bromo-n!2 m component component butyric actd ~ 1 1 td
V O L U M E 23, NO. 9, S E P T E M B E R 1 9 5 1
stock solutions needed to adjust the test solution to the desired pH, and diluting to volume. Buffer D was prepared by dissolving the calculated amounts of sodium acetate and potassium chloride in 450 ml. of distilled water. The pH was measured continuously while standard acetic acid was added until the desired pH was obtained. The solution was then diluted to 500 nil. Buffer E was prepared by diluting the necessary amounts of the two components and of potassium chloride to 400 ml. The pH of the solution was then adjusted by adding stock solution of the requisite components. On dilution to 475 ml., the pH was again adjusted. The solution was then diluted to 500 ml. Buffer G was similarly prepared. Procedure. Solutions for electrolysis were prepared by pipetting 5 ml. of 10 millimolar a-bromo-n-butyric acid stock solution into a 100-ml. volumetric flask and diluting to the mark with the buffer being investigated. The electrolysis cell was rinsed with a portion of this solution and another portion was added
to the h a 1 bubbler in the nitrogen purification train. Nitrogen was then passed through the solution for 10 minutes, after which the tube was withdrawn and nitrogen allowed to pass over the surface of the solution during electrolysis. Values of t , the lifetime of the mercury drop, were determined for the limiting current portion of each curve. The pH of the test solution was measured after electrolysis. The polarographic values and constants were calculated in the usual manner. The experimental results obtained are summarized in Table I1 and Figures 1 and 2. Complete data containing individual averaged results may be obtained from the authors. Calculation of Ionic Strength. Values of the ionic strength were calculated on the basis of the constants given in Table I. Using the values of the ionization constants of the particular buffer component being considered, the concentrations of the various ionic species derived from the main buffer component were calculated. After determination of the concentrations of the ions, the ionic strength was calculated by the relation:
\There p is the ionic strength, M is the actual calculated concentration of the ion in moles per liter, and z is the number of unit charges on the ion. Where the concentration of a particular species was found to be very slight, as in the case of t h e a - b r o m o - n butyric acid, the contribution of such species was neglected in the calc u l a t i o n of t h e ionic strength.
Figure 2. Variation of pH with E0.s a t Known Values of Ionic Strength Curves are identified by Roman numerals; values i n parentheses indicate ionic strength. Broken lines indicate probable location of curves. Short lines at pH 8 t o 9 indicate values i n the NHICI-NHIOH buffer system. Line I represents low concentration of main buffer component; all others represent high concentrations.
The change in Eo of a-bromo-n-butyric acid was found to be dependent upon the nature of the buffer components, the concentration of the buffer components, the pH of the solution, and the ionic strength of the solution. In general, a variation in the concentration of the buffer comp o n e n t s of a n y one buffer system caused no appreciable difference in the slope of the curve of EO5 vs. ionic strength; this slope shows a slight but systematic change xith variation of pHe.g., buffers B and D of Figure 1. I o n i c strength generally seemed to have only slight if any effect
on the diffusion current constant. There is a variation of this constant with pH, which may be associated with a change in the magnitude of the diffusion coefficient in solutions of different buffer systems or buffer components. Effect of Ionic Strength. Variation in ionic strength had a marked effect on the EO6 of a-bromo-n-butyric acid for most of the buffers investigated. In general, buffers used in the alkaline region caused a greater shift of EO6 with change in ionic strength than buffers used in the acidic region. While an increase in ionic strength shifted EO6 to more negative values in the low pH region, the opposite was found to be true in the remaining pH region. With the sodium citrate-hydrochloric acid buffer a t pH 2.0 and with the acetic acid-sodium acetate buffer at pH 3.5 and 3.7, E o 6 was practically independent of ionic strength. The variation of EO5 with ionic strength was not the same with all buffers, nor WLLSthe variation