HENRYLINSCHITZ A N D KYOSTISARKANEN
4s25
pure ether and 1: I for the pure alcohol, thus can be explained. The cornparison of the present data with the average values reported by Gutowsky is generally good. The deviations that exist are Drobablv clue largely to medium effects since GLtowsk; examined mainly pure organic liquids while our data are-~-for two of the compounds-for infinitely rlilute solutions in the “spherical” solvmt* CCl,,.
[COXTRIBUTIOS F R O M THE
CHEMISTRY DEPARTMENTS
Vol,
so
Acknowledgments.-It is a pleasure to thank Professor J. S.Waugh and Mr. Richard Fessenden for their cobperation and advice. This work was supported in part by the C . S. Altninic Energy Commission ~~
( 4 ) A.
~
h. Tiothncr-By
atid
I?. IC. Glick,
TtIIs JOURNAL,
78, 1071
CAMBRIDGE 39, MASS.
OF SYRACUSE
UNIVERSITY AXD
BROOKIIAVEN
NATIONAL LABORATORY]
The Absorption Spectra and Decay Kinetics of the Metastable States of Chlorophyll A and B1 R Y HENRYLINSCIIITZ? AND KYOSTI SARI;AKEN RWFXVEII :IPRII.’7, 1958 Using flash-illumination apparatus with improved time-rrsolution, the alxorption spcctra and decay kinetics of the mctastable states of chlorophyll A and B have been studied in qyritlinc antl benzene solution at 25’. The spectra are peculiarly diffuse and a measurable absorption extends heyond 7500 A . Tlie decay kinetics obcy the following ratc law: -tlC*/dt = ktC* k?(C*)l k3C*C, ( C * = conccntration of the metastal)lr, C, = ground state). Values of the k’s are presented and the results are discussed. Tlie fast k r proccss does not nppcnr to hc due simply to effects of pnrarnagnctic molecules on the triplet-singlet conversion.
+
+
Introduction The formation of a long-lived excited state of chlorophyll and its function as the energy-storing and catalytic agent in photosensitization have been postulated for some time, on the basis of much indirect evidence. This has included data on the kinetics of sensitization, fluorescence quenching and steady-state b l e a ~ h i n g . ~ -Recently, ~ direct demonstrations of this metastable state (probably the triplet) have been given by flash-excitation techniques, and preliminary estimates of its absorption spectrum and lifetime have been obtained.“-’O It is desirable to extend the spectral data, both in wave length coverage antl precision, particularly in view of recent ideas on the mechanism of energy accumulation and transfer in photosynthesis. 11~12 Moreover, the study of the kinetics of conversion of the triplet t o the ground state bears directly on the general and important problem of the nature of radiationless transitions in complex molecules. Existing kinetic data on the tleactiv at’ion rcactions of the metastable state are very rough. This uncertainty is due mainly to the relatively long (1) This work was assisted b y a g r a n t from t h e Ti. S. Atomic Energy Commission t o Syracuse KJniversity (Contract No, AT-(30-1)-8?0), and was carried out in p a r t while 11. 1. was on snhhntical IM\Y nt
Brookhaven National T,ahoratory, (2) D e p a r t m e n t of Chemistry, Brandeis Kiniversity, Walthnm. Mass. ( 3 ) E. Rabinowitch, “Photosynthesis,” Interscicnce Publishers, Inc.. New York, N. Y., 1956, Vol. TIT. Chap. 35, and earlier voliimeq. (4) J. F r a n c k , Ann. Rrv. P l m t PhysioZ., 2, 5.7 (1051). ( 5 ) R. Livingston a n d V. A. R y n n , T H I S J ~ I J R V A ~ ,75, , 2170 (1853). ( 6 ) R. Livingston, G. Porter ani1 11. Windsor. A’afikre, 173, 48.7 (I!),;$). (7) E. W. Abrahamson an11 13. I.inschitz, J . Cliem. Pltss.. 23, ?1!l8 (1955). ( 8 ) R. I,ivingston, T H r s J O U R N A L , 7 7 , 2170 (195.5). (9) H. Linschitz and E. W. Ahrahamson. “Research in Photosynthesis,” Interscience Publishers, Tnc., S e w Ynrk, 1057, p. 31. (10) R. Livingston, ref. 9,p. 3. ( 1 1 ) J. F r a n c k , ref. 9, p. 142. ( 1 2 ) I;. I,. Allen antl J. Pranck, Arch. Biochenz. and Biophys., 68, 124 (1955).
