COMMUNICATIONS TO THE EDITOR
Oct., 1963
. 45
-3
Wavenumbers x I O
30
35
40
r
70
1
2243
boron “p”-orbital chromophore, implying a t least a charge transfer type transition. This cannot be correct since the transition energies do not show the required correlation of decreasing with increasing ionization potential of the aromatic ring. Therefore, the transition must be one involving the nonbonded electrons of the oxygen or nitrogen atom. It can be considered as an intramolecular charge transfer transition from the oxygen or nitrogen to the vacant p-orbital of boron. The excited state can be represented in valence bond structure as R2B-=X+-R where X is some atom with a nonbonded pair of electrons. I n support of this argument, it has been observed6 that the oxidation productgof trimethylboron, undoubtedly a mixture of methyl boronates and borinates, show an intense absorption a t 260 mp, which can only be due to the boron oxygen chromophore. (6) A. G. Davies, D. G. Hare, and L. Larkn-orthy, Chem. Ind. (London), 1519 (1859). (7) Department of Chemistry, Pennsylvania State University, University Park, Pennsylvania.
3500
2500
31000
2000
Angstroms. triphenylboron; - -.- Fig. I.-Ultraviolet spectra: triphenylboron-ammonia complex; - - . - benzene ( e i s the molar extinction coefficient).
DEPARTMEKT OF CHEMISTRY FLORIDA STATEUNIVERSITY TALLAHASSEE, FLORIDA RECEIVED AUGUST5, 1963
B. G. RAMSEY’ J. E. LEFFLER
~
the cells loaded in a glove box under nitrogen. On exposing the solution to air, the intramolecular charge transfer disappears almost instantly with the simultaneous appearance of a maximum in the 2400-A. region, characteristic of a boron-oxygen bond. I n spectrograde acetoiiitrile purged with nitrogen, the spectrum of “triphenylboron” shows only weak benzenoid absorptions a t 2670 and 2750 A., and an intense maximum a t 2250 A.,but no maximum or shoulder near 2400 A. Similar effects to the above were found in the infrared spectra of triarylboranes. Spectra of solutions prepared and loaded under nitrogen in glove boxes do not show a strong maximum near 1350 cm.-l, a characteristic absorption of the boron-oxygen bond. On exposure of the solutions to air, the 1350-cm.-l band rapidly appears. The ultraviolet spectra of triphenylboron, triphenylboron-ammonia complex, and benzene in niethylcyclohexane are shown in Fig. 1. The absence of the intramolecular charge transfer transition in acetonitrile solution is best explained by the formation of an acetonitrile-triphenylboron complex, since the highly sterically hindered tri-l-naphthylboron and trimesitylboron have the expected charge transfer transitions in acetonitrile. This observation is not surprising in view of the fact that acetonitrile complexes with a variety of other boranes. It is interesting to note that the tripheiiylcarbonium ion shows no evidence of such strong complexing with acetonitrile, probably because of the lack of the electrostatic bonding contribution present in (CaHa)3B--N+=CCI-Is and the loss of charge delocalization in going from (CoH5)3C+to (CsH5)sCN+=CCHs. I n addition to triphenylboron, Mikhailov4 has reported the ultraviolet absorption maxima of a large number of alkoxy, hydroxy, and amino arylboranes. He has assigned this transition to the aromatic ring ( 5 ) L. J. Bellamy, J. Chem. Soc., 2412, (1958).
ELECTROCHEMICAL GENERATION OF SOLUTION LUMINESCENCE
Si?”: We wish to indicate in this paper some of the steps involved in the anodic generation of luminescence from luminol (3-aminophthalhydrazide) and also discuss the decay and lifetime of the luminescent emission. Electrochemically generated luminescence has been observed by Harvey’ and Vojiir.2 No detailed studies have been made. The proposed mechanisms of chemically produced luminescence have been reviewed recently by White.3 We oxidized luminol a t a platinum electrode in either the presence or absence of oxygen and in 0.1 F NaOH or Na2C03. The cell configuration with appropriate electrodes was similar to the one previously described4; circuits were conventional. Emitted light was monitored by an 1P21 photomultiplier tube. Luminol was prepared and purified according to a published procedure.5 The luminescent spectrum of luminol obtained electrolytically in the presence of oxygen was found to be identical with the chemically produced spectrum.6 The identity of the two spectra indicates that the overall luminescent reaction and the species involved are most likely to be identical in the two cases. The initial step leading to chemiluminescence appears to be an electrochemical oxidation of luminol. The rate of this initial step can now be conveniently controlled electrochemically. Rletal catalysts or strong oxidizing agents3 are also no longer needed to initiate luminescence. (1) N. Harvey, J. Phys. Chem., 33,1456 (1629). (2) V. Vojiir, Collectton Czech. Chem. Commun., 19,862 (1954). (3) E. H. White, “Light and Life,” W. D. SIoElroy and B. Glass, Ed., Johns Hopkins Press, Baltimore, Md., 1981,p. 183,and pertinent references therein. (4) J. N. Pitts, Jr., H. W.Johnson, Jr.. and T. Kuwana, J. Phys. Chem.. 66,2456 (1962). (5) E.T. Seo and T. Kuwana, J . Electroanal. Chem., in preas. (6) H. H. Seliger, ref. 3, p. 200.