uniform over the ionic strength range for every buffer investigated. K t h the sodium citrate-hydrochloric acid buffer st pH 1.0, 3.0, 3.1, and 4.5, the hydrochloric acidpotassium chloride buffer a t pH 1.0, the acetic arid-sodium acetate buffer a t pH 5.3, and the disodium hydrogen phosphatesodium hydroxide buffer a t pH 10.4, 11.5, and 12.4, the shift of EO5 was appreciable a t low ionic strengths but less pronounced as the ionic strength was increased. With the sodium citratehydrochloric buffer a t p H 3.1 and 4.5, as the ionic strength increased, Eo5 shifted to more negative values a t first and then, with continued addition of potassium chloride to increase the ionic strength, shifted to less negative values. The general shape of the Eo5 os. ionic strength curves produced by varying ionic strength is in agreement with the results obtained for cadmium by DeFord and Andersen (2). A slight shift was observed in the value of a, the apparent number of electrons involved in the potential-determining step, as the ionic strength of the solution was varied. The value of CY was mlculated by means of therelation: E l l 4 - E314 = 0.056/a. In most c a w , the value of CY increa-4 as the ionic strength increased. The variation reached a maximum in the unbuffered solution. In a few cases, the variation of a ~ i t ionic h strength was random. In all cases, a was less thuii 1, which indicated that the reaction a t the electrode was irreversible. The calculation of n in the IlkoviE equation gave a value of 2. Effect of B d e r Component Concentration. Care was taken that the concentrations of buffer systems used provided adequate buffering capacity. Buffer systems were used within the pH ranges recommended and, in general, within the pH range of pK, f. 1. Buffer systems which have been described as being sluggish in reaching equilibrium were avoided; some question in this respect has been raised concerning the borate system which was investigated, as it has been extensively used in polarographic studies. It can be seen from Figure 1 that a variation of the concentration of the buffer component itself will cause a shift in the Eos even though solutions of equal ionic strength and pH are electrolyzed. If pH and ionic strength are kept constant and a change in EO5 still occurs, this change is presumably due to some type of interaction between the reducible species and the main component of the buffer. This behavior can he noted a i t h several of the buffers investigated. In the case of the sodium citrate-hydrochloric acid a?d the hydrochloric acid-potassium chloride buffers, the essential difference betx-een these two buffers a t pH 1.0 is the presence of undissociated citric acid in the former. This undissociated citric acid makes no contribution to the ionic strength. Consequently, the shift of EO6 to more negative values with the sodium citrate-hydrochloric acid buffer is due to some effect other than that of ionic strength. This effect may possibly be due to some form of interaction. Similar effects have been found with other reducible compounds (3). In order to stress the significance of the experimental data, the following points should be emphasized:
ANALYTICAL CHEMISTRY In the absence of predominant ionic strength effects, if any interaction occurs between the buffer components and the rt+ ducible species, the EO.& will be shifted to more negative values. At low pH values, a plot of Eo.5 us. ionic strength &OWN greater curvature for the lower concentration of the main buffer component, while the op site is apparently true in respect to the curvature at high pgvalues. Dependin on the magnitude of the shift of as a function of the main tuffer com nent concentration and the de ree of curvature existing for high and low concentrations, $e two curves may or may not intersect as illustrated by buffers D a t pH 5.3, E a t pH 6.0, and J a t pIf 12.4, and indicated by buffers C a t pH 2.2, 3.6, 5.7, and 7.9, and D a t pH 3.6 and 5.0. At the intersection point of all curve3 on Figure 2, which has been designated as the “iso-Ep point,” the EO.5 is independent of ionic strength and main buffer component concentration. In alkaline solution, the increase in ionic strength with increasing pH, due to dissociation of the buffer components to form more highly charged anions, causes the EO to become less negative; this is opposed to the usual effect of the Eo5, becoming more negative as the pH is increased. In the more acidic region, the two effects reinforce each other.