duration of the exciting flashes which have been used so far, comparable in fluid solvents to the lifetime of the inetastable state itself. The overlapping of excitation and deactivation not only greatly complicates the kinetic analysis, but with oscillographic recording of absorption changes, necessitates large scattered-light corrections over most of the decay curvc, with consequent further loss of accuracy. 111 this paper, we present new data on the nictastable states of chlorophyll A and B , taken with apparatus providing an exciting flash of much shorter duration than those previously used. .\bsorption spectra have been measured over the range 3500’7.500 A. The kinetic data establish t h a t only one metastable species trf chlorophyll is present in pyri(line or benzene solution, ant1 permit the evaluation or estimation of three rate constants involved in the transition to the ground state. The improved tiinr rcsolution also makes feasible the direct study of bimolecular reactions betwerii the mctastahlc qtatc and added qucnehers. Experimental 1. Apparatus.l3--C1iaiiges in absorption of a 3tcady light bcam, passed through the sample, were rrieasurcd by a photocell and oscilloscope ns function of time after excitation, the flash being triggered a t a suitable interval after starting thc smccp. T h e apparatus is shown diagrammatically in Fig. 1. Mullard “Arditron” lamps, Type LSD-2, were used as the flash source.14 These are short, wide tubes, with a straight 40 mm. spark gap, filled t o 1 atm. with argon containing a small addition of hydrogen and, when fresh, have a breakdown potential near 11,000 volts. This permits t h r use of high-voltage low-capacity storage condensers to minimize flash duration and obviates the need for auxiliary spark gaps. Triggering of these high-pressure lamps is facilitated by using copper electrodcs, activated b y impregnation with (13) This a p p a r a t u s was developed and built a t t h e B r o o k h a r r n National Laboratory, in collahnration with Mr. Lloyd Nevala a n d Mr. Fred Merritt, of t h e Instrumentation Section. It is a pleasure to acknowledge their indispen?:ililr assistance, as well a s advice from Mr. W. Higinbotham. (13) J \V. RTitrhrll. ’ / ‘ ? , i ? i v i l l r c n i R n g , .9oc. (I,o?rdnn), 14, 91 (1948).
Sept. 20, 1958
SPECTRA OF
THE
METASTABLE STATES OF CHLOROPHYLL A AND B
Fig. 1.-Diagram of flash-illumination apparatus: 2 , d.c. zirconium arc; L, collimating lens; S,shutter; C, cell containing test solution; F T , flash tube; R, elliptical reflector; B, baffle box; 0, mounting for filter or calibrated screen; L’, condensing lens; M, n~onochromator;PPVI, photomultiplier tube; CF, cathode folloii er; CRO, oscillograph; P, timing-pulse generator; D, delay; H, high-voltage firingpulse generator; Tr, trigger lead. potassium. I n our apparatus, the storage capacity (per flash) was further decreased by distributing the necessary energy among four Arditrons, which were mounted, respectively, a t the outer foci of a four-leaved elliptical reflector. The cylindrical absorption cell, containing the solution t o be irradiated, was supported at the common central focus of the four-leaved ellipse, with its axis parallel to those of the surrounding flash-tubes. Each lamp was powered by its own low-inductance, 0.5 pf. condenser, isolated from the others by resistors, and operated at 9,000-10,000 volts from a common high-voltage supply. The reflector unit was fixed in a rigid metal frame between the storage capacitors, which were lamp terminals with short strap leads. The lamps were mounted in the reflector on bases which were adjustable for optimum focus. A powerful firing pulse was found necessary t o ensure simultaneous flashing of all four lamps. This mas generated by discharging a 0.01 pf. capacitor, charged t o 9000 v., through a hydrogen thyratron and the primary of a special pulse transformer. The hydrogen thyratron was itself fired by an initiating trigger pulse from the timing circuit. T h e resulting output firing pulse (greater than 30 kilovolts, with rise time less than 0.1 psec.) was then distributed t o each of the four lamp trigger electrodes through 100 ppf. high-voltage condensers. At 9000 volts or higher, using fresh lamps, firing of all four lamps occurred with a time spread less than 0.2 ~ s c c . The output trigger circuitry was mounted in the frame directly above the reflector assembly, while the timing pulses were provided by a Grass “Stimulator” (Model S4B). T h e measuring light source was a 100-watt zirconium arc (Sylvania, Type T-100), operated from a d.c. generator. Its light was collimated by a short-focus achromatic lens and passed through a shutter, the absorption cell, through a cylindrical lens and into a Bausch and Lomb 250 mm. grating monochromator. T o increase the ratio of collimated measuring light t o scattered light from the flash, the distance between the absorption cell and monochromator condensing lens was made large (70 cm.) and a series of baffles were placed in this interval. Calibrated screens and filters could also be inserted in this portion of the optical path. Light from the monochromator exit slit was measured by a n endwindow photomultiplier tube (RCA, Type 6217), whose output was d.c. coupled through a cathode-follower t o a wideband oscilloscope (Tektronix, Type 531). With the cathode follower, photo-cell load resistors up t o 10 Kohms could be used without serious distortion of the flash profile (2 psec. decay time). I n the kinetic measurements, involving much longer decay times, load resistors of 25 or 50 Kohms were used. The photomultiplier power supply was designed to render negligible a n y shifts in dynode bias during the measurement. Absorption cells were made of Pyrex tubing, 50 mm. optical length and 13 mm. i.d. A narrow side-arm, placed at
