Spectrochimica Acta Part A 57 (2001) 1781 1791 www.elsevier.

com/locate/saa

Raman and FTIR spectroscopies of uorescein in solutions

Lili Wang *, A. Roitberg 1, C. Meuse, A.K. Gaigalas
Biotechnology Di6ision, National Institute of Standards and Technology, 100 Bureau Dri6e, Stop 8312, Gaithersburg, MD 20899 -8312, USA Received 29 December 2000; accepted 13 January 2001

Abstract Raman and Fourier transform-infra red (FT-IR) spectroscopies of uorescein in aqueous solutions have been investigated in the pH range from 9.1 to 5.4. At pH 9.1 uorescein is in the dianion form. At pH 5.4, uorescein is a mixture of monoanion ( 85%), dianion and neutral forms (together 15%). The uorescence quantum yield drops from 0.93 for the dianion form to 0.37 for the monoanion form. The Raman and FT-IR studies focused on the frequency range from 1000 to 1800 cm 1 which contains the skeletal vibrational modes of the xanthene moiety of uorescein. At pH 9.1, the spectroscopic feature of uorescein dianion are consistent with a picture of an electron delocalized among the xanthene moiety and two identical oxygens attached to opposite ends of the xanthene moiety, forming a very symmetric structure. The characteristic of uorescein dianion is the presence of the phenoxide-like stretch at 1310 cm 1. At pH 5.4, uorescein monoanion has lost the symmetric structure characteristic of the dianion. The spectra of the monoanion have distinctive contributions from the phenolic bend at 1184 cm 1. The assignments of the vibrational bands shown in Raman and FT-IR spectra are given based on both literature and the ab initio calculations at the HartreeFock level with HF/6-31 + +G* basis set. Excellent correlation is found between the experimental and calculated spectra. 2001 Elsevier Science B.V. All rights reserved.
Keywords: Fluorescein dianion and monoanion; Raman; FT-IR; Ab initio calculations; Fluorescence quantum yield

1. Introduction Fluorescein is the most widely used extrinsic uorescence probe in bioscience [1]. The popularity is associated with its long existing history, high extinction coefcient, large uorescent quantum
* Corresponding author. Tel.: + 1-301-9752447; fax: + 1301-9755449. E -mail address: lili.wang@nist.gov (L. Wang). 1 Present address: Quantum Theory Project, University of Florida, P.O. Box 118435, Gainesville, FL 32611-8435, USA.

yield, and well-developed conjugation chemistry to biomolecules. The drawbacks include its susceptibility to photobleaching, and pH-sensitive uorescence, although advantages may be taken of the later feature to probe local pH values [2]. Surprisingly, only a few articles have dealt with the fundamental aspects of uorescein. The thermodynamic and photophysical properties for seven protolytic forms of uorescein were obtained for pH values ranging from 0 to 10 by two different methods [3,4]. When the solution pH drops from 9.0 to 5.4, for instance, the protolytic

1386-1425/01/$ - see front matter 2001 Elsevier Science B.V. All rights reserved. PII: S 1 3 8 6 - 1 4 2 5 ( 0 1 ) 0 0 4 0 8 - 5

