DAEHAN HWAHAK HWOEJEE (Journal of the Korean Chemical Society) Vol. 18, No. 5, 1974 Printed in Republic of Korea

Triphenylphosphine Phenylimide 의 전기화학적인 환원

朴鍾民* , Wilson M. Gulick, Jr.

미국 후로리다주립대학교 화학과 (1974. 7. 13 접수 )

Electrochemical Reduction of Triphenylphosphine Phenylimide

Chong Min *Pak and Wilson M. Gulick, Jr.

Department of Chemistry^ Floridia State University^ Tallahassee,

Florida 32306, U, S. A.

요 약 . Triphenylphosphine phenylimid合의 비수용액에서의 전기화학적인 환원반응을 polaro- graphy, cyclic voltammetry, controlled-potential coulometry 및 electron spin resonance 법 을 써 서 고 찰하였다 . 이 유기 인화합물은 cme-electon transfer 에 따라서 anion radical 형성되 나 순간일 뿐이 고 protonation 과 재차 one-electon reduction 결과 인과 질소사이의 이중결합이 끊어진다 . 그 결과 아 닐린이 주요 반응생성물로서 발견되었다 . 또 한편 동반하는 화학반응결과 생긴 주산물의 하나인 triphenylposphine oxide 의 환원결과 인과 페닐사이의 단일결합이 끊어지는 것도 관찰할 수 있었다 .

Abstract. The electrochemical reduction of triphenylphosphine penylimide in nonaqueous media has been examined by polarography, cyclic voltammetry, controlled-potential coulometry and electron spin resonance spectroscopy. The reduction of tiiphe교 ylphosphine phenylimide proceeds by a one- electron transfer to form anion radical which undergoes both protonation and a second one-electron reduction followed by cleavage of the phosphorus-nitrogen double bond. Aniline is a major product. The cleavage of a phosphours-phenyl bond was also observed after reduction of triphenylphosphine oxide which is one of the major products of the chemical reaction which follow the primary process.

synthesis and ligands in coordination chemistry. 1. Introduction The range of application of phosphorus com­ Phosphorus compounds have become increas­ pounds in modern technology is extremely broad ingly important as intermediates in organic and varied. Since phosphorus is sometimes *Present Address: Department of Chemistry, Soong found in minute quantities often in the form of Jun University, Taejon 300, Korea. Author to whom labile or nonvolatile compounds, the detection, correspondence should be addressed. Presented at the 33rd National Korean Chemical Society Meeting, assay, and identification of its compounds re­ Seoul, Korea (Apr, 26~27, 1964). quire considerable skill, forcing the chemist to

—341 — 342 朴鍾民 , Wilson M. Gulik, Jr. push his analytical tools to their limits of per­ 2. Experimental formance. Free radical intermediates have frequently 2.1. Chemicals been postulated for reactions involving or음 ano- Acetonitrile and tetraethylammonium perchlo­ phosphorus compounds. However, compara­ rate (TEAP) supporting electrolytes were pre­ tively few phosphorus-containing radicals have pared by standard literature methods.12 PPPI been prepared in solution under conditions which was prepared by reaction of phenyl azide permitted their study in detail by esr. As early (in hydrocarbon solution obtained from Pfaltz as 1953 Heine and coworkers1 reported that the and Bauer, Inc., Flushing, N. Y.) and tri­ reaction of phosphine oxides with alkali metals phenylphosphine in ether solution. This re­ led to colored, paramagnetic solution; however, action is known as Staudinger reaction.13 a review2 of the literature of phosphorus esr in After three recrystallizations from ?z-hexane, 1966 counted only 16 publications. Since that the resulting crystals melted at 131 〜131.5° time, a number of additional studies have ap­ which is in agreement with the melting point peared. Cowley3 has reported several radicals reported by Wiegraebe and Bock.14 Other che­ containing phosphorus, Lucken4 has observed micals were available from standard commercial phosphorus redicals in irradiated solids, and sources. several radicals have been detected as an adjunct 2.2. Electrochemistry to the electrochemical studies by Bard and co- Polarographic measurements were made with workers5"'7, Recently Matcalfe and Waters8 have the solid state polarograph, employing oper­ reported the formation of radicals from phospho­ ational amplifier circuiting with a three electrode rus and phosphoric acid esters in a fast flow configuration, described previously.15 For rapid­ system and Dimroth9 has prepared a novel het­ sweep cyclic voltammetry, a Tektronix type erocyclic phorphorus radical. The first indication 201—W oscilloscope equipped with type D of hindered intramolecular motion in a phospho­ plug-in aplifiers was employed. Pictures were rus radical has been detected recently by recorded with an oscilloscope camera and a Rieker10 via line-width variations. type 2620 Polaroid attachment. At slow-sweep From the industrial point of view, triphe- rates. Moseley model 2D—2M X—Y recorder nylphosphine phenylimide (PPPI) was chosen as was used. A Hewelett—Packard model 3300 A phosphorus—nitrogen compound; considerable function generator with a model 3302 A trigger interest has arisen recently in polymers which served as the signal generator. Controlled- have the general formula (N™ P (OR) 2) 11 Some potential electrolysis were carried out using a of the desirable properties of these materials Princeton Applied Research (PAR) model 173 have been attributed specifically to the presence potentiostat equipped with PAR model 177 of the phosphorus—double bond(一 current readout module. Coulometry was carried N=P—) which is, it is pointed out,11 isoelec- out using a nitrogen— coulometer as tronic with the —Si—O— skeletal unit of the described by Lingane.16 Additional cyclic vol­ silicon series, and the two units behave simi­ tammetry experiments were carried out usin흠 the larly in many instances. We have now examined signal generator and oscilloscope described above the mechanism of the reduction of PPPI, which in conjunction with the PAR 173. In these is subject of this report. experiments, a PAR model 176 current to