4827
one end of the cell, led to a n ampoule in which the degassed dye solution was prepared and sealed off from the vacuum line. For work at high concentrations, smaller cells were used. 2. Profile of Exciting Flash.-At 10,000 volts and 0.5 pf., with low-inductance circuitry, the Arditron lamp provides a flash of about 2 psec. duration, measured t o the halfdecay point, in the blue region of the spectrum (Fig. 2A). Somewhat longer durations are obtained at longer wave lengths (2B). T h e profile also shows lesser peaks, corresponding to the oscillatory discharge. For comparison, we note t h a t in their original work on this subject,6 Livingston and Ryan used 200 psec. flashes (measured t o half-decay point) while Porter and Windsor’s equipment for studying triplet states of aromatic hydrocarbons16 gave 35 psec. flashes at discharge energies comparable to those used here. However, for precise kinetic work, the minimum resolution time is much longer than a n y of these figures indicate, since raregas discharge lamps all show a very long tail, particularly at longer wave lengths (Fig. 2C). While the use of dissociable gases in the filling renders this tail relatively weak, the absolute intensity is still such that appreciable light is emitted at times as long as 20-fold half-decay time, and the integrated light output for the long tail may be a considerable fraction of that in the main peak.
Fig. a.-Light-flash profiles at various wave lengths: A and B, 4psec. per grid; C, l0psec. per grid; A, 400 mp; B, 650 mp; C, 650 mp. 3. Procedure.-The measuring procedure was generally Sweeps (and flashes) similar to t h a t described previously were photographed with the shutter in the measuring beam first open, then closed. Since scattered or fluorescent light from the flash is the same for the two sweeps, the “closed shutter” trace, including the flash profile, represents the base line (zero transmission) for the measurement, and changes in the vertical separation of the two sweeps measure changes in transmission. For times longer than about 20 microseconds after flash initiation, no scattered light cor(15) G . Porter and hl. Windsor, F n r o d o y SOC.Disc., 17, 178 (1954).
3828
HENRY LINSCIIITZ AND KYOSTI SARKINE~S
rection was necessary at all, except in t h e very far red, and the base-line could be taken t o be entirely level. Calibration of the oscilloscope deflection in terms o f per cent. transmission was initially done b y recording blank (no flash) sweeps, in which a solvent-filled cell a n d a calibrated neutral filter were substituted for the test solution. Since closely spaced neutral filters m a y be used for calibration, this procedure allows small changes in transmission t o be spread over the full oscillograph scale, b y suitable adjustment of slit width a n d gain and permits measurement at high optical densities. Transmissions obtained in this way generally agreed well with those given b y conventional spectrophotometers. However, small differences in placement of absorption cells, window quality, etc., led occasionally t o considerable errors, due t o the long optical path between absorption cell and detector. Further small errors were caused by variation in intensity of t h e measuring source and in sweep position during the time required t o change cells arid filters. I n establishing t h e spectra of t h e metastable state, more consistent results were obtained when t h e oscilloscope scale readings on a single sweep, immediately before and after the flash, were used t o determine only the relative changes in transmission. Let Ri and K 1 be the oscillograph deflections measuring transmitted light intensity before t h e flash and a t time t after the flash, respectively. Thus, the change in absorbance is A D = log ( R i / R t ) . :Idding this AI1 t o the original absorbance of the sample, as measured iiidependently (Cary, Model 11) then gave the absorbance of the excited solution at time t after the flash. T h e scale reading corresponding t o initial excitation (representing maxiniuni conversion) was found b y a short extrapolatioii of the tlcflected trace, back t o the instant of triggering the flash. To facilitate this, measurements intended t o establish tlie nietnstable spectra were made with relatively fast sweeps, to flatten t h e decay curve. This technique permits g o i d prccision in measuring AD'S, up t o about 0.8. T h e spectra of t h e excited solutions immediately after the flash were corrected t o 100% conversion by assuming t h a t the new absorption had no peak in the region of the original
Vol.