Effect of pH. On plotting Eo.6 us. pH (Figure 2) for lines of equal ionic strength, afamily of &shaped curve^ is obtained which intersect a t pH 4.8. This S-shaped relation of Eo5 and pH is characteristic of acids (3). The significance of the common point, the iso-E, point, is not clear in reference to the fundamental physical processes involved; its significance for experimental work is evident. At this pH the observed EO is independent of the nature of the buffer components and of the ionic strength. Analytical measurement should be made ut this pH value unless other facturs--e.g., need to separate EO.& v a l u e s 4 i c t a t e a different pH region. Behavior in Nonbuffered Solution. Solutions of a-bromo-nbutyric acid in distilled water containing potamium chloride but no buffer system were electrolyzed in order to observe the behavior in unbuffered solutions (W in Table I and Figure l ) . The pH of these solutions was 3.5. The change in EO6 with ionic strength was found to be slight and the EO6 of a-bromo-n-butyric acid a t any ionic strength mas less negative than in buffered solutions of equal pH, indicating the lack of interaction between electroactive species and supporting electrol-vte component. CONCLUSIO3
There ale regions (Figure 1 ) for each buffer over nhich thc change in Eo5 is relatively slight for a given changr in ionic strength. It is obvious that the nieitsurement of En i n these ranges of ionic strength will yield values uhich are inwe valid when used to compare the results of various investigators than results obtained at other levels of ionic strength. For the compound used in this investigation, a-bromo-n-butyric acid, the particular region of the ionic strength at which to compare Eo a t any pH is between 1.0 and 1.5, as this region would encompass sections of each curve where the change in EO5 with ionic strength is relatively slight. It is recommended that in future investigations of the polarographic behavior of organic compowds, the test solutions examined be made up to definite ionic strength, the value of R-hich should be stated in the same manner as the routine statement of the capillary constants Such B procedure mwuld increase the facility of duplication and comparison of results. .Slthough the present paper does not deal with any specific analj-tical methods, the more precise evaluation of several of the factors affecting the results obtained in observing organic compounds polarographically has Pmphasized the need of controlling such factors when developing an analytical procedure. The need is obvious of operating in a region of ionic strength and buffer concentration where the wave is relatively unaffected by such factors. Where large samples are taken for the determination of minor components, the effect of the contribution of the sample to the ionic strength must be considered and a method of calibration used which is a t the same level of ionic strength as that expected in the actual analytical sample solution. ~
V O L U M E 23, NO, 9, S E P T E M B E R 1 9 5 1
The author& wish to express their gratitude to the Atomic Eiiergy Commission a n d to the Research Corp. for grants-in-aitl which supprted t,hi. research. LITERATURE C I T E D (1j Albright, C., mLtciTer's thesis, The Pennsylvania State ('olic~uc~.
1950. (2) DeFord, D. D., nritl Aiidersen, D. L., J . Am. Chew. Sac., 72, ,3918 (!950). (3) Elmng, P. J.. T~orc.ritha1.I., aiid G a m e r . M. K., Ibid.,73, 1717 (1951).