Journal of the Korean Chemical Society Triphenylphosphine Penylimide 의 전기화학적인 환원 343 voltage converter was used and positive feedback voltammetric experiment, fr毕hly prepared so­ IR compensation was employed. lutions were used, identical with those employed The electrolysis cell was a three compartment in polarography. The same cell as used for vessel for polarography and cyclic voltammetry. polarography was employed except for the use Anaqueous saturated calomel electrode (see) was of a mercury-coated platinum wire or platinum used as the reference electrode which was iso­ disk working electrode. After each solution was lated from the working electrode compartment purged with nitrogen the potential range was by means of a fritted disk, the auxiliary elec­ swept back and forth by means of the triangle trode was a platinum wire and a dropping mer­ wave generator connected to the summing point cury electrode was the working electrode. For of the control amplifier in the polarograph. cyclic voltammetry, a mercury-coated platinum17 The frequency range of 0. 01 〜10 cps was usually or a platinum-disk electrode replaced the drop­ investigated, with the voltammograms being ping mercury electrode. The mercury-coated recorded either on X—Y recorder (>1 cps) or platinum wire electrode was made by sealing by photography from the oscilloscope for the platinum wire into a soft-glass tube, as described higher sweep rates. The precision of potentials previously.18 The auxiliary electrode was a measured from chart recording is estimated at platinum wire and the reference electrode was the ±5 mV level. Larger uncertainties, esti­ an aqueous see. The working electrode in the mated at 5 % were encountered in oscilloscopic controlled-potential electrolysis was a mercury measurements. pool of approximately 7 cm2 area which was Controlled-potential Electrolysis. When stirred continously during the controlled-potential coulometric measurments were made, cathodi- electrolysis. The reference (see) and auxiliary cally reactive impurities were destroyed by (mercury pool) electrodes were separated from pre-electrolysis; a measured volume (usually 25 the working electrode compartment by means m/) of background solution (0.1 M TEAP一 of, respectively, a glass fritted disk and a MeCN) was introduced into the working bridge containing solvent and supporting elec­ electrode compartment of a three compartment trolyte between two fritted disks. cell. This solution was electrolyzed at the po- Polarography and Cyclic Voltammetry. In tental which was employed for the controlled- a typical polarographic experiment, a freshly potential electrolysis of the compound until a prepared solution of TEAP (0. 1 M) and start­ steady-state current was reached. After pre­ ing material (ca. 1 〜4 mA】) in acetonitrile electrolysis, a weighed amount of the sample (25 mZ) was transferred to the electrolysis sufficient to make the solution ca. 1 〜4 mAf was cell and the three electrodes were intro­ introduced into the working electrod compart­ duced. The solution was then purged for ment. After dissolution, electrolysis was recom­ 30 minutes with prepurified nitrogen which menced at the same potential, stirring rate was passed first over hot copper wool and and flow rate of nitrogen. The temperature was then through purified acetonitrile. The pre­ not controlled. However, cell heating was not saturation step minimized loss of solvent from observed to any great extent. Completion of the cell. An atmosphere of nitrogen was then electrolysis was confirmed by obtaining a steady­ maintained over the solution as the vol- state current. tammograms were recorded. In a typical cyclic 2.3. Product Analysis

Vol. 18, No. 5, 1974 344 朴鍾比 , Wilron M. Gulik, Jr.