1782

L. Wang et al. / Spectrochimica Acta Part A 57 (2001) 1781 1791

form of uorescein changes from pure dianion to mainly monoanion form ( 85%). The uorescence quantum yield decreases from 0.93 for dianion to 0.37 for monoanion. Similar drop in quantum yield is often observed when uorescein is conjugated to biomolecules. When uorescein is attached to biomolecules in aqueous environment, the uorophore experiences mostly an amphiphlic environment. Sawyer and co-workers investigated the effect of mixing organic solvent with water on the spectral properties for uorescein dianion [5]. It has been found that the spectral shifts in absorption and emission maxima reect both general solvent effect through solvent-water hydrogen bonding and specic effect due to solvent-uorescein hydrogen bonding. Their work is the expansion of the original investigation made by Martin in 70s [6] and is signicant in that it incorporates water, the most important biological medium. In the present study, we use Raman and Fourier transform-infra red (FT-IR) spectroscopies to probe structural differences in uorescein dianion and monoanion forms. The changes of the vibrational bands were monitored as a function of solution pH. We suggested assignments to the observed vibrational modes based on both literature and the ab initio calculations. A few articles dealing with IR spectra of solid uorescein and its derivatives serve as important reference for this work [7 9]. The original motivation for this work was the intuition that the monoanion and dianion forms of uorescein possess very different symmetry, which will be reected in the vibrational spectra taken at different solution pH. According to [3,4], the quantum efciency of uorescein monoanion is 0.37, which is much lower than that of the dianion (0.93). The competing non-radiative decay of uorescein is through internal conversion processes to the ground state and depends on the vibronic coupling between the excited and ground states. The vibronic matrix elements have selection rules, which constrain many of them to zero for highly symmetric molecules. The understanding of the vibrational modes of both ground and excited states is essential for understanding of the internal conversion processes. Moreover, when uorescein is attached to proteins the interaction between

uorescein and amino acids, such as tryptophan, tyrosine, arginine, and histidine, may cause changes in molecular symmetry of uorescein similar to that induced by the protonation described above. Voss, Jr. and his coworkers have investigated such interaction intensively in the model system of uorescein and its monoclonal antibody 4-4-20 [10 12]. Hence, the current work is also relevant to understanding why conjugation of uorescein to proteins leads to signicant decrease in its quantum efciency at the molecular level.

2. Methods

2.1. Experimental details

Fluorescein powder (acid form) was obtained from Molecular Probes and used as received2. Boric acid and sodium hydroxide were analytical reagents from Mallinckrodt. Sulfuric acid (99.999%) and hydrochloric acid (99.999%) were both purchased from Aldrich. Water was deionized with a Milli-Q purication system. The borate buffer solution (0.1 M) was adjusted to a nal pH 9.1 using sodium hydroxide. For Raman measurements, uorescein was rst dissolved in borate buffer solution to make up a nal concentration of 0.8 mM. Sulfuric acid was then used to adjust the solution pH as needed. For FTIR measurements, hydrochloric acid was used for pH adjustments in order to minimize interference from HSO 4 species. Raman measurements were carried out in uorimeter quartz cuvette setting using the 633 nm ( 30 mW) He Ne laser as the light source. The signal was detected at the right angle by combination of a monochromator (Model 270M from SPEX) and a thermoelectrically cooled CCD detector (Spectrum One from SPEX). The nal spectrum was obtained by subtracting the contriCertain commercial equipment, instrument, and materials are identied in this paper to specify adequately the experimental procedure. In no case does such identication imply recommendation and endorsement by the National Institute of Standards and Technology, nor does it imply that the material or equipment is necessarily the best available for the purpose.
2

L. Wang et al. / Spectrochimica Acta Part A 57 (2001) 1781 1791

1783

bution of the medium at the same pH from the measured Raman spectrum of uorescein. The spectral wavenumber was calibrated using benzonitrile, a consensus Raman shift standard for spectrometer calibration. The infrared experimental setup is based on a Bruker Equinox 55 spectrometer (Billerica, MA) with an MCT detector. The internal reection accessory utilized Harrick Scientic Corporation (Ossining, NY) Fast IR 45 reection optics, and a single reection Basics Fresnel trough mounted AMTIR optics card (International Crystal Laboratories, Gareld, NJ) with an infrared refractive index of 2.5. Data collection was automated to allow 5 min of dry air purging (FTIR purge gas generator, Whatman, Haverhill, MA) before two separate acquisitions, each 10 min long, were run at 4 cm 1 resolution using the Opus software of the Bruker Equinox 55. These separate scans allowed the acquisition of water vapor spectra during both the sample and background acquisitions. These water vapor spectra were averaged and subtracted from the average sample spectrum to minimize water vapor contributions to the spectrum. The data work-up was performed in the Bruker Opus software and water vapor subtraction was performed in Grams/32 version 5.01 (Galactic Industries, Salem, NH).

basis set [13]. All calculations were performed using Gaussian 98 [14] along with a implicit solvation model based on a self-consistent reaction eld (SCRF) approximation (Onsager model) [15,16]. The computations were carried out for uorescein dianion (charge = 2) and monoanion (charge = 1). The geometries were locally optimized and frequencies were computed at the same level of the theory on the optimized coordinates. Infrared and Raman intensities were computed as the derivative of the dipole moment with respect to a normal mode change and the derivative of the polarizability with respect to a normal mode change, respectively. The modes with larger IR or Raman intensity (above 10% of the maximum) in the same region as the experimental measurements were inspected using the program Molden [17] and compared with the experimental patterns.