so
red band. .I straight line was drawii between experimental Iwiiits near or a t tlie isoshestic poiiits lying on either side of the Origiii~tlp e d i and the residual red absorption attributed to unclianged elilorupli~-ll. This fractional conversion ivas t heii used t o correct the observed s p e c t r u ~ ~ iTo . establisli the inctastable spectrum in the blue and green regions, dilute clilorophyll solutions (-10-6 -21) were used, for which the j,. Tile correction \vas tlius iriitial convcrsioii wis over $)i)' stndl, : I I I ~ thc cnlculntion cuuvergeti imiiiediately. Tlic 5pectr:t iri the ret1 i r e r c oljtciiiied a t solneivliat liigher C O I I centrations. \\.itliin t h e precision [ i f these iiie:isurcn:ents, tlie mctast:ihle spectrurn obeys Rcer's Inn.. Sample col1ceritr:i tiiiiis and gain Irere adjusted iu the fitlal ruiis o n the h s i ? o f 1)relitniiiary nieasurernei~tsto give maximum on-scale tlcflectirms xt each n a v e length imd a t reascinable monochromator slit widths. Linearity of the phutometric system and t h e base w a s checked frequently, using calibrated screens and :I precision time-rnark generator. For rneasureiiieiit, eiilarged images of tile oscillogra1t1s r e r e projected onto graph paper and traced. A smooth curve, xver:iging o u t noise, then was drawn through each iweep profile, ant1 tieflectiini? tlcterriiinetl from this. Solutioii temperatures were rnaiiitainctl a t 25 =tI " , 11)equilil?ratirtg the ithorption cell in a therniostnt inilnediatel!. before e:icli SilcJt. 1Ic:ttiiig effects rluc t o the flush were slight. 1 general iden r;f apparatu.; perfornlatice, typical s are shown iu Fig. 3 . .At 480 IIIN, for esainple, full-scale deflccticiiis could he obtained covering the trxnqmission range ()-.?r (', with :I !iiiinochroln:rtr,r slit xvidth of 40 o of about 20. higher transiiditiotis were of course possible. vc length ( ; ~ o o .$.I slit miiltIis of 100 A . were required. In degassed pyridinc solution, tlie photo-1,leachitlg reaction was found t o he highly reversible, less than 576 decompiisitioii (judged from absorption spectrum) resulting after 100 flashes. Iii degassed benzene. tlie bleaching reaction is soinewhat less reversible 4. Materials.-Tliiiiplleiie-free benzene (Mallinckrodt) \vas shaken with H2SO:, water, dried with S a 8 0 1 and distilled from C;tH2. General Chemical reagent-grade pyridine was allowed to stand over Ha0 and distilled from fresh DnO. These solvents Irere degassed on t h e vaeuul~iline IIY cycler of freeziug, pumping arid tllawitig, and then severxl trap-to-trap distil1atio:is over CaH2 and BaO. Chlorophyll A and I? were prepared from fresh spinach by the rnethod of Jncobs, et 0 2 . , ' 6 anti stock solution in purified toluene stored :it -20". Toluene l r a s used rather than ether t o rninimize peroxide contamination. To prepare a test sample, n suitable volume of stock solution was placed in t h e ampoule carrying the absorption cell, the stock solvent taken off o n the vacuuni line, freshly degassed benzene or pyridine distilled in, and t h e preparation sealed off. The lifetimes ( k , belolr) v-ere found t o be highly sensitive t o the quality o f the x v x u t n (oxygen quenching) and otil>- solutions sealed off under "sticking-vacuurn" were used. Final concentratiotis were measured spectrophotorIietricall~,using itio1ar estinctioii co-efficients determined in separate experi111e111.;.