Analysis of electrolysis products were carried solvent was evaporated from each eluant under out using gas chromatography, voltammetry and reduced pressure and the resulting material was thin-layer chromatography (TLC) followed by examined by spectroscopic methods. spectroscopic methods. The gas chromatograph Aniline. After the electrolysis, the catholyte was a Pe나 cin—Elmer model 990 equipped with was analyzed without workup by gas chroma- thermal conductivity detector. The infrared t。흥 raphy using 4. 0 ft column containing 5. 0 % spectra were taken on either model 21 or model diethylene glycol succinate on Chromosorb W 257 Perkin-Elmer $고 ectTometers, using standard at 125° with a helium flow rate of 50 mZ/min. techniques; the ultraviolet spectra were recorded Aniline was eluted with a intention time of 5. 0 on a Cary model 14 using quartz cells; the minutes. Standard solutions containing 100 NMR spectra were determined either on a mg of aniline in 25.00 mZ of acetonitonitrile Varian model A 60 or 90 MHz Bruker model was prepared and aliquots of these standard H FX一 10 spectrometer； the low resolution mass solutions were employed to obtain analytical spectra were taken on a Nuclide instrument; the curves. The respective peak height was plotted high resolution mass spectra were taken on an against the vzeigbt of aniline in the standard Associated Electronics Industries MS 902 instru­ solution. In order to investigate further evidence ment equipped with a computer for data re­ for aniline the following experiments were carr­ duction and elemental analysis printout. ied out. The GC effluent was trapped from the Small analytical TLC plates were prepared by exit port of the gas chromatograph in a melting dipping microscope slides in a slurry of silica point tube cooled with liquid nitrogen. The gel containing an inorganic fluorescent indicator trapped effluent was examined by UV, IR, and and gypsum (Merck GF一 254) in 2 • 1 chio- NMR spectroscopic experiments. roform and methanol. Preparative plates were Triphenylphosphine (TPP) and Triphenyl­ prepared by coating 20 x 30 cm glass plates with phosphine Oxide (TPPO). After the eletrolysis aqueous slurry of silica gel GF—254 (60 g silica acetonitrile was removed from the catholyte by gel—120 mZ water-shake for about 30 second s- evaporation under reduced pressure and the enough for two plates). The wet plates were remaining residue was extracted with benzene. air-dried for overnight and were then activated The benzene extract was analyzed by thing­ in oven at 105° for 2 hours. The base line was layer chromatography. Development of the fixed at a distance of 20 mm from the edge of plates with a solvent mixture (benzene一 acetone, the plates, and a benzene extract of the elec­ 75/25) was carried out by the ascending method trolysis products was applied to the plates. The and led to the formation of four separate zones. chromatograms were developed in one-dimension After the plates were removed from the clevcL (by the ascending method) in a glass tank oping chamber, the two zones of dry gel (Rf containing solvent mixture. After the plates values of 0. 40 and 0. 82) were scraped, respec­ were removed from the developing chamber tively, from the plates and then eluted with they were air-dried and the dry gel of each chloroform. The solvent was evaporated from zone was scraped from the plates and was placed each eluant and the resulting residue was exam­ on a column, the bottom of which had been ined by UV, IR and NMR spectroscopic me­ plugged with a sintered glass frit. The gel of thods. Quantitative estimations of these com­ each zone was then eluted with solvent. The pounds were performed as follows. A know n

Journal of the Korean Chemical Society Triphenylphosphine Penylimide 의 전기화학적인 환원 345

■amount (100〜200 以 )of the benzene solutions obtained by UV absorption spectrophotomety. was placed on the preparative plates. After In the analysis of the catholyte for aniline development of the plates with the solvent by gas chromatography, the appearance of an mixture described above the two zones of dry additional peak with retention time of 7.2 min­ gel were scraped, 고 espectively, from the plates utes was discovered. In order to investigate the and then leached from the silica gel with chlo- nature of this peak, the GC effluent was trapped 고 oform. The chloroform was evaporated from and was found to be a very hygroscopic, white each eluent and the remaining residue was ex­ crystalline solid, melting point 75°. The isolated tracted with acetonitrile. The acetonitile solution white solid was examined by UV, IR, NMR and was analyzed by UV absorption spectropho­ mass spectroscopic methods in an effort to elu­ tometry. cidate the structure. Diphe효 yphosphinic Acid. The product solu­ 2.4. Water Analyses tion was evaporated under reduced pressure and Water concentration in acetonitrile was ana­ was then extracted with benzene. The resulting lyzed by gas chromatography on a 3. 0 ft Po­ residue was treated with hot distilled water. rapak Q column at 200° with a helium flow Cooling this solution caused precipitation of the rate of 20 mZ/min and a retention time of 0. 6 supporting electrolyte which was filtered out. minutes. Calibration curves were obtained from The aqueous solution was extracted with ether. mixture of acetonitrile and water; the water The aqueous portion from which ether solution content of the solvent was determined by the was separated was acidified with 2. 0 M hydro­ method of standard additions. chloric acid and extracted with benzene. The 2.5. ESR Spectroscopy benzene solution, after reduced pressure evapo­ ESR spectra were recorded using a Varian E— ration, gave a white solid which melted at 189 12 spectrometer system and the intra muros19 〜190° which is in agreement with the melting electrolytic technique of free radical generation. point of an authentic sample of diphenylphos- phinic acid. The resulting white solid was then 3. Results and Discussion examined by TLC, UV, IR, NMR and mass 3.1. Polarography and Cyclic Voltammetry spectroscopic methods. The white solid was The polarographic reduction of PPPI in aceto­ identified as diphenylphosphinic acid by the nitrile solution containiri 응 0.1M TEAP showed results of TLC and spectroscopic examinations. two waves; one well-defined and the second ill- Quantitative analysis for diphenylphosphinic defined, with half-wave potentials of — 2.50 acid were carried out by dissolving the white and —2.66 Nvs. see. The variation of the limi­ solid in methanol and measurenent of the UV ting current with the mercury head applied to spectrum. the dropping mercury electrode was studied for Benzene. Bensene analyses were performed by the first wave. As shown in Table 1 the wave gas chromatography of the product solution drawn height varied approximately with the square ， ： 〃 from the catholyte compartment. A3 x - root of the mercury head. This behavior sug­ column packed with Porapak Q at 195° with a gests that the wave is result of diffusion-con­ helium flow rate of 20 mZ/min indicated the trolled process. The value of Eg—E* for the presence of benzene with a retention time of 10 first wave, —17 mV, suggests an irreversible minutes. Additional evidence for benzene was reduction.

VoL 18, No. 5, 1974 346 朴鍾民 • Wilson M. Gulik, Jr.