3. Results The protolytic reaction involving uorescein dianion and monoanion is shown in Scheme 1. The reaction is governed by pKa of the OH group attached to xanthene moiety (6.43). At pH 9.1, uorescein exists purely in the form of dianion and at pH 5.4, uorescein monoanion form is dominant by 85% with dianion and neutral forms accounting for 15% [3,4]. Although uorescein monoanion does not exist purely in aqueous solutions, its characteristics will domi-

2.2. Computational details

Quantum ab initio calculations were done at the Hartree Fock level using the HF/6-31 + +G*

Scheme 1. A diagram of the chemical changes which take place during the protolytic reaction between dianion and monoanion forms of uorescein. The reaction is driven by changes in pH.

1784

L. Wang et al. / Spectrochimica Acta Part A 57 (2001) 1781 1791

Fig. 1. (a) Raman and (b) FTIR spectra of uorescein taken at different pHs. The buffer contribution has been subtracted. The uncertainties of the measurements are 1.5% for Raman and 5.0% for FTIR, respectively.

nate at pH 5.4. At pH 9.1, the uorescein molecule is expected to be very symmetric due to the presence of two identical oxygens on the xanthene moiety. To emphasize changes in molecular symmetry during the protolytic reaction, the Raman and FTIR studies focus on the frequency range from 1000 to 1800 cm 1 which contains the skeletal vibrational modes of the xanthene moiety of uorescein. The torsional motions between the xanthene ring and benzoate are below 1000 cm 1. The Raman spectra acquired in aqueous solutions at two different pHs are displayed in Fig. 1a. With the decrease of the solution pH, systematic spectral changes may be seen. The intensities of the vibrational bands at 1170, 1311, and 1499 cm 1 decrease signicantly and those of the bands at 1184, 1330, and 1414 cm 1 increase when pH changes from 9.1 to 5.4. The variations in the band intensities reect molecular structure change, which will be discussed later. The band frequencies and their tentative assignments based on literature are given in Table 1. These vibra-

tional frequencies were obtained through tting using PeakFit version 4.0 from SPSS Inc. with FWHM of 20 cm 1. Raman intensities of active vibrational modes are sensitive to the p-electron system of uorescein, which includes two oxygens and the xanthene moiety. The FTIR spectra have contributions both from the xanthene moiety including two oxygens and the benzoate moiety (Scheme 1). The IR spectra of uorescein in aqueous solutions at three different pHs are given in Fig. 1b. As the pH of the solution decreases, intensities of the bands at 1310 and 1326 cm 1 drop considerably. The shift in band frequency also occurs, for example, from 1171 cm 1 at pH 9.1 to 1183 cm 1 at pH 5.4. In addition, three vibrational bands at 1116, 1215, and 1393 cm 1 are clearly shown in the spectra, but they are absent in Raman spectra (Fig. 1a). The signal-tonoise ratio for the data at pH 5.4 is somewhat smaller than those at higher pHs. We believe this is due to the smaller solubility of uorescein at

L. Wang et al. / Spectrochimica Acta Part A 57 (2001) 1781 1791

1785

low pH as well as poorer subtraction of water contribution. Fluorescein monoanion is less hydrophilic and attracts fewer water molecules to the surface than the dianion. To illustrate a consistent spectral change, the spectrum taken at pH 5.7 was included. The band frequencies and their assignments for FTIR spectra are also summarized in Table 1. To perform the ab initio calculations, we rst focused on the geometry of both forms of uorescein. Fig. 2 presents a diagram of the local minima obtained from the calculation. The geometry

of uorescein dianion at pH 9.1 has D2h symmetry (this symmetry was not imposed on the calculations). The xanthene moiety including the two oxygens is fully symmetric and planar. The bond lengths and angles are the same with regard to the symmetry plane. The two carbonyls at both ends of the xanthene ring are neither single nor double C O bonds. The benzoate moiety is fully planar and perpendicular to the xanthene ring. By comparison, uorescein monoanion has lost the symmetry with respect to the dianion. One side of the xanthene moiety now has a pure carbonyl group