Results and Discussion
Fig. 3.-Typical oscillograms: total sweep time, sec.; slit width, 2 mp. Upper picture, Shot h-0. 26-11, chlorophyll X in pyridine; 1.73 X 10-6 M ; wave length, 442 nip. Lower picture, Shot S o . 35-3, chlorophyll B in pyridine; 3.25 X 1 0 - 6 M; wave length, 513 mp. In each picture, upper trace is with shutter open, lower trace with shutter closed (base-line). Sharp break marks instant of occurrences of flash.
1. Spectra of Metastable Forms.-The absorption spectra of the metastable states of chlorophyll h and B are shown in Figs. 4 and 5 , and extinction coefficients are given in Table I. Based on the reproducibility of our results, the extinction coefticients probably are accurate to 5% or better in the green region of the spectrum. Under the Soret band and in the orange and red, accuracy is somewhat less. The spectra of X and B are closely similar, the main peaks (at 4G2 and 485 5 mp, respectively) lying just a t the long-wave side of the corresponding Soret bands. In the green, both spectra show broad plateaus, tailing off into gradually weakening absorption which extends into the
*
(16) 15. E. Jacubs. A . li. V a t t e r ;inrl A . S . H d t , A y r h . Biochem. aizd E i o ~ h y s . 53, , 228 (1951)
Sept. 20, 1958
SPECTRA OF
THE
METASTABLE STATESOF CHLOROPHYLL A AND B
red a t least as far as these measurements can follow. The most remarkable feature of these spectra is their extremely diffuse character. I n view of the sharpness of the original bands, this can hardly be due to vibrational structure, and presumably several overlapping electronic bands are involved.
TABLE I SPECTRA OF METASTABLE CHLOROPHYI.L .I PYRIDINE Molar extinction coefficients: ---Chlorophyll
400 462 505 530 0.8
0.2
500 600 700 Wave length, mp. Fig. 4.-Absorption spectra of chlorophyll A and its metastable state, in pyridine; 2.1 X 10-8 M ; cell length, 5 cm. : 0, corrected spectrum, using relative calibration of oscillograph deflection ; 0,corrected spectrum, using absolute (neutral screen) calibration; 8 , original uncorrected absorbance, immediately after flash. 400
t
500 600 700 Wave length, mp. Fig. 5.--Absorption spectra of chlorophyll B and its metastable state, in pyridine; 2.1 X 10-6 M ; cell length, 5 cm.: 0, spectrum corrected to 100yo conversion; 63, original uncorrected absorbance, immediately after flash. 400
A-
a X 10-4
A , mr
min. max. min. mas.
2.57 3.20 1.74 1.78
-cy
=
(l/cZ)
ASD
B,
IN
log10 (10/1)
Chlorophyll B -
a X 10-4
X, mp
380 400 445 465 485 550 605
4829
max. min. max. min. mas. mas. max.
1.97 1.89 2.60 2.41 3.47 2.15 1.21
In the “B” spectrum, the indications of a small peak a t GO5 mp are outside our experimental error. However, the suggested peak near 675 mp is somewhat more suspect, since the shape of the spectrum in this region is sensitive to the assumptions made in correcting to 100%) conversion. There is no question whatever of the enhanced absorption in the far red. These new data permit a review of earlier sugg e s t i o n ~that ~ ~ certain ~ ~ ~ general relationships exist between the spectra of the metastable forms and the phase-test intermediates (“enolate ions”) of chlorophyll. The phase test products17 show diffuse absorption in the same regions found here for the metastable molecules, and i t is striking that the enolate ion of chlorophyll-B has a very broad and weak absorption extendiag from the main peak in the green to about 7800 A.18 At low temperature, chis sharpens somewhat to a broad flat band a t 7400 A.I8 While these similarities are marked, differences also appear. The far red absorption band of the chlorophyll-A phase-test product is much sharper than the long, flat tail of the corresponding metastable form, and the relative intensities of the blue and green peaks of the two intermediate forms do not agree. 2. Decay Kinetics; Evaluation of Rate Constants.--Typical examples of the extensive kinetic data taken in this work are given in Table 11. Trace deflections above the base line, as measured on the projected oscillograms, are listed for various concentrations, wave lengths and times t , after an arbitrary zero time to. This was taken a t 50-150-p sec. after the flash to avoid any possible tailing effects and scattered light corrections. For use in calculation, Table I1 also gives values of a i , defined below. The data are conveniently handled in terms of a plot of log AD,/AD vs. t, in which AD, and AD are the changes in absorbance a t times to and t, respectively, relative to the absorbance of the solution before flashing. Figure G shows such typical curves for chlorophyll B in benzene. Since AD is proportional to the concentration of metastable molecules, it is evident from Fig. 6 that the kinetics of the recovery reaction deviate markedly from first order. Experiments in which flash intensity and total dye concentration were varied indicated t h a t higher-order terms involving the excited state were mainly responsible for this behavior. The data therefore were fitted to a rate expression as (17) A. Weller, THISJOURNAL, 76, 5819 (1954). (18) H. Linschitz, in preparation.