Triangular sweep cyclic voltammetry of 1. 0 ip/nFA(Da)1/2,*C where ip is the peak cur­ mM PPPI in 0.1 M TEAP—MeCN solution rent, v is the sweep rate *C is the initial con­ at a mercury-coated platinum wire cathode centration of electroactive substance, and the yielded one reduction peak with peak potential, parameter a is defined as(i=hFp]RT；八 Ep= —2. 65 Nvs. see. No anodic peak correspond­ 3.2. Controlled-Potential Coulometry ing to the oxidation of the product of the In view of the isolation of products similar cathodic process was observed on the reversal to those produced by hydrolysis of PPPI we veri­ of the scan at any sweep rate or any switching fied that hydrolysis does not occur under our potential, suggesting a totally irreversible pro­ experimental conditions. An UV spectrophoto- cess. From the experimental point of view, a metric examination of a solution of PPPI in cathodic ece (election transter-chemical reaction­ acetonitrile showed absorption spectra with electron tranfer) system can be classified accord­ maxima at 252 and 227 nir, which did not ing to the number of cathodic and anodic 아 ｝ ow in the UV absorption spectra of TPPO in waves which are oberved. Further classification according to the reversibility or irreversibility Table 1. Polarographic data" for the reduction of of the two charge transfers can be made by TPPP. examination of the anodic waves as scan rate is varied, since anodic waves are never observed Mercury head A (Corr), cm (A/cm1-2) for irreversible charge transfers20. From exami­ nation of cyclic voltammograms as a function 91.3 0. 37 79.8 0. 35 of sweep rate from 0. 06 to 0. 60 volts per sec 70.1 0.28 using X—Y recording, the current function 61.9 0. 27 (proportional to ip!is found to decrease as 。Data for the first wave. the sweep rate is increased. This fact may indi­ b The solution was 0. 1 M TEAP in CH；CN, 가 nd： cate a chemical reaction coupled with the contained 0. 69 mM PPPI. electrochemical reaction at these sweep rates (Table 2). The current function is defined as

Table 2. Cyclic voltammetric data11 for the reduction of TPPI in acetonittrile^.

Sweep rate Without IR Compensation With IR Compensation

(volts/sec) —E T 源 ―%。 財

0. 06 2. 68 75.0 307.4 2. 68 80.0 327.9 0. 12 2. 69 95.0 275.4 2.69 100.0 289.9 0. 18 2. 71 100.0 237.0 2. 71 105.5 250. 0 0. 24 2. 72 110.0 225.4 2. 72 115.0 235. 7 0. 30 2. 72 115.0 211.0 2. 72 120.0 219.8 0. 60 2. 73 120.0 155.4 2. 73 130-0 168.4 Data for the first sweep. The wsking electrode was a mercury-coated platinum wire electrode, auxiliary electrode was a platinum wire and aqueous see was used as a reference electrode; b The solution contain­ ed 1. 40 mM PPPI and 0. 1 M TEAP; c Volts vs. see; d /(A; e se서 /'' Volts^i/2

Journal of the Korean Chemical Society Triphenylphosphine Penylimide 의 전기화학적인 환원 347 acetonitrile. After letting the solution sit over­ mid키 e part of the electrolysis the current decr­ night we have found no change in the absorp­ eased exponentially eventually reaching a steady tion maxima. state current. These facts suggest that a chemi­ Controlled-potential coulometry of PPPI at a cal reaction which follows the charge transfer mercury pool cathode on the diffusion current constantly regenerates an electroactive species. plateau of the first wave at a potential of The original colorless solution changed to yel- —2.60 V vs. see was carried out in order to low-organge at the beginning of the electrolysis; find the number of electrons transferred during then the color turned to yellow during the reduction of the compound and the nature of middle part of the electrolysis. Finally, the the reduction products. Because of the closeness solution became colorless again at the end of of the background discharge potential (—2. 85 V. the electrolysis. All log i vs. time plots obtained vs see), conditions for controlled-potential coulo­ as various concentrations are distinctly non­ metry are not the most favorable. The most linear. The curves show concave deviation from unfavorable factor is, however, the small sep­ linear behavior and their slopes are essentially aration of the two reduction waves; E/ of the concentration indepedent. wave 2(—2. 66 V) is separated by only 60 mV The electrochemical behavior of the reduced from the cathode potential. As shown in Table solution was investigated by means of polaro- 3 the final steady state current in coulometry graphy and cyclic voltammetry. The first cathodic was approximately 2〜8 % of the initial value wave has disappeared and the product of the and also the steady state current is much higher reduction gives rise to a new reduction wave than background value (ca. 0.05 mA). This with an Ei/2 of 一 2. 57 V. The anodic wave appears to be due to products which are them­ height of approximately 20 % of the cathodic selves electroactive at the working electrode wave corresponds to oxidation of a product of potential. In every case, the number of electrons chemical reactions. It was suspected that this involved in the reduction was found to be new reduction wave was due to the presence of approximately four at this potential. triphenylphosphine oxide. The following proce­ The variation of electrolysis current with time dure was used to confirm the presence of TPPO was not exponential. It was found that near the following the reduction of PPPI. Approximately beginning o£ the electrolysis the current showed 1.0 mAf solution of TPPO was prepared by nearly a linear decay with time. During the addition of authentic TPPO to the reduced

Table 3. Coulometric data for the reduction of TPPI in acetonirile.