Table 1 The vibrational bands (in cm1) observed in Raman and FTIR spectra of uorescein in aqueous solutions at two pHs and their assignments based on both literature and ab initio calculations pH 9.1 FTIRa 1115 1171 1170 1183 1215 1310 1330 1393 1311 1330 sh 1326 1393 1327 1215 1184 C OH (phenolic) [9,18] C O C stretch of XR [7] Phenoxide ion stretch conjugated with XR stretch [7] XR C C stretch [20] (a) Aromatic skeletal C C stretch [8], (b) symmetric COO stretch [7] (a) XR skeletal C C stretch conjugated with symmetric COO stretch [8], (b) symmetric COO stretch [9] Ramana pH 5.4 FTIRa 1116 Ramana Aromatic C H in plane bend [9,21] Asymmetric CCH bend, some C O (XR) stretch, minor C C stretch CCH bend (H26, H28, H29, H30) CCH (H26, H28, H29, H30) and C OH bend CCH bend, COC stretch and bend (H26, H28, H29, H30) CCH bend, CC stretch and C O (phenoxide) stretch CC stretch (XR stretch) Symmetric COO stretch CCH bend and CC stretch CO and CC stretch (quinone-like) Assignmentsb,c [literature] Assignmentsb,d (theory)

1460

1463 sh

1464

1414 1465 sh

1497 1544 w

1499 1546

1494 w

1497 XR C C stretch [19,20]

1573

1575 sh

1557 w 1584

1556

1596 1636 1636

XR C C stretch [19,20] (a) XR skeletal C C stretch containing conjugated carbonyl band and asymmetric carboxylate stretch [8], (b) asymmetric COO stretch [7,9] XR skeletal C C stretch containing conjugated carbonyl band [8] XR C C stretch [20]

Central ring breathing, CC stretch (C12 C13, C5 C6) CC stretch (C13 C14, C9 C11, C3 C4, C1 C6) C O and CC stretch (C7 C8, C1 C6) Asymmetric COO stretch

CC stretch (C13 C14, C2 C3, C5 C6) Quinone-like stretch, very symmetric CO and CC stretch

w and sh refer to weak and shoulder, respectively. XR is abbreviation of the xanthene ring. c The interpretations of some vibrational bands were somewhat different in various literature. We labeled them as (a) and (b). d Atom numbering is given in Scheme 2.
b

1786

L. Wang et al. / Spectrochimica Acta Part A 57 (2001) 1781 1791

Fig. 2. Geometries of uorescein dianion and monoanion forms at the local energy minima obtained from the ab initio calculations. The dotted line in the dianion structure denotes the presence of molecular orbitals, which are delocalized over the three-ring xanthene moiety. The angle shown below each structure is the angle between the plane of the xanthene moiety and the plane of the benzoate ring.

(C O bond) while the other has a single C O bond (phenolic). The ring with the OH group attached to is fully aromatic whereas the other ring adopts a quinone form. As a result, the benzoate moiety has a 70 dihedral angle with respect to the xanthene moiety, leaning towards the phenolic ring. Under the geometries of the two uorescein forms described above, the frequencies of the normal modes were computed. In agreement with the experimental measurements, the modes between 1000 and 2000 cm 1 were selected, based on IR and Raman intensity criterion (above 10% of the maximum), as candidates for assignment and comparison with the experimental values. Of these, we focus on modes that change signicantly upon pH changes and/or whose assignments are not found in literature. Table 2 gives the frequencies of the assigned modes obtained directly from the theory along with the actual experimental frequencies for both pHs. A linear regression of these frequencies yields an excellent correlation between experiment and theory. At pH 9.1, we found exp = theory 0.935 68.1 cm 1 (R = 0.998) and at pH 5.4, exp = theory 0.916 28.1