COMMUNICATIONS TO THE EDITOR
2244
LE.
+ O2
- k2
Vol. 67
k3
products
[intermediate]
k4
4 >
[intermediate] +PI*
-0
+ Pz
J
c t
W
20 c W
n 0
a c W 0
W J
TIME (msac.).
Fig. 1.-Electrode potential and relative luminescence following electrooxidation o luminol: trace A, electrode potential as a function of time (left-hand scale); trace B, relative light intensity as a function of time (right-hand scale).
If a charge of electricity was discharged rapidly from a capacitor to the platinum anode, the potential of the electrode and the light emission responded as shown in Fig. 1. This figure is a typical photograph of a dual trace oscilloscope sweep. Luminescence is clearly initiated a t the time of luminol oxidation, as seen from the comparison of the leveling-off point of the electrode and the rise of poteiitial (trace A) a t ca. $0.45 luminescent intensity (trace B) . The luminescent intensity rises linearly with time and rapidly attains a region of ca. constant intensity. When the electrooxidized luininol is all consumed, the luminesceiice decays by a first-order rate. The half-life of this decay is 6.8 f 0.2 msec. (temp. 21 f lo)in solutions of 4 X F luiiiiiiol and 0.1 F Na2C03. Temperature dependence of rate gave a heat of activation of ca. 22 kcal. /mole. Sext, chronopotentiometric experiments were performed to gain additional iliforniation about the sequence in the reactions involving luminol and oxygen. Briefly, luminescence appeared in solutions containing 10-3 F luminol, 0.1 F Na2C03,and dissolved oxygen when the electrode potential reached a region where luminol mas oxidized. In the absence of oxygen (ca. M or less), however, luminescence did not appear until the electrode potential was sufficiently positive that oxygen was being concurrently generated with the oxidation of luniiii01.~ Lumiiiol must be oxidized prior to the reaction with oxygen to produce luminescence. I n stirred solutions containing sufficient oxygen, the light intensity was proportional to the current level except in the high current regions, where light intensity was independent of current. In this latter region, the rate of mass transfer of luminol t o the electrode limits the oxidatioii rate. The general scheme for the electroluiniiiescencc, based on our experimental results and others13is 17.
ki
LH- +LH.
+ e-
(7) '1'. liuuana, J . Blectroanal. Chem., in press.
The luminol anion, LH-, undergoes electrooxidation to form the radical, LH.. The rate, kl, corresponds to the rate of electrooxidatioii of luminol. The radical, LH., reacts rapidly with molecular oxygen to produce unstable oxidation product (s) and/or intermediate(s). The unstable intermediate, by decomposition or by an intramolecular reaction, produces an excited state fragment or molecule PI*,which in turn undergoes deexcitation to the ground state via fluorescence. The magnitude of the decay rate and the heat of activation indicates that the rate-determining step for the luminescence is the decomposition of the unstable intermediate (h.5 > k4). The lifetimes of excited state species which undergo transitions to the ground state via fluorescence are several powers of texis less than the observed time. The observed first-order decay is, therefore, not due to the fluorescent species itself. Phosphorescence is probably not being ohserved. A triplet molecule will be rapidly quenched by the oxygen present in the solutions.* Preliminary data do not support the mechanism of Bernanose, et aL,9 \Tho proposed the formation of a biradical through a bimolecular reaction of radicals (LH .). We favor a nieclianisni iiivolviiig a similar intermediate as proposed by White.3 It is particularly attractive since it does not impose energy transfer to neutral luminol but does produce a product (aminophthalate ion) whose fluorescent spectrum in base is similar to the lumiiiol luminescence.G Also, the rapid growth of the light emission followed by a period in which the light intensity remains almost constant (trace B i n Fig. l),with the exception of a slight decrease until the first-order decay was reached, favors a scheme where a transient intermediate approaches a steady-state concentration. The products, PI aild Pz,are suspected to be the aminophtl-date ion and nit r ~ g e n respectively. ,~ Seither product, however, has been directly related to the luninescent reaction. Voltammetric and electron spin resonance studies, as well as the isolation of intermediates and products, are in progress to verify the quantitative details of mechanism. (8) C. Reid. "Excited States in Chemistry and B~olog)," Butterworth Scientific Publications, London 1957, p. 100 (9) A. Bernanose, Thr. Bremer, and 13. Goldfinger, BulE. soc. ch?m. B e l ~ e s66, , 260 (1947).
THEODORE KUWANA DEPARTMENT O F CHEMISTRY BARRYEPSTEIX UNIVERSITY OF CALIFORNIA RIVERSIDE, CALIFORKIA EDDIET. SEO RECEIVED A U G L ~27,I ~1'363