(4) l ~ ~ l v i i i 1' g., J . , and Tang, C. S., Ibid., 72,3244 (1950). ( 5 ) Furxnall, s.H., and Stone, K. G., Ibid., 70, 3055 ( 1 M ) . (6) Hodgmaii, c'. I)., ed., "Handbook of Chemistry m d Physics," 26th rd.. ('leveland, Ohio, Chemical Rubber Pllblishing Co., 1942. ( 7 ) l i o i i i y ~ t h y ,J . C ' . , Jfalloy, I:., and Elving, P. J., Ax.4~.CHEM., accepted for publication. (8)Stewart, 1'. E., aiid Boiiner, W. A , Ihid.,22,793 (1950). (9) Stone, K. G., J . F'lcctrochen?.,Sot., 97,63 (1950). 1)wember 1, 1950. No. 10 in a series on the polarographic behavior of organic roiripound.?. Previous papers have appeared in .IXALYTICAI, C H m i I s l i ~ Y .J. o i i r . i i n 1 a f t h e American Chemicnl S o c i e t y , etc. RECEIVI.IJ
Derivative Polarographic Titrations CH4HLES N. REILLEY, W. DONALD COOKEL,AND N. HOWELL F U R M l K Princeton University, Princeton, IT. J. ' h i s method of differential polarographic titrations
r e u l t e d from a s t u d j of t h e fundamentals of endpoint indication for coulometric titrations. From theoretical considerations it was predicted that if a pair of platinum electrodes are polarized by a small constant current, ca. 2 microamperes, the electrode w i l l give runtinirous e.m.f. readings corresponding to t h e slope of a polarographic curve a t its zerocurrent axic. If the systems are both reversible, differentiaY e.ni.f. peaks will be obtained a t the end point. and a aiiccession of differential end
STEFtEST in dvnvativc titrations has recently been revived
( 3 , 4 , I d , 1 6 ) K a t t and Otto (16) used a concvntmtioii cell cffcxct for potentiornetric titrations involving solutions of metals in liquid ammonia Delahny (3,.$)described an electrical method for differentiating the voltage-volume relationship, u-lug the diffcrentiation prop.rtit.6 of a condrnser ( I ) . This lattrr method has the dieadvantap of requiring a constant flow rat(, o f titrttting ngcnt and therefore of having no sustaincd mvter rc3:~dingfor :L 1
I'rcent address Coruell Univerjity, Ithacs, K. Y.
0 c Y
c e o Y
givcsn v o l u ~ n cof ~ th(7 titr:iting agent. The origirid differential method :I> tlt.viwd ~ O (X2 ) employed a dual titmtion where one titration na? kept .;lightly in advance of the other. -i more practic:tl method dewloped by MacInnes and Jonea (10) utilized the concentration vo1t:igt~obtained by removal of a mll portion of thv titrated solution 1x.forrS w c h addition of reagent. Miiller ( 1 1 ) employcd a simihr cchtmie. These last two whernes require a change ot rclfewnce solution before each addition of reagrnt. Some dihsinlilar bimetallic electrode eyPtems give results t h a t simuhte these differential potentiomctric titr:ttions ( I , $ , 17). Delahay ( - 5 ) has n w n t I \ discushetl the Foulk and Bawden 3'dc~;d-stop''inc.chanisni (6) by coneideration of tht. Phaptv O f various polarization curves. ti polarized bimetallic system using two similar electrodes n as suggested by Willard and Fenwick (18). Van Same and Fenwick (16) discussed VOLTbDL the behavior of two platinum electrodes polarized by R 0.5-volt sourcc. The method de-rrihctl in this paper is a polarized system similar t o thnt (Jf Fenwick e l al., but cam is taken to polarize the electrodes with constant current. Tlie electrochemical phenomenon giving rise to the curves i c esplained in tern- of polarographic behavfor. I t is hoped that the esplanation will rliminatr the, hcretoforc empirical nature of the method and assist the analyst in Rpplying this technique to new situations
points w i l l be r r a l i m c i in t h e titration of a succession of substance&. If one of the systems is irreversible, the electrodes w i l l s h o w a sharp increase or decrease in e.1n.f. at the end point. The conclusions we= verified experinlentally. This method provides simplicitj of apparatus, con tinirous indication, sharper breakfi mer the conventional potentiometric metho d s in mine cahes, rapid attainment of stable readings, eliniinatioo of a reference half-cell, and ability to detect a siicce4on of end points simply and witho u t plotting.
Figure 1. Typical Polarograms of a Solution of Ferrous Ion Being Titrated by Ceric Ion A. E.
Initial solution of ferrous ion Solution half titrated
Solution at end point Solution after addition of excess d
Typical polarograms taken a t a platinum electrode in a stirred solution of a polarographically reversible system (ferric-ferrous) titrated by a reversible system (ceric-cerous) are shown in Fig-