cm 1 (R = 0.998). The parameters in the above correlations reect the limitations of the current theoretical models. The experimental (line graph) and theoretical (column) spectra are shown together in Fig. 3. Good correlations can be seen in both IR and Raman spectra, and at both pHs.
Table 2 Comparison of the vibrational modes (in cm1) observed in Raman and FTIR spectra and obtained directly from ab initio calculations for uorescein in aqueous solutions at two pHs pH 9.1 exp 1115 1170 1215 1310 1330 1393 1460 1497 1544 1573 1635 theory 1256 1337 1366 1486 1494 1551 1632 1674 1737 1755 1810 pH 5.4 exp 1116 1184 1215 1326 1393 1414 1464 1494 1557 1584 1596 1636 theory 1240 1332 1363 1487 1551 1565 1615 1675 1710 1755 1777 1830

L. Wang et al. / Spectrochimica Acta Part A 57 (2001) 1781 1791

1787

Fig. 3. Experimental (line graph) and theoretical (vertical lines) spectra of uorescein, (a) Raman at pH 9.1; (b) FTIR at pH 9.1; (c) Raman at pH 5.4; (d) FTIR at pH 5.4. The theoretical bands were shifted using the procedure discussed in the text.

4. Discussion The assignments for the individual modes obtained from the theory are included in Table 1 with some descriptions of the vibrations by atom numbering (Scheme 2). For the purpose of visualization, as well as providing a basis for the notation in Table 1 we discuss in detail a small number of modes, some of which are displayed in Fig. 4. When solution pH drops from 9.1 to 5.4, the protolytic form of uorescein changes from pure dianion to mainly monoanion ( 85%) [3,4]. The structure of uorescein changes from symmetric (D2h symmetry) xanthene ring conjugated with two identical oxygens to less symmetric xan-

thene moiety with a hydroxyl group attached to one side and a pure carbonyl group on the other side. The corresponding calculated structures are shown in Fig. 2. The change in the ionic state is accompanied by a redistribution of one electron charge among the nuclear centers comprising the xanthene moiety. The charge redistribution results in changes of the force constants, which describe the coupling of the various nuclei. Thus we expect to see frequency changes upon change of ionic state as well as appearance of new modes associated with the new hydrogen atom. The change in molecular symmetry will lead to changes in the intensities of the bands in the observed vibrational spectrum.

1788

L. Wang et al. / Spectrochimica Acta Part A 57 (2001) 1781 1791

The vibrational band at 1184 cm 1 is present only at pH 5.4, and the band at 1171 cm 1 is present only at pH 9.1. This behavior is observed in both the Raman and FTIR spectra. The molecular calculations show that the bands at 1184 cm 1 at pH 5.4 and at 1171 cm 1 at pH 9.1 involve considerably CCH bends on the xanthene ring (H26, H28, H29, and H30, Table 1 and Fig. 4). We suggests that the addition of the C OH bend at low pH results in the shift in frequency from 1171 cm 1 at pH 9.1 to 1184 cm 1 at pH 5.4. This assignment is consistent with those in the literature, which stress only the contribution from phenolic vibration [9,18]. In the report of Davies and Jones, the band at 1179 cm 1 was related to the absorption of hydroxyl group because this band was absent in the IR spectrum of sodium phenoxide but present in the IR spectrum of phenol [7]. Similar arguments can be made for the band at 1326 cm 1 which is present at pH 5.4 and the band at 1330 cm 1 which is present at pH 9.1. Both bands most likely result from xanthene CC stretch. The interpretation of the data in the region around 1330 cm 1 is complicated by the presence of a intense band at 1310 cm 1 present only at high pH 9.1 (Fig. 1a). Davies and Jones indicated that this band was related to phenoxide ion stretching conjugated with xanthene ring vibration [7]. The tentative assignment is consistent with the phenomena shown in Raman and FTIR spectra (Fig. 1) in that this band

Fig. 4. Vibrational patterns of the modes, 1170 and 1636 cm 1 for uorescein dianion; 1184, 1414, and 1636 cm 1 for monoanion. The arrows represent the relative displacement of the nuclei for that specic vibrational mode.

