JOURNAL OF RESEARCH of the National Bureau of Standards-A. Physics and Chemistry Val. BOA, No. 3, May- June 1976 Fluorescence Efficiency of Laser Dyes *

K. H. Drexhage

Research Laboratories, Eastman Kodak Company, Rochester, New York 14650

(April 9, 1976)

The Auorescence effici enc y of xanthene dyes. oxazine dyes, and 7-aminocoumarins is di scussed. Relations with the mol ecular structure are pointed out and dependence on solvent and temperature is explained. Several new Huorescence standards are sugges ted.

Ke y words: Aminocoumarins; carbazin e dyes; deuterium effect; Auorescence quantum yield; laser dyes; mol ecular structure; oxazine dyes; quenching; xanthe ne dyes.

1. Introduction

The recent development of the dye laser [1] 1 has opened up an important new field of applications for organic dyes. This has led to a renewed interest in the theory of nonradiative transitions in dyes and also to the synthesis of new highly fluorescent dyes. In this article we review briefly the relations between fluor· escence and molecular structure in the most important classes of laser dyes: xanthenes, oxazines, and 7·amino· coumarins. Following this discussion specific sug· gestions for improved fluorescence standards are Depending on the end groups, in particular number and type of the substituents the maximum of the made. 2 R, main absorption band falls somewhere in the range 480-580 nm. The transition moment is parallel to the 2. Xanthene Dyes long axis of the molecule [2]. With R '= H and amino end groups the dyes are called pyronins. Due to a con· The chromophore of xanthene dyes has typically venient syntheses with phthalic anhydride many the following structures xanthene dyes have R I = carboxyphenyl and are called fluorescein or rhodamines (fig. 1). The methyl substi· tuents of Rhodamine 6G have practically no influence on the optical properties of the dye except for the dimerization in aqueous solution which we are not concerned with here. The absorption maximum of the main band occurs almost at the same wavelength in pyronins and rhodamines, if the end groups are identical. The absorption band near 350 nm is stronger '" Paper prese nt ed al th e Workshop Seminar 'Standardi za tion in Spectrophotom etry and in rhodamines than in pyronins, and rhodamines show Luminescence Measure me nt s' held at Ih :! National Bureau of S tandards, Gaithersburg, Md .• Nove mber 19-20. 1975. a slightly larger Stokes shift than pyronins. The , Figures in brac kets in di cat e th e literature references at th e end of this paper. chemical stability of rhodamines is generally suo 2 In ord er to describe mat erial s and eXI)crimcnl al proce dures adc(luately. it is occasionally necessa ry to id entify co mm ercial products by manufacturer's name or label. In no in stance perior, as pyronins in alkaline solution are readily does such id entificati on imply endorse ment by th e Nati onal Burea u of Standard s, nor does it imply that th e particular product or equipment is necessaril y the bes t available for that oxidized by dissolved oxygen to form a colorless, purpose . blue fluorescing xanthone.

421

206-267 OL • 76 - 3 V c--.;;-o '?" ~ OH H, ttl dyj/ I / H N // 0 // N/ H/ ' H Rho dam ine 110 Rhodamine 6G Rhoda min e B 510 nm 5 30 nm 554 nm

FIGURE 1. Molecular structure of rhodamine dyes; absorption maximum in ethanol.

Rhodamines like Rhodamine B react to pH variations concentration (except for aqueous solutions, where ag­ in an interesting manner. The carboxyl group is com­ gregation takes place, and for strongly acidic con­ pletely protonated in acidic solution, but dissociates in ditions, where protonation of the amino groups occurs). alkaline solution_ The negative charge has an inductive Because of the symmetry of the rhodamine chromo­ effect on the central carbon atom of the chromophore. phore, there is no dipole moment parallel to its long The maxima of the main absorption band and of the axis. As a consequence, the Stokes shift is small, and fluorescence are shifted to shorter wavelengths, the fluorescence band overlaps strongly the absorption and the extinction coefficient at the absorption maxi­ band (fig. 2) [10]. Also, there is only little variation of ab­ mum is slightly reduced. While these effects are rather sorption and fluorescence maxima with solvent polarity weak in aqueous solution [3], the wavelength shift [4, 5]. Provided there are no heavy-atom substituents, amounts to about 10 nm in polar organic solvents the rate of intersystem crossing from S 1 is very low (methanol, ethanol, etc.) [4, 5, 6, 7]. When Rhodamine in xanthene dyes, which is in accordance with our loop B, which is usually obtained as the hydrochloride, rule [11]. Phosphorescence is weak even in low temper-