Concn. 「％、 潛 PPPI (mA) (mA) 4^xiool (mM)

1.00 12. 0 1. 05 8. 75 4- 07 1. 71 22.5 1.75 7.78 4.04 1. 79 27. 0 1. 20 4. 45 4. 01 2. 40 43.0 1.05 2.44 3.74 3. 35 57. 0 1. 50 2. 63 4. 03 Faraday per mole of PPPI.

Vol. 18, No. 5, 1974 348 朴錘民 • Wilson M. Gidik, Jr. solution and polarograms were taken under the examined by UV absorption spectrophotometry. same conditions. The reduction wave exhibited The UV spectrum consists cf three sharp intense by the added TPPO occurs at the same potential peaks at 272, 265 and 222 nm. Comparison of as the new reduction wave with half-wave these values with the spectral values of authentic potential of —2. 57 V and causes an increase in TPPO (272, 265, 259 and 223 nm) indicated the wave height. This investigation suggested that the reduced s시 ution contained TPPO as a the formation of TPPO in the product solution. major species. Further evidence that one of the When the reduction was examined by cyclic major products of the controlled-potential elec- voltammetry the voltammogram showed three tr사 ywis of PPPI was TPPO was obtained by reduction peak potentials at —2.13, —2. 71 and TLC followed by spectroscopic examinations. —2. 76 vs. see. To determine the identity of Through the spectroscopic examination of two the new systems, a comparison of the peak major spots and comparison of authentic potentials with those of various known com­ ples, it was ascertained that the spot with Rf pounds was made. The reduction potential at value of 0.40 was TPPO and the other spot —2. 71 V agreed well with the TPPO system. (•R/=0. 82) was TPP. The voltammetric data of TPPO showed that Because the latter spot appeared to be too the oxide is reducible at the potential employed large compared to the yield of TPP and TPPO for controlled-potential coulomentry. The re­ obtained from quantitative analyses and to gain duction of TPPO probably proceeds by a one- fuHher insight into the spot, two-dimensional electron reduction to the anion radical.5 Since TLC technique was used. After the chromato­ TPPO was found to be approximately 25 % grams were developed in one dimension with yield from the electrolyzed solution (Table 4), benzene一 acetone the plates were removed from reaction of the anion radical with solvent to the developing chamber and were air-dried for regenerate TPPO probably occurs. The nature several hours. The plates were then rotated ■of the other peaks were not investigated further, through 90 degrees, immersed in the second but they are assumed to be caused by minor solvent (dichloromethane), and developed. Thus products of undefined side reactions. via two-dimensional TLC experiments, we found After eletrolysis, the reduced solution was that the high Rf value spot (R尸 =0.82) was

Table 4. Controlled-potential electrolysis results of triphenylphosphine phenylimide

.Concn* Yield, moles product/moles starting material, % PPPI (mM) Diphenylphosphinic TPP TPPO Acid Aniline

1.67 10. 7=1- 26. "2.俨 34.4 >.0. 5’ 70. 9_~0. 9b 2. 26 14. 7-0. 6 23. 7±1. 9 32. 3 二 2.0 74. 7二 5. 3 3. 50 15. 8±0- 7 18. 6工 L 4 26. I . zl.4 77. 3二 L 1 4. 76 12. 5±0. 5 18. l rzl. 0 단 2. 3 r：5. 2 82.6-5.8 a The acetonitrile solution contained 0-1 M TEAP and 10. 9 mA/ water. Each concentration of the solution was run two or three times. b Uncertainties are standard deviations from at least four measurements. These reflect only the precision of the final analytical step, not the precision of the over-all work-up.

Journal of the Korean Chemical Soci< iy Tripheny[phosphine Penylimide 의 전기화학적인 환원 349 resolved to two spots (Rf value of 0.03 and formed by loss of a proton from the molecular 0.77). The combined TLC and spectroscopic ion m/e=218). Subsequent loss of water results examinations of these two spots showed that in the fragment Ci2H8PO+ (m/e= 199). These 하 ley are TPPO(Ry=0. 03) and TPP(K須 =0. 77). combined spectroscopic results provide persuasive Examinations of the authentic TPP by two- evidence for the formation of diphenylphosphinc dimensional TLC experiments suggested TPP acid in the electrolyzed solution. is not oxidized on the plate. Presuma바 e, The identification and quantitative analysis TPP and TPPO form a complex in the first of aniline were carried out by gas chromato­ solvent mixture. graphy. The gas chromatographic peak from the The reduction of PPPI was carried out with catholyte was trapped and the efRuent was dis­ consumption of le, 3e and 4e in order to solved in cyclohexane and was analyzed by UV ascertain the number of faradays required to absorption spectrophotometry. This UV spectrum consume all the starting material. The existance contained absorption maxima at 287 and 233 nm. of the starting material in the incompletely They are the same as the absorption maxima reduced solutions, up to and including 3e re­ for aniline.23 Confirmation of aniline was ob­ duction, was found by using TLC. These ex­ tained via another UV spectrum; the effluent was periments also showed that TPP and TPPO were dissolved in 2.0 A/ hydrochloric acid and the produced in every case. UV spectrum of the acidic solution showed In order to investigate additional products, absorption maxima at 259, 253, 248 and 243 the residue remaining after reduced pressure nm. These spectral values are w시 1 matched evaporation of the catholyte was examined. with those for authentic aniline hydrochloride. The white solid obtained by solvent (benzene) The formation of benzene is established by extractions of the residue was examined by gas chromatography and absorption spectropho- TLC and it was found that the compound was tometric analysis of the electrolyzed solution; resolved at Ay value of 0.15. A methanolic phenyl radicals formed by cleavage of preceding potassium hydroxide solution of the white solid electrolysis products may abstract a hydrogen was analyzed by UV absorption spectrophoto­ atom from the medium to form benzene. Pre­ metry; this solution exhibited four intense absorp­ vious workers5 have reported that the phenyl tion peaks at 272, 265, 259 and 223 nm. In radicals dimerize to form biphenyl when TPP the absence of potassium hydroxide, the weak is reduced in DMF. Results of analysis of elec­ peak at 253 nm was not visible. Comparison of trolyzed solution by 응 as chromatography show these values with the spectral values of phos­ that biphenyl is not present in our system at a phorus compounds reported by Jaffe, et al.22 detectable concentration. The limit of detection indicated that the methanol solution contained for biphenyl was found to be 2. 0 x 10~3// mole. diphenylphosphinic acid. IR and NMR exami­ It was not possible to characterize completely nations of the compound further confirmed the the hygroscopic white crystalline solid which presence of diphenylphosphinic acid. The spectra we isolated during the GC analysis for aniline. agreed well with the spectra of authentic ma­ The following results, however, confirm that terial. The mass spectrum of this compound it is a derivative of (C6H5) 2?02-. The physical showed a base peak at m/e=217 which corres­ properties of the unknown compound suggested ponds to the primary frament (C6H5) 2POO+ a salt as a first guess, such as possibly an