Scheme 2. The convention for the numbering of atoms of uorescein molecule. The numbers are used to identify specic nuclei in the theoretical descriptions of the vibrational modes.

gradually disappears when the solution pH drops from 9.1 to 5.4. In the process, the percentage of uorescein dianion decreases signicantly. The ab initio calculations show that one negative charge is delocalized at the xanthene moiety including the two oxygens and the band at 1310 cm 1 results from both CCH bend and stretch of the delocalized xanthene moiety as a whole. The presence of a small peak at 1310 cm 1 at pH 5.4 is consistent with less than 15% dianion form which has a intense band at 1310 cm 1. We next consider the bands at 1544 cm 1 for pH 9.1 and at 1557 cm 1 for pH 5.4, which we suggest result from similar vibrational modes of the xanthene moiety. The different external groups, such as oxygens and hydroxyl group, might cause the difference in vibration frequencies. This postulation is in part supported by the work of Majoube and Henry who studied FT-Ra-

L. Wang et al. / Spectrochimica Acta Part A 57 (2001) 1781 1791

1789

man, FTIR, and surface-enhanced Raman of rhodamine 6G [19]. Fluorescence of Rhodamine 6G results from the delocalized p-electrons among xanthene ring and ethylamino external groups as compared with xanthene ring and two oxygens in the case of uorescein. Hence, one expects the pattern of the xanthene ring modes to be similar in Raman spectra of both uorescein and rhodamine 6G. The shift in mode frequency may reect the difference in the external groups. The authors reported xanthene ring modes of rhodamine 6G in the similar frequency range (at 1363, 1569, and 1647 cm 1 in FT-Raman). Comparing their results with ours, it indeed suggests that different external groups would give rise to the difference in the vibrational frequency of the xanthene ring mode. Additionally, the ab initio calculations show that the band at 1544 cm 1 for pH 9.1 is due to xanthene CC stretch and the band at 1557 cm 1 for pH 5.4 involves both C O and CC stretch (Table 1). Hildebrandt and Stockburger reported the three stretching vibrations of xanthene ring at 1324, 1548, and 1629 cm 1 by means of Surface Enhanced Resonance Raman measurements of uorescein monoanion adsorbed on colloidal silver [20]. The mode at 1414 cm 1 is present only at pH 5.4, its intensity was equivalent to the noise level at pH 9.1. Our calculations illustrate that the band is associated with uorescein monoanion and is due to CCH bends and CC stretch of the xanthene moiety. The mode is neither symmetric nor anti-symmetric (Fig. 4) and therefore can not have an analog in the dianion form which is highly symmetric. Since no reference has been found about this mode, we have assigned it to CCH bend and CC stretch on the xanthene moiety based solely on the theoretical calculations. In contrast to modes discussed above, the band at 1636 cm 1 remains the same for both mono and dianion forms of uorescein. As shown in Fig. 4, the modes at 1636 cm 1 are highly symmetric for both uorescein forms and hence show reasonable Raman intensities. Two additional vibrational bands at 1115 and 1215 cm 1 are only present in IR spectra. The intensities of the two bands are constant in the entire pH range chosen for FTIR measurements. According to Davies