170i. so
HENRYL,rNscrmz AND KYOSTISARKANEN
'IS30
TABLEI1 FLASHBLEACIIIXC OF CHLOROPHYLL A ASD l3. TYPICAL EXPBRIMESTAL DATA ,
Chlorophyll A in pyridine--
Run nn. (mi4 Concn., AI X I O G to, pscc.
16-13 47.i
26-11 442
3.00 1 Oil
inn
1.73
26-8 671 1.x
480
ion
iijit
Time, p e c .
Chlorophyll B in benzene---
30-17 458 2.16 150
11(1
31-3 647
2.16 100
In
5. (i9
(i. 86 f i . 50
TO1
100
f i , iI7 (i.39
5
150 200 2 50 300 400 500 600 700
0 . ti4
800
. .
li.22 5.98 5.58 5.24 4.96 4.73 4.39 4 13 :3 93
..
7.05 7.36 7.60 7 80 7 90
9.77 6.69 7.33 7.79
3 82 9.2[i 8.7.5 3 ni;
x.:n
2 82 11.08 0 33 8.12 7.23 6 .fi0
8 .(?7
(i.08
8.91 !I 07 9 19
5.69
,
8. 11
7.71) 7.20 6.79 6.46 6.20 5.75 6.40 5.11
5.08
4.68 4.m
*-,74
10 82 10 , :HI 9.86 9.40 9 . It\ 8 ,G,5 8.27 7.97 7.72 7.36
7 os 6 87 ii . 70
4 l?
4.92
29-11
30.16
,5on
500 2.16
11 100
100
K"
Deflection from 7ero trnnsmis4on line,
2.83 7.27
Bcfnre flash 0 25 50 75
9 03
10 2G 3 19
ii 4-1 ii, G 5
6.84 7 ill
1 XP
7.17 7 . '13 7 . (i?
3 . 93 6.67 7.23
..
7.81
Y67
7.w s.20
8.00 8.48
8.39 S . 52 8 04
8.82 9.n; 9.2ii
7 78
7 78
ii .SD
x
10"' 8.51 3.21 3 , 36 All mensurcinents in 5 cm. cells, escept 31-17 ( 1 . 5 cm.). ah
31-17
')
9.72 1.92 For 5 cm. cell.
r, 2s
Co = total chlorophyll concn
Txt
C* = concn of metastable chlorophyll C, N*,
= concn. of unexcitcd chlorophyll
an = mnlnr extinction cocfiicicnts =
1
= lcngth of absorption cell
CHL. B
1 loglo To/Z CI
0
. a
t
0.0
We assume
+ kz(C*j2 + k3C'C,
dC* - k,C' tl I
(1)
But
c,
=
co
-
no
a
0
c*
/
..
-
-
-1
04
= t7xAD
--
where ax = (N*
/
1 - nn)Z
02
The rate expression then is cl(AD) -_ d1
(ki -k ksCo)AD
Letting A = ki
+
/+
0.6-
0
/
P
(3
and C'
- BENZENE
-
-
-I
,=--I i
if /
-2
16 X
--- I IO
M
x IO-^
M
4
r7h
(k?
-
/??)(AD)'
(21
I
1 1 200
400
600
TIME, MICROSECONDS.
+ k?Co
Fig. 6.-Plot of log ( A D ~ / A Dus. ) time for typical decay: chlorophyll B in benzene; Shots No. 30-16 and 29-11; h 500 mp.