Vol. 18, No. 5, 1974 350 朴鍾民 * Wilson M. Gulik, Jr. ammonium perchlorate, but mass spectrum (MS) der of the molecule we obtain a 20 eV MS; showed no substantiation for the presence of per­ however, the greatest m/e of any prominence chlorate. The mass spectrum of this compound were still 217〜219. Since a molecular ion does showed very pronounced peaks at m/e=217, not appear, we compared the MS of the unkno­ 218 and 219 with relative intensities very similar wn with that of authentic (C6H5) 2POOH and to those for authentic (C6H5)2POOH. In addi­ then pi시 ced out some of the prominent peaks tion, high resolution mass spectroscopic data which are not present in MS of authentic sam­ gave excellent fits on these peaks for (C6H5)2- ple. It was hoped that these fragments would zO zO /OH give a clue to the structure of rest of the mole­ P+( (CfiH5)2P+^ and (C6H5)2P+(f cule (see Table 5). Although here is not e- 0, XOH 、OH nough information for a definite structure deter­ (see Table 5). The UV spectrum was also very mination, there is some indication from the com­ similar to (CfiH5)2POOH and the NMR spectrum position of these fragment ions that the rest of showed the presence of pheyl group; however, the molecule does contain nitrogen. In addition both the UV and NMR spectra were distinctly the IR spectrum showed a broad absorption in different from those for (C6H5)2POOH. the region of 2. 93〜3. 02 Thus, we conclude that the unknown com­ Assuming aniline and the unknown compound pound is a derivative of (C6H5)2PO2-; it could have approximately equal GC response on an be a salt of the anion or an ester. The material area and mole basis, the incompletely character- was dirs^lved in 1.0 M potassium hydroxide ized material appears to be present in sufficient 산 solution and was len extracted with ether. If quantity to account for the phosphorus not the material is a salt, presumably the anion h)und in other products. Assuming that the (C6H5)2PO2- would stay in aqueous phase and UV spectrum of the unknown compound is due free base would be extracted into ether. Both entirely to (C6HQ2PO2一 and the molar absorp­ phases, however, showed idential UV spectra, tivity of the absorption peak at 264 nm is the a result which suggests the ester postulate may same both the unknown compound and (C6H5) 2~ be correct. In order to characterize the remain­ POOH, the incompletely characterized material

Table 5. Summary of physical and spectral data for the incompletely characterized compound.

1. physical Properties Very hygroscopic, white crystalline solid. Soluble in water, acetonitrile and aceton. Insoluble in benzene, chloroform and carbon tetrachloride. Melting point 75° 2. Spot test Positive for phosphorus by ammonium molybdate— 3. TLC R/=Q. 03 on the silica gel GF—254, solvent mixture is benzene/acetone (75/25). 4. GC 5. 0 % diethylene glycole column at 125° with a retention time of 7. 2 minutes. 5. NMR (acetone— 1.89 (s') Integration gives a 13 : 10 ratio for singlet ($) and multi- (w. TMS) 6. 41 〜6. 80 (m) plet (m) peaks. 6. UV (MeCN) (nm) 238, 289 7. IR (KBr) 2. 93〜3. 02 S) 6.10 〜6.14 7. 25 8. 92