ttl + H

is dissolved in ethanol, the relative amounts of the two ature glasses and has a lifetime in the order of 0.1 s, forms of the dye are determined by this acid-base which is short compared with, for instance, acriflavine equilibrium. The dissociation increases on dilution and some aromatic compounds. causing a shift of absorption and fluorescence to The fluorescence of rhodamines is quenched ex­ shorter wavelengths_ This concentration dependence ternally by I - and SCN - , less efficiently by Br­ of the spectra is therefore not due to a monomer­ and Cl- . No quenching was observed by CI04 - and dimer equilibrium, as was claimed in recent papers BF 4 - [4, 5]. The quenching process apparently in­ [8, 9]. In less polar organic solvents (e.g_, acetone) volves a charge transfer from the anion to the excited the unprotonated (zwitterionic) dye undergoes a revers­ dye molecule. It is not a heavy-atom effect. In polar ible conversion to a lactone: solvents like ethanol, Rhodamine 6G-iodide is com­ pletely dissociated at a concentration of 10- 4 molll and no quenching takes place, because the lifetime of the excited state is of the order of a few nano­ seconds, much shorter than the diffusion time the quencher would need to reach an excited dye molecule. In less polar solvents like chloroform the dye salt does not dissociate and its fluorescence is totally quenched. The perchlorate of Rhodamine 6G be­ haves strikingly differently: its fluorescence efficiency is independent [12] of solvent polarity (0.95 based on the value 0.90 for fluorescein). It is, however, reduced by heavy-atom solvents (0.40 in iodomethane), and the Because the conjugation of the chromophore is inter­ fluorescence of this dye is completely quenched by rupted. this is a colorless compound. On addition of nitrobenzene, presumably due to the high electron acid or complex-forming metal ions the chromophore affinity of this solvent. is regenerated. If the carboxyl group is esterified as in There are two structural features that influence the Rhodamine 6G, all these reactions do not take place, rate of internal conversion in xanthene dyes: mo­ and the absorption maximum is independent of pH and bility of the amino groups and hydrogen vibrations. 422

__I ------~---

15

,- E u f, ,- , \ Q) , \ (5 10 , \ E , \ , \ , \ .,. , \ \ 'Q \ \ \U \, 5 , ,, \ , ,, \ \ '" 4 00 6 00 700 Wavelength

F'GU HE 2 . Rhodamine 6C in ethanol.

Ahsorption s pectrum (E mo lar decadic extinction coeffi cient): (IU antum s pectrum or flu orescence (arbit rary units).

We found that the fluorescence e ffi ciency of rhoda­ dyes, it is noti ceable for instance in Rhodamine 6C. mines that carry two alkyl substituents at each nitro­ On solution in O-d euterated ethanol, H is exchanged gen, e.g. , Rhodamine B, vari es strongly with solvent for D in the amino groups of the dye. With the greater and tem perature [4 , 5, 11]. We ascribe this to mobility mass of deuterium th e nonradiative decay is less ]jkely of the amino groups. If the amino groups are rigidized and the fluorescence e ffi ciency increases from 0.95 as in the new dye Rhoda mine 101 (A abs=577 nm in to 0.99. ethanol) the quantum effi ciency is practically unity,

3. Oxazine Dyes When the central carbon atom in the xanthene chromophore is replaced by nitrogen, the chromophore Rhodamine 101 of an oxazine dye is obtained. The central N-atom acts as a sink for the 7T-electrons, causing a wavelength shift of about 80 nm to the red. As within the xanthene class the absorption of oxazines also shifts to the red with increasing alkylation of the amino groups (fi g. 3). Apart from the pronounced red shift of the 50 - 5 I transition there is little difference between the elec­ tronic transitions of oxazines and rhodamines [2J. The electron-withdrawing effect of the central N­ independent of solvent and temperature. It is inter· atom causes the amino groups to be more acidic. Thus esting that the amino groups are not mobile when they addition of a little base to a solution of Oxazine are less than fully alkylated (Rhodamine 6C, Rhoda­ 4 in ethanol changes the color from blue to red due to mine llO). However, in such dyes there is a proba- deprotonation of one of the amino groups: -- bility that the electronic excitation energy is funneled Oxazines with fully alkylated amino groups do not into N-H vibrations, causing nonradiative decay to undergo any change on addition of base. On the other 50. Although this effect is rather weak in xanthene hand, all oxazines are protonated in strong acids.