and Jones, the band at 1215 cm 1 in IR spectra is due to the absorption of the ether linkage, C O C [7]. Since the band is absent in Raman spectra of the two uorescein forms, the vibrational mode ought to involve an asymmetric motion of the xanthene moiety. The similar phenomenon has also been seen for the band at 1115 cm 1, which was assigned to aromatic carbon hydrogen in plane bending [9,21]. The results from the theory are also consistent with the literature assignments (Table 1). The vibration at 1115 cm 1 is mainly due to asymmetric CCH bend and the band at 1215 cm 1 involves COC stretch and bend, and CCH bend as well. Although the IR band involving asymmetric carboxylate stretch at 1580 cm 1 was reported consistently in literature [7 9], the band associated with symmetric carboxylate stretch was in disagreement. In two reports [8,9], the band at 1460 cm 1 was assigned to the symmetric carboxylate stretch. Davies and Jones, however, stated the vibrational band at 1398 cm 1 as the symmetric carboxylate stretch based on the observation of the symmetric carboxylate stretch at 1411 cm 1 in sodium benzoate and at 1418 cm 1 in sodium acetate [7]. The carboxylate frequencies are notorious for being unreliably predicted by theoretical methods, unless a good description of the solvent is used. An Onsager model is simply not enough. We have carried out extensive calculations for a number of model systems (acetate, benzoate) at different levels of theory (AM1, HF/ 6-31G*, HF/6-31 + + G*, and B3LYP/6-31 ++ G*) with an increasing number of water molecules bound to the carboxylate. We see a down shift in the predicted frequency for the asymmetric stretch and a up shift for the symmetric stretch. The values of these frequencies stabilize, in every case, after about ten water molecules are added to the system. At that point, the magnitude of the shift is 50 cm 1. Hence, we have identied the symmetric and asymmetric carboxylate modes through visual inspection by either adding 50 cm 1 (in the case of symmetric stretch) to or subtracting 50 cm 1 (in the case of asymmetric vibration) from the theoretical values. We have found that the carboxylate symmetric stretch is at 1393 cm 1, and the asymmetric vibrations are at

1790

L. Wang et al. / Spectrochimica Acta Part A 57 (2001) 1781 1791

1573 cm 1 for uorescein dianion and at 1584 cm 1 for monoanion, respectively. The result of the asymmetric stretch at 1580 cm 1 is consistent with those reports on IR spectra of solid uorescein [7 9]. The symmetric stretch at 1393 cm 1, on the other hand, is in agreement with various reports on benzoate vibrational spectra [22 24]. We have further performed FTIR and Raman measurements of sodium benzoate and sodium salt of benzoic-carboxy-13C acid in aqueous solutions. In the case of sodium benzoate, the bands at 1388 and 1542 cm 1 are due to the symmetric and asymmetric carboxylate stretch, respectively (the results will be published elsewhere). Interestingly, both symmetric (1393 cm 1) and asymmetric ( 1580 cm 1) carboxylate stretch exhibit high intensities in IR spectra but none or low intensity in Raman spectra in the present study. The theory further shows that both stretches are conned to the benzoate moiety. The molecular geometries for both uorescein dianion and monoanion, shown in Fig. 2, support these assignments in that the benzoate moiety has an either 90 (in the case of dianion) or 70 (in the case of monoanion) angle with respect to the xanthene moiety and is, therefore, an independent entity.

1215 cm 1). The ab initio calculations of uorescein dianion and monoanion show excellent correlation to the experimental spectra. Though challenging, the theoretical work gives not only complete assignments but also detailed vibrational structures. The present study is the rst step in the attempt to understand the internal conversion processes in uorescein. The vibronic matrix elements are products of electronic wave function and the changes in the electronic wave function due to nuclear displacements. The symmetry of the normal modes will be one factor in determining the magnitude of many of the matrix elements.

Acknowledgements The authors are indebted to V. Vilker for the encouragement and support of this work.

References
[1] J.R. Lakowicz, Principles of Fluorescence Spectroscopy, second ed., Kluwer Academic/Plenum Publishers, New York, 1999. [2] R.P. Haugland, in: M.T.Z. Spence (Ed.), Handbook of Fluorescent Probes and Research Chemicals, sixth ed., Molecular Probes Incorporation, Eugene, OR, 1996. [3] R. Sjo back, J. Nygren, M. Kubista, Spectrochim. Acta Part A 51 (1995) L7. [4] N. Klonis, W.H. Sawyer, J. Fluoresc. 6 (1996) 147. [5] N. Klonis, A.H.A. Clayton, E.W. Voss Jr, W.H. Sawyer, Photochem. Photobiol. 67 (1998) 500. [6] M.M. Martin, Chem. Phys. Lett. 35 (1975) 105. [7] M. Davies, R.L. Jones, J. Chem. Soc. (1954) 120. [8] R. Markuszewski, H. Diehl, Talanta 27 (1980) 937. [9] I.M. Issa, R.M. Issa, Y.M. Temerk, M.M. Ghoneim, Egypt. J. Chem. 17 (1974) 391. [10] L.K. Denzin, M. Whitlow, E.W. Voss Jr, J. Biol. Chem. 266 (1991) 14095. [11] L.K. Denzin, G.A. Gulliver, E.W. Voss Jr, Mol. Immun. 30 (1993) 1331. [12] J.N. Herron, A.H. Terry, S. Johnston, X. He, L.W. Guddat, E.W. Voss Jr, Biophys. J. 67 (1994) 2167. [13] W.J. Hehre, R. Ditcheld, J.A. Pople, J. Chem. Phys. 56 (1972) 2257. [14] Gaussian 98, Revision A.7, M.J. Frisch, G.W. Trucks, H.B. Schlegel, G.E. Scuseria, M.A. Robb, J.R. Cheeseman, V.G. Zakrzewski, J.A. Montgomery, R.E. Stratmann, J.C. Burant, S. Dapprich, J.M. Millam, A.D.