and Bx = ox (k? - k s )
the integrated rate law is then (3)
A and B may be separated by first evaluating A from the asymptotic first-order slope, as AD goes to zero. However, the second-order term is relatively large, and very long cells must be used to obtain sufficiently accurate data. For uniform illumination under these conditions lamp size must be increased, with accompanying increase in flash duration and loss of time resolution. Accordingly we writc cquation 2 as
The timc derivative is obtained directly by drawing tangents to the plot of log (AD,/AD) us. t (Fig. 6). These rates are then plotted against the corresponding values of AD to obtain A and Bx. For any given sample, measurements a t various wavc lengths should be represented by a family of straight lines having the same intercept A , a t AD = 0, but with different slopes corresponding to the wave length dependence of ax. Conversely, measurements a t a single wave length but with increasing Co should yield a family of lines of constant slope but increasing intercept. The variation of A with COthus gives k1 and k3. Then, since the spectral data (corrected to lOO7, conversion) give ax a t the wave length of the measurement, and BA a t the
same wave length is known, kz can be calculated. Representative results of such treatment are shown in Figs. 7 and 8, in which eq. 4 is applied to data on chlorophyll A and B in pyridine and benzene respectively. Among others, runs given in Table I1 and Fig. 6 are included in Figs. 7 and 8. In agreement with the proposed kinetic scheme, satisfactorily converging straight lines are obtained for measurements a t various wave lengths and constant concentrations. This has been verified throughout the entire visible spectrum and constitutes very strong evidence that, a t least in pyridine and benzene, only a single long-lived intermediate is formed by light excitation.
I 30
3.0
x
i
C H L . B IN BENZENE 500rnu
0
2 I6 X I 0 6 M I IO x I O - 5 ~
0.2 0.4 CHANGE IN A B S O R B A N C E .
Fig. 8.-Dccay
,
l
,
l
,
l
,
l
,
0.2
0.4 CHANGE IN ABSORBANCE.
l
0.6 AD,
kinetics of chlorophyll A in pyridine; application of eq. 4.
Runs a t constant wave length give generally parallel lines, displaced to larger intercepts with increasing concentration (Figs. 7, 8) again in agreement with eq. 4. The consistency of the data with the kinetic scheme can be checked further by comparing k2’s obtained from measurements a t different wave lengths. Satisfactory agreement is found. The rate constants are summarized in Table 111. TABLE I11 SUMMARY O F RATECONSTAXTS Chlorophyll A Chlorophyll B Pyridine Benzene Pyridine Benzene k a , 1. mole-’ sec.-1 ki. sec. - 1
2 X IO7 670 X 442 671 470
5 X 107
440 kp X 1 0 - p , I . kr kz X 1.51 433 1.98 1.48 665 1.95 1 . 5 4 470 2.33
2 X lo7 310 mole-1 sec.-l k2 X 473 1 . 6 1 657 1 . 5 4 480 1.53
510 Av. k z , 1. mole-1 sec. --I
1.51
2.07
I
---
A,V 3.0X 1076M
l
‘
IO
440 1.7 X 10-6M
IO-~M
I
I
2.0
CHL. & I N PYRIDINE
Fig. 7.--Decay
4831
SPECTRA OF TIIE bfETASTAULE STATES O F ~IILOROPIIYLL.A AND
Sept. 20, 1058
9 X 107 330 X
458 647 500
kz 2.23 2.09 2.26
1.64
1.58
2.19
To obtain appreciable shifts in intercept requires such “high” concentrations that, with mini-
0.6
AD,
kinetics of chlorophyll I3 in benzene; applicatioii of eq. 4.
mal flash energies, one is soon liniited by inconiplete conversion and resulting concentration gradients. Since only small changes in A with Co can be observed, and k2 is much larger than ks, the error in ka is large, the indicated values being little better than order-of-magnitude estimates. However, we believe that the increase in ka in going from pyridine to benzene is real. In obtaining kl, most reliance for the extrapolation was placed on the line of smallest slope (Figs. 7, s). The convergence of the other lines is good, however, and we estimate that the error in the k l ’ s is probably not worse than 20%. In the case of chlorophyll B in benzene, a rather large correction for the klCo term is involved, and the error in kl may be somewhat greater. The effects of strongly quenching impurities, most seriously oxygen, will be to increase the observed kl, so that the values given may, if anything, be too high. The various values of k2 check well for each solution, inasmuch as they include also errors in measurement of uh for the different wave lengths. The relative values of the kz’s are probably reliable to *lo%,, and the absolute error should not be much larger. Previous measurements on triplet-state decay in fluid solvents have been of too low precision to establish clearly the higher-order terms in the rate law.5-1ut15 These terms are not unique to chlorophyll.’O Similar kinetics, in particular similar values for kz, are found in the decay of triplet states of anthracene and porphyrins,2o The rate constant kl of course measures, in principle, the sum of radiative and radiationless transition rates. However, even in rigid media a t low temperature, the phosphorescence of chlorophyll B is weak and that of chlorophyll A is doubtfuL2I It is thus most likely that in fluid pyridine
*
(19) Private communications from Profs. G. Porter a n d R. Livingston indicate t h a t recent independent measurements in their respective laboratories give results in general agreement with those presented here. (20) H. Linschitz and L. Pekkarinen. in preparation. (21) R . S. Becker a n d M. Kasha, TUISJ O U R N A L , 77, 36139 (1955).