423 H3CONXXCH3 H-.JBDNU .& 1.& /H H, (B .& 1.& /H HS CZ,(B DNU.& 1.& / C2H S H/ N 0 N'H NON N O N HsC{ 'CzHS HSC( 'C2HS Oxozlne 118 Oxozine 4 Oxoz ine I 6 45 nm 588 nm 611 nm

~ " N '" ~ /N '" ~::::-N '" H,(BD 1.& /H H, (BD 1 .& /H HsCz--(BO 1 .& / H W,N 0 N"H NON, H C/ N 0 xXN"H HsC{ xX CzH 5 2 xX s Cresyl Viole! Oxozine 170 Nil e Blue 601 nm 6 27 nm 635 nm

FIGURE 3. Molecular structure of oxazine dyes ; absorption maximum in ethanol.

Some oxazine dyes have a structure modified by an As a consequence these compounds show a large added benzo group (fig. 3). This causes a slight red Stokes shift, in particular in polar solvents (fig. 4). shift of absorption and fluorescence. Furthermore, Absorption and fluorescence maxima depend much the shape of the absorption spectrum is different more on solvent polarity than is the case with xan­ than in other oxazine dyes and depends on the temper­ thenes and oxazines [4, 5]. A large number of highly ature [1 3]. These effects are probably caused by steric fluorescent derivatives have been synthesized in re­ interference of the amino group with a hydrogen of the cent years [4,5, 14-18]. Here we can give only a few added benzo group [11]. examples (fig. 5). Absorption and fluorescence of The triplet yield of oxazines is generally very low in 7-aminocoumarins are not influenced by a small accordance with the loop rule_ No phosphorescence has amount of base present in the solution. However, been observed in these dyes. The fluorescence quench­ they react with strong acids in a variety of ways ing processes, discussed for xanthene dyes, are also [4, 5, 11]. The amino group of Coumarin 1, for in­ found with oxazines. Owing to the smaller energy stance, is easily protonated and thus decoupled from difference between 51 and 50, internal conversion plays the chromophore. As a consequence the absorption a greater role in oxazines than in rhodamines. Thus the band at 373 nm disappears. This reaction is inhibited, fluorescence efficiency of Oxazine I-perchlorate is less if the amino group is rigidized as in Coumarin 102. than 0_1 in ethanol, but much higher in dichloromethane In the excited state 51, the keto group is more basic and in 1,2-dichlorobenzene_ Likewise, the effect of than the amino group and thus is protonated if suf­ hydrogen vibrations is much more pronounced. The ficient acid is present so that the diffusion is faster fluorescence efficiency of Oxazine 4 is a factor of 2 than the decay of 51. Frequently a new fluorescence higher in O-deuterated ethanol than in normal ethanol. band appears due to the protonated form. Most coumarins are very soluble in organic solvents, but insoluble in water. However, derivatives have been reported recently that are highly soluble in water 4. 7 -Aminocoumarins (fig. 6) [18].

The most important laser dyes in the blue and green region of the spectrum are coumarin derivatives that have an amino group in 7-position: 5. Fluorescence Standards The optical properties of an ideal fluorescence stand­ ard should be independent of the environment (solvent composition, temperature, etc.). This means, among -- other things, that the compound should not be in­ volved in chemical equilibria. Alkaline solutions are to be avoided, because their alkalinity changes B gradually due to absorption of carbon dioxide. Further­ more the fluorescence should not be quenched by oxygen and the fluorescence efficiency should be in­ The chromophore is not symmetric as in the xanthenes dependent of temperature variations. The commonly and oxazines discussed above. The ground state used fluorescence standards fail badly on one or more can be described by structure A, the excited state of these requirements. A number of better standards 5 I by B. While there is some dipole moment in the can be suggested on the basis of the foregoing ground state, it is much greater in the excited state. discussions. 424 I' 1\ I \ I \ ,. /'\ Coum 102 I \ Coum I E I \ I \ <.> r \ r \ ., I \ .,

r" \ r \ r \ I \ I \ r \ Coum 153 : " Coum 6 I1-' , I I I , I I .,

FIGURE 4a, b, c, and d. 7-A minoco umarins in ethanol.

A bso rption spec trum (IE mol ar decadi c ex ti nction coe ffi cient ); quant um spec trum of fluoresce nce (a rbitrary units).