Journal of the Korean Chemical Society Triphenylphosphine Penylimide 의 전기 화학적 인 환원 351 accounts for 22. 5 % of the total phosphorus on It is also evident that the anion radical is con­ the basis of a simple calculation of 4. 76 mM sumed by some fairly fast reaction since no PPPI. The malar absorptivity used was 1170 reverse current was obtained in cyclic voltam­ which was calculated from a standard solution metry. The experimental results are best explain­ containing authentic (C6H5)2 POOH in acetoni­ ed by assuming that the anion radical formed trile. The absorbance of the unknown at 264 can immediately add a second electron at the nm in ether solution was 0. 78 and in aqueous applied potential to form PPPI dianion. Further, solution was 0. 74. The total volume of both the dianion may react with a protonating solutions was approximately 40 mZ. A summary agent (HS) present in the medium, such as traces ■of the physical and spectral data for the com­ of water, the solvent itself, etc. pound is given in Table 5. ((C6H5)3P-NC6H5r+ e The results described here show that the over­ —((C6H5)3P-NC6H5]= (2) all reduction indicates an n appearent of 4 dur­ ing coulometric reduction, and that the primary ((C6H5)3P^NC6H5r+ 2HS ■electrode process is followed by chemical reac­ —(C6H5) 3P+H2NC6H5 + 2S- (3) tions which produce products electroactive at the Another possible reduction pathway could be working cathode potential. In addition to this, postulated by assuming that the anion radical the qualitative picture obtained from cyclic may undergo chemical reaction with a proton­ voltammetry showed total lack of anodic current ating agent. The following reduction mechanism corresponding to the reoxidation of the reduced can be suggested: products and the current function decreases as 〔 (GM5) 3P=NC*5 〕十 + HS—스 the scan rate increased (Case VI in the work ((CgHs) 3P NCgHg] H • + S~ (4) of Nicholson and Shain21), all of which are consistent with an ece-type mechanism20. The ((C6H5)3P^NC6H5]H- 十£= quantitative results for controlled-potential elec­ l(C6H5)3P^NC6H5]H- (5) trolysis are presented in Table 4. In addition 〔(C6H5) 3P-NC6H5：H- + HS—T to these, the products included benzene and an (C6H5)3P + H2NC6H5+S- (6) incompletely characterized compound. The formation of TPP and aniline indicate Clearly, the mechanism of the reduction of that the reaction occurs by cleavage of the PPPI in acetonitrile is more complicated than phosphours一 nitrogen double bond, while the that of simple organic compounds (e. g., aromat­ nitrogen一 phenyl linkage remain intact. The ic hydrocarbons, azocompounds24). The voltam­ known reduction potential of TPP shows that metric and controlled-potential electrolysis it is not reduced at the cathode potential and results suggest a reduction mechanism possibly no reduction of aniline is observed in the 0 to of the following type. For the first step in the —2. 8 V vs. see potential range. reduction of PPPI a mechanism in which a one- The triphenylphosphine oxide isolated from electron transfer leads to the formation of anion the electrolyzed solution may be formed by a redical can be suggested. reaction of TPP with the peroxide anion25 (c6h5)3p-nc6h5+^—c(c6h5)3p=nc6h5]- formed by the reaction of oxygen.

(1) O2 + 2。■ * Or (7)

Vol. 18, No. 5, 1974 352 朴鍾民 • Wilson M. G'jlik, Jr. hydrofuran(THF) at —10° resulted in a "疋 (GHQM + C侦一스 (GH5)3P = O + O= (8) solution with a 2S-!ii3e csr spectrum. By con­ Little is known concerning above reactions trast, potassium reduction in DME resulted in a beyond the fact the product indicated has been red-brown solution with a 10-line spectrum. detected. The voltammetric data support the They reported that the species in THF solution view that TPPO proceeds by a one electron is TPPO and the 10-line spectrum has a very reduction of the anion radicals,27. similar appearance to that reported for (CgH5)厂 (C&压)3P=O + c=〔 (C6H5)3P=O) - (9) P? by Britt and Kaiser.30 The difficulty in observing an esr spectrum of Controlled-potential coulometry and product TPPO anion radical3;2S and our results indicate analysis suggest that the reaction of TPPO that the TPPO anion radicat is unstable. It anion radical is as follows. seems reasonable that previous nonaqueons elec­ 〔(GH 私 P=。〕=+HS—> trochemical studies on organo-phosphorus system (c6h5)2poh+c6h5-+s- (10) *5 show that if the phosphorus atom is actively Diphenylphosphinic acid is probably formed involved in the electroactive group, the product during the work-up of the electrolyzed solution. is generally unstable and decomposes rapidly. Diphenylphosphinous acid is known to be un­ Since the current function of the cathodic stable and presumably reacts with oxygen to wave of PPPI decreases with increasing sweep produce diphenylphosphinic acid. rates in cyclic voltammetric data, it seems likely A similar cleavage of a phenyl group from the that a chemical reaction is 시 osely coupled to phosphourus— linkage to phenyl radical the primary electroreduction. and diphenylphosphide anion has been pro­ (C6H5)3P-NC6H5 — e posed by Santhanam and Bard5 in the reduction ((C6H5)3P=NC6H5}- (1) of TPP. Analogous phenyl cleavage of TPPO ((C6H5)3P=NC,sH5r + HS 一 by sodium was reported by Hoffman and Tesch [(C6H5)3P-NC6H5JH-+S- (4) 2气 This is an excellent example of the direct comparison of electrochemical and chemical (CbH5)3P=NC6H5]H- + e 二 = reactions. They suggested that after the initial ((C6H5)3P^NC6H3JH- (5) formation of a radical anion-either absorbed at An additional ece type process involving TPPO the metal surface or in low concentration in can be postulated to account for the observation solution~it collapses to phenyl radical and phos­ of 3〉2) phide or phosphinite anion. The predeminent ) ) ， portion of the cleaved phenyl group forms ben­ (C6H5 3P=O + # = *5(C 3P = O (9) zene. Thus the formation of diphenylphosphinic (c6h5)3p-o-+hs ―느 acid and benzene is in agreement with the pro­ (C6H5)2POH+C6H5+S-* (10) ducts obtained by the reaction of sodium with zO TPPO in 1, 2~dimethoxyethane(DME). (C6H5) 2P\ *+ = Products (11) 、OH Cowley and Hnoosh3 also suggested that the nature of free radical species derived from TPPO We conclude that the reduction of PPPI can depends on the alkali metal and the solvent. be characterized as both the ec and ece mecha­ Reduction of TPPO with potassium in tetra­ nisms, where an ec mechanism involves reac­

