Fluorescence Quantum Yields of a Series of Red and Near-Infrared

ARTICLE pubs.acs.org/ac

Fluorescence Quantum Yields of a Series of Red and Near-Infrared Dyes Emitting at 600-1000 nm Knut Rurack* and Monika Spieles Division 1.5 Bioanalytics, BAM Bundesanstalt f€ur Materialforschung und -pr€ufung, Richard-Willst€atter-Strasse 11, D-12489 Berlin, Germany

bS Supporting Information ABSTRACT: The determination of the fluorescence quantum yields (QY, Φf) of a series of fluorescent dyes that span the absorption/excitation and emission ranges of 520-900 and 600-1000 nm is reported. The dyes encompass commercially available rhodamine 101 (Rh-101, Φf = 0.913), cresyl violet (0.578), oxazine 170 (0.579), oxazine 1 (0.141), cryptocyanine (0.012), HITCI (0.283), IR-125 (0.132), IR-140 (0.167), and four noncommercial cyanine dyes with specific spectroscopic features, all of them in dilute ethanol solution. The QYs have been measured relative to the National Institute of Standards and Technology’s standard reference material (SRM) 936a (quinine sulfate, QS) on a traceably characterized fluorometer, employing a chain of transfer standard dyes that include coumarin 102 (Φf = 0.764), coumarin 153 (0.544), and DCM (0.435) as links between QS and Rh-101. The QY of Rh-101 has also been verified in direct measurements against QS using two approaches that rely only on instrument correction. In addition, the effects of temperature and the presence of oxygen on the fluorescence quantum yield of Rh-101 have been assessed.

F

luorescence spectroscopy is an invaluable and ever more popular tool in analytical and bioanalytical chemistry.1-5 This trend is basically fuelled by the wealth of information that is accessible through the measurement of fluorescence parameters.6,7 A fluorescence signal is initially determined by the wavelength of excitation, λex, used in a particular experiment. Once excited, a fluorophore emits its characteristic spectrum with a certain intensity distribution and over a certain period of time during which the fluorophore remains on average in the excited state. This time period is expressed by the fluorescence lifetime, τf, which is characteristic for a given emitter in a given environment. The intensity of the fluorescence depends on the ability of the fluorophore to convert a number of absorbed into emitted photons in a particular environment, i.e., the fluorescence quantum yield Φf. If fluorophores are confined in their environment, either accidentally in a frozen or highly viscous sample or purposefully in a designed assay, the emission of such an ensemble is commonly not isotropic any more, and the fluorescence anisotropy, r, can also be used to obtain information on the state of the system. In addition, detection of the fluorescence excitation spectrum with a fluorescence instrument set at a certain emission wavelength, λem, provides valuable information on the spectral characteristics of the absorbing species. Besides this parametric versatility, the popularity of fluorescence measurements is connected to exceptional instrument sensitivity, offering the possibility to measure in spatially resolving modes and permit remote sensing or measurement applications. Last but not least, a wealth of organic, inorganic, or hybrid fluorophores or luminophores with very diverse and tailored properties has been developed over the past century many of which are commercially available.8-13 r 2011 American Chemical Society

Of all the fluorescence parameters introduced above, the fluorescence intensity is perhaps the quantity that is most often interrogated in (bio)analytical applications. Exploitation of the fluorescence lifetime, which lies characteristically in the lower nanosecond time regime for organic dyes, requires more sophisticated instrumentation, and assessment of the fluorescence anisotropy is restricted to more specialized areas of application. The measurement of changes in the fluorescence spectra also does not come close in prevalence to intensity measurements since it requires the use of a diffractive and/or wavelengthtunable element. When dealing with intensity-based applications, knowledge of the fluorescence quantum yield of a fluorophore is often indispensible. On one hand and in combination with the molar absorption coefficient at the excitation wavelength used, ε(λex), this quantity is a measure for the strength of the fluorescence signal to be expected in a particular application; the product ε(λex) Φf determines the brightness of a fluorophore. On the other hand, the measurement of Φf is an important aid in the development of fluorescent dyes, indicators, tags, or materials. The measurement of fluorescence quantum yields can be carried out in an absolute or relative fashion. The absolute determination of Φf can be performed either directly with optical methods measuring the emitted photons in absolute radiometric units or indirectly with calorimetric methods assessing the amount of energy funneled through nonradiative pathways. Both approaches Received: May 21, 2010 Accepted: December 20, 2010 Published: January 20, 2011 1232

dx.doi.org/10.1021/ac101329h | Anal. Chem. 2011, 83, 1232–1242

Analytical Chemistry

ARTICLE

have certain advantages and disadvantages.14 The main advantage of the direct method is that it offers to obtain both the fluorescence quantum yield and the fluorescence emission spectrum. However, absolute measurements of Φf are an inherently difficult task seldom performed routinely because they require the establishment of a complete photon balance and thus special instrumentation.15,16 Much more widespread is the measurement of relative Φf as introduced by Parker and Rees in 1960.17 Here, an unknown sample is measured against a fluorescence standard, i.e., a fluorophore with a known Φf preferably under identical conditions with a conventional fluorometer. The fluorescence quantum yield is then calculated according to eq 1, taking into account the absorbances Ax(λex) (= εx(λex)cxl; cf., the Beer-Lambert law, eq 2) of both sample (x = i) and standard (x = s) at the respective excitation wavelength, through the corresponding absorption factor fx(λex),18 the integral fluorescence Fx(λem), and, if different solvents are used, their refractive indices nx. R F i ðλem Þ ni 2 i s fs ðλex Þ R λem Φf ¼ Φf ð1Þ fi ðλex Þ λem