H" ~ I HSC2'- ~ I N 0 0 N ~ 0 0 W .... N 0 0 ro HSC{ ro Coum . 120 Coum . I coOCoum . 102 354 nm 373 nm 390 nm

'i"0 s,9 (JClN HSC2'- ~ I N ~ 0 0 C(5C'C'OC'"'N ~ 0 0 N 0 0 cy) Hs C{

Cou m. 153 Co um . 314 Cou m. 6 423 nm 436 nm 458 nm

FIGURE 5. Molecular structure of 7-aminocoumarin derivatives; absorption maximum in ethanoL.

425 is superior to qumme bisulfate in every respect. It can be used in almost any solvent except water. Rhodamine 6G-CL04• As was pointed out previously H~"- NA)lO~O in this article, the frequently used standard Rhodamine H C/ B undergoes acid-base reactions that affect its optical 2 "­ S03No properties. Its fluorescence efficiency depends strongly on type of solvent and temperature. There­ Coum. 175 Coum. 386 fore it is not surprising that the quantum yield values 353 nm 388 nm reported in the literature vary considerably. As dis­ cussed above, Rhodamine 6G-perchlorate is superior FIGURE 6. Molecular structure of water-soluble 7-aminocoumarins; absorption maximum in water. to Rhodamine B, because its quantum efficiency has the value 0.95 almost independent of solvent and temperature. The dye chloride is readily available in Coumarin 1. This compound (7-diethylamino-4- rather pure form. It can be further purified by column methylcoumarin) is the most readily available 7-amino­ chromatography on basic alumina with ethanol or coumarin. It is easily purified by recrystallization methanol as the solvent. The perchlorate is insoluble from ethanol. It is highly soluble in most organic in water and is easily prepared by adding HCl04 solvents, but not in water. The fluorescence quantum to an aqueous solution of the dye chloride. It can be efficiency is about 0.5 in ethanol, generally higher in recrystallized from alcohol-water mixtures. The ethyl less polar solvents, and is almost independent of ester perchlorate of Rhodamine 101 has the same temperature. There is little or no quenching by oxygen. useful properties as Rhodamine 6G-perchlorate, while A disadvantage is the possible protonation in acidic absorption and fluorescence are shifted about 50 solvents. If this is a problem, Coumarin 102 or nm to the red. However, this dye is not yet com­ Coumarin 153 may be used. The latter compound is merciallyavailable. particularly interesting because of its large Stokes­ Oxazine 170-CL04 • Of all available oxazines this shift and the very broad fluorescence spectrum. The derivative has the highest fluorescence efficiency fluorescence efficiency of Coumarin 102 was measured ( ~ 0.5 in ethanol). Its fluorescence properties are as 0.6 and that of Coumarin 153 as 0.4 (both in ethanol). closely related to those of Rhodamine 6G-CI04. Another compound, possibly useful as a standard, Absorption and fluorescence are shifted by about is Coumarin 6 (quantum efficiency 0.8 in ethanol), 100 nm to the red (fig. 7). As in the case of Oxazine which absorbs at longer wavelengths than most 4, the tluorescence efiiciency increases nearly a factor other coumarin derivatives. The latter compounds of 2 on deuteration of the amino groups. Apart from are commercially available, but the price is still high. this effect, it is almost independent of solvent and However, this should not be a deterrent, as the price temperature. However, variations of temperature have will certainly come down and generally only a few some effect on the absorption spectrum due to the milligrams are required. We feel that Coumarin 1 annellated benzo group. In this respect Oxazine

10 r------r------,------~------r------,------

[\ 1\ I \ I \ I \ I \ I I I I I I I I \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ ,, '" 400 500 700 sao Wav elength (nm)

FIGURE 7. Ox azine 170 in ethanol.

Absorption spectrum (E molar decadic extinction coefficient ); quantum spectrum of fluorescence (arbitrary units), 426