Journal of the Korean Chemical Society Triphenylphosphine Penylimide 의 전기 화학적 인 환원 353

tion of the electrogenerated product to a non­ 13. H. Staudinger and J. Meyer, HeL Chlm. Adat electroactive species, while an ece mechanism 2, 635(1919); H. Staudinger and E. Hauser, implies the product of the chemical reaction is Helv. Chim. Acta. 4, 861 and 887(1921). reduced at the potential where PPPI is reduced. 14. W. Wiegraebe and H. Bock, Chem. Ber., 101, 144(1968). 15. W.M. Gulick, Jr., J. Amer. Chem. Soc., 94, References 29(1972). 1. F. Hine, H. Plust and H. Pohlemaum, Z. 16. J. J. Lingane, "Electoanalytical Chemistry” 2 nd Anorg. Allgem. Chem., 272, 25(1953). Ed., P. 456 Intersicence Publishers, New York, 2. W. M. Gulick, Jr. and D.H. Geske, J. Amer. N.Y., 1968. Chem. Soc.t 88, 2928(1966). 17. T.L. Marple and L.B. Rogers, Anal. Chem., 25, 3- A.H. Cowely and M.H. Hnoosh, ibid., 88, 2595 1351(1953); S.A. Moros, ibid., 34, 1584(1962). (1966). 18. C.M. Pak and W.M. Gulick, Jr., Electrochimica 4. E.A.C. LucRen and C. Mazeline, J. Chem. Soc., Acta, 1025(1973). (A) , 439(1967) 19. A.H. Maki and D.H. Geske, J. Chem. Phys,, 5- K.S.V. Santhanam and A.J. Bard, J. Amer. 30, 1356(1961). Chem. Son, 90, 1118(1968). 20. R.S. Nicholson and I. Shain, Anal. Chem,, 37, 6. K. S. V. Santhanam L. O. Wheeler, and A.J. 178(1965). Bard, ibid. 89, 3386(1967). 21. R.S. Nicholson and I. Shain, Anal. Chem., 36, 7. K. S. V. Santhanam and A. J. Bard, J. Elec- 706(1964). troanal. Chem., 25, App. 6〜9(197。). 22. H.H. Jaffe and L.D. Freedman, J. Amer. Chem. 8. A.R. Metcalfe and W.A. Waters, J. Chem. Soc. Soc., 74, 1069(1965). (B) , 340(1967). 23. Sadtler UV spectrum 1129, Sadtler Research 9. K. Dimroth, N. Greif, H. Perst and F. W. Laboratories, Inc., Philadelphia, Pa. 19104. Steuber, Angew. Chem,, 79, 58(1967). 24. J.L. Sadler and A.J. Bard, J. Amer. Chem. Soc., 10. A. Rieker and H. Kessler, Tetrahedron, 24, 90, 1979(1968). 5〜33 (1968); A. Rieker, Z. Naturfors사 ！ ., 216, 25. D. T. Sawyer and J. L. Roberts, Jr., J, Electro- 647(1966). anal. Chem., 12, 90(1966). 11. See for example Chem. and Eng. News, P. 34, 26. S. Wawzonek and J.H. Wagenknecht, "Polar。- ； Jan. 13, 1969 S. H, Rose. J. Polym. Sic. Part graphy 1964” Macmillan, London, P, 1035, 1966. B. 6, 837(1968); M. V. Lenton? B. Lewis and 27. M.I. Kabachnick, V.V. Voevodskii, T. A. Mast- C. A. Pierce, Chem. Ind. (London), P. 1387, ryukova, S.P. Solodovniku and T. A. Melenteva, ； 1964 F. Goldschmidt and D. Dishon, J. Polym. Zh. Obshch. Khim., 34, 3234(1964). Sci., 3, 481(1948); H.R. Allock, R.L. Kugel 28. G. M. Kosolapoff, “Organophosphorus Com- and K.J. Valan, Inorg. Chem., 5, 1709(1966). pounds**, P. 171, John-Wiley and Sons, Inc.f 12. D.H. Geske, J. Ragle, M. Bambenek and A. New York, 1950- Balch, J. Amer. Chem. Soc.t 86, 987 (1964)； 29. A.K. Hoffnann and. A.G. Tesch, J. Amer. Chem. K. Kawata and D.H. Geske, ibid., 86, 2101 Soc., 응 1, 5519(1959). (1964)； L. R. Faulkner and A.J. Bard, ibid., 30- A.D. Britt and E.T. Kaiser, J. Phys. Chem., 90, 6284(1968). 69, 2775(1965).

Vol. 18, No. 5, 1974