5. Conclusion Raman and FTIR spectra of uorescein in aqueous solutions have been measured at pH ranging from 9.1 to 5.4. In this pH range uorescein changes from pure dianion form at pH 9.1 to mainly monoanion form at pH 5.4 ( 85%). At pH 9.1, the spectroscopic feature of uorescein dianion is that of a highly symmetric structure consisting of a xanthene moiety with two identical oxygens. At pH 5.4, uorescein monoanion is signicantly less symmetric. The systematic investigation of the vibrational structure change in uorescein upon pH change yielded modes whose frequency changes upon protonation (1171 1184, 1330 1326, 1544 1557 cm 1), modes that appeared in only one ionic species (1310 cm 1 dianion, 1414 cm 1 monoanion), and modes which remained constant over the pH variation (1115,

L. Wang et al. / Spectrochimica Acta Part A 57 (2001) 1781 1791 Daniels, K.N. Kudin, M.C. Strain, O. Farkas, J. Tomasi, V. Barone, M. Cossi, R. Cammi, B. Mennucci, C. Pomelli, C. Adamo, S. Clifford, J. Ochterski, G.A. Petersson, P.Y. Ayala, Q. Cui, K. Morokuma, D.K. Malick, A.D. Rabuck, K. Raghavachari, J.B. Foresman, J. Cioslowski, J.V. Ortiz, B.B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaromi, R. Gomperts, R.L. Martin, D.J. Fox, T. Keith, M.A. Al-Laham, C.Y. Peng, A. Nanayakkara, C. Gonzalez, M. Challacombe, P.M.W. Gill, B.G. Johnson, W. Chen, M.W. Wong, J.L. Andres, M. Head-Gordon, E.S. Replogle, J.A. Pople, Guassian, Inc., Pittsburgh PA, 1998. [15] L. Onsager, J. Am. Chem. Soc. 58 (1936) 1486. [16] M.W. Wong, K.B. Wiberg, M.J. Frisch, J. Chem. Phys. 95 (1991) 8991. [17] G. Schaftenaar, J.H. Noordik, J. Comput. Aided Mol. Des. 14 (2000) 123.

1791

[18] D. Lin-Vien, N.B. Colthup, W.G. Fateley, J.G. Grasselli, The Handbook of Infrared and Raman Characteristic Frequencies of Organic Molecules, Academic Press, San Diego, 1991. [19] M. Majoube, M. Henry, Spectrochim. Acta 47A (1991) 1459. [20] P. Hildebrandt, M. Stockburger, J. Raman Spectrosc. 17 (1986) 55. [21] N.B. Colthup, L.H. Daly, S.E. Wiberley, Introduction to Infrared and Raman Spectroscopy, third ed., Academic Press, San Diego, 1990. [22] J.H.S. Green, W. Kynaston, A.S. Lindsey, Spectrochim. Acta 17 (1961) 486. [23] W. Lewandowski, H. Baranska, Vibrational Spectrosc. 2 (1991) 211. [24] Y.J. Kwon, D.H. Son, S.J. Ahn, M.S. Kim, K. Kim, J. Phys. Chem. 98 (1994) 8481.