------4-CI0 4 would be preferable _ As pointed out earlier, operates up to 700 nm. It is much more sensitive than these dyes cannot be used in basic solution_ Being equall y concentrated solutions of methylene blue perchlorates, they are practically insoluble in water. as this has a lower flu orescence efficiency. HexamethyLindodicarb ocyanine (HID C) and Hexa­ The flu orescence spectra and effi cie ncies of the methyLindotricarbo cyanine (HITC). Very few dyes s uggested dyes need to be determined accurately. To are known that are relatively stabl e and flu oresce well us this seems to be of greate r benefit than perpetuat­ in the infrared region of the spectrum. Of th e com­ ing bad sta ndards whose only justification today is me rcially avail abl e materi als, the cyanine dyes their common usage. 1,1' ,3,3,3' ,3 ' -hexamethylindodi carbocyanine (HIDC) iodide and 1,1' ,3,3,3' ,3' -hexameth ylindotricarbocya­ 6. References and Notes nine (HITC) iodide stand out on both counts. Con­ trary to the other compounds [11 Dye Lasers (F. P. Schafer, Ed.) Topics in Applied Physics, Vo l. I (S pringer, Berlin, 1973). [21 J akobi , H. , a nd Kuhn, H., Z. Elektroc hem. Ber. Bunsenges. Phys. Chem. 66,46 (1962). HIDC (n=2) [31 Rametle, R. W. , and Sande ll , E. B., J . Am. Che m. Soc. 78, 4872 (1956). HITC (n=3) [41 Dre xhage, K. H. , VII Inte rn . Quant. Electron. Conf. , Montreal, Canada, paper B.5 (1972). [51 Drexhage , K. H., Laser F'o c us 9 (3),35 (1973). [6) Ferguson, .]. , a nd Mau, A. W. H., Chem. Phys. Lett. 17,543 (1972). suggested here, the flu orescence e ffi cie ncy of these [7] F'erguson , J.. a nd Mau, A. W. H. , Austral. J. Che m. 26,161 7 dyes depends markedly on type of solvent a nd te mper­ (1973). ature. It appears to be highest in sulfoxides (dimethyl [81 Selw yn, J . E., and Steinfeld, J. I. , J. P hys . C hem. 76, 762 (1972). sulfoxide, te tra methylene s ulfoxide). As shown in (9) Wong, M. M. , and Schell y, Z. A. , J. Phys. Chem. 78, 1891 fi gure 8, the flu orescence of RITe exte nds to 900 nm. (1974). The photoc he mical stability is muc h hi gher than in ll OI The absorption of Rhodamin e 6G (2g/1 in ethanol), whose rdated thi acarbocyanines. A concentrated solution maximum is at 530 nm, is still noti ceable at 633 nm, far beyond t.h e flu orescence maximum of 555 nm: a 5 mW He-Ne of HIDC-iodide in dimethyl sulfoxide is currently laser excites ye ll ow a nti-Stokes flu orescence that can be in use in our laboratory as a photon counter that observed vi s ua ll y.

3o,------,------,------,------,------,------,

'E HIDe (\ u 1\ , 20 1\ Q) I \ o I \ E \ \ \ \ "a_ \ 10 \ \ '" , , \ \ ~0~0~~~300~~--~40~0~--~5~00~--~~~--~7~070--~'-8~OO Wavelenglh (nm)

30,------,------,------,------,------r------,------~

'"E HITC /\ u I \ 'Q) 20 I \ (5 I \ I '. E I \ \ r \ \ Q \0 \ \ IV \ \ \ \ \ \ ~0~0~~~3~0~0====~4~00~=-~5~O~0~~~6~0~O----~7'~O~--~8JO~O----~~O Wavelength (nm)

FIGURE 8a. a nd b. HIDC and HITC w ethanoL.

Absorpti on s pec trum (E mo lar decadic extinction coeffi cient): quantum spectrum of flu orescence (arbitrary un it s). 427 [11] Drexhage, K. H., Structure and properties of laser dyes, in: [15] Drexhage, K. H., and Reynolds, C. A., VIII Inte rn. Quant. Dye Lasers (F. P. Schafer, Ed.), Topics in Applied Physics, Electron. Conf., San Francisco, Calif., paper F.l (1974). Vol. 1, p. 144 (Springer, Berlin, 1973). [16] Schimitschek, E. J. , Trias, J. A., Hammond, P. R., and Atkins, [12] This state me nt has to be taken with a grain of salt. We found R. L. Opt. Commun. 11, 352 (1974). indeed variations up to 10 percent with solvent. But the [17] Reynolds, C. A., and Drexhage, K. H., Opt. Commun. 13, accuracy of our quantum effici ency measurements was prob· 222 (1975). ably le ss than this, when solvents with widely varying reo [18] Drexhage, K. H., Erikson, C. R., Hawks, C. H., and Reynolds, fractive index were involved. We feel, however, that the C. A., Opt. Commun. 15, 399 (1975). accuracy of the measurements was sufficient to justify the interpretations given here and elsewhere in this article. [13] Cacoin, P., and Flamant, P., Opt. Commun. 5,351 (1972). [14] Tuccio, S. A., Drexhage, K. H., and Reynolds, C. A., Opt. Commun. 7, 248 (1973). (Paper 80A3-894)

428