4832
HENRY LINSCH~TZ AND KYOSTI SARKANEN
Vol. 80
or benzene, the triplet-singlet conversion is essen- of ions and electroris.'" At high excitation densitially radiationless, and we assume that k I refers to ties, along the track of a particle, mutual quenching this latter process. by excited states may be expected, and consequent Making rough estimates of the collision fre- decrease in chemical yield. Similar effects may also quency in solution, kz corresponds to a quenching apply to the interaction of two excited singlets. efficiency per collision of about 0.1, an excited However, in ordinary photochemical situations, molecule being much more effective than an un- under steady illumination in solutions containing excited one in inducing transitions to the ground reasonably reactive substrates, the k r process will state. The slight increase in k 2 in going from pyr- play no serious role. idine t o benzene can be most simply explained by &?th regard to ka, our data do not permit any the slightly lower viscosity of benzene. Kinet- thorough comparison of the quenching reactions ically, one cannot distinguish between mutual between ground state molecules and excited triplets quenching of both metastable molecules in collision or singlets, respectively. Thus, while the kinetics or the quenching of only one. If the excited states oi' fluorescence self-quenching have been studied are triplets, mutual quenching would appear rea- with good precision,'j we cannot, in the case of k:+, sonable on the basis of the possibility of over-all adec1u:~telydistinguish between bimolecular collision spin-conservation in the triplet-singlet reaction. processes and morc complex interactions involving Alternatively, the high quenching efficiency of a possible long-range energy exchange. Howcollision with a triplet compared to that with a ever, the order of magnitude of ka raises certain singlet might be attributed t o the paramagnetism problems in connection with the function of of the quencher. T o check this point, lifetime chlorophyll in photosynthesis. Even a relatively measurements have been made on metastable small k : i , combined with the extremely high local chlorophyll and anthracene in solutions containing chlorophyll concentrations in aiuo, will mean rapid paramagnetic metal ions.22 While strong quench- quenching. This of course has long been appreing is found for some ions (Ni'+, C o + r , Cu++), ciated for excited singlets, but the problem now others show very much weaker interactions (Mn+-, appears for triplets as well. The explanation of Xd+++). I n addition, measurements on quench- fluorescence self-quenching as an enhancement of ing rate constants by Fujimori and Livingston?" singlet--triplet conversion offers no solution since show that various non-paramagnetic substances the resulting triplet will itself be quenched. If o11e interact with the triplet as strongly as does oxygen. cares to extrapolate from these dilute solution It is evident that other factors than paramagnet- measurements to highly condensed chlorophyll, a isin must be invoked t o explain these results, and lifetime of about 10- sec. may be estimated for the i t is likely that different mechanisms will apply to triplet. These tneasurenients thus further e n different specific situations. In the particular case phasize the long-standing problem of how sclfof quenching by excited states, one may suggest, quenching is avoided in the chloroplast. Othcr for example, in addition t o the mechanisnis al- mechanisms for efficient eiiergy-transfer presuinready cited, a reversible dismutation reaction ably must he involvecl in the pigment "crystal." leading back to the ground state, or a collision of Acknowledgment.-~ I t is a pleasure to thank our the second kind, in which one excited molecule colleague, Dr. E. IV.Abrahamson, for his assistance passes t o the ground state while the other is raised in the initial stages of this research. t o a still higher excited state. Under certain conditions, the large value of k . ~ may have important consequences, apart from its ('21) H 1,inichitz 11 G. Berry and L). Schwritzer, TIIISJ O U R N A L , theoretical interest. For example, in radiation (1'35.1) chemistry, high yields of triplet states are expected, 7 6(, 25833 3 ) W. F. \Vatson a n d 12. Livingston, J . Chef%. P h y s . , 18, 80% due either t o direct excitation or to recombination ( l ! l , j O ) . ~
Lliih ( 2 2 ) A more detailed reixirt ( 2 3 ) 11. I'ujimori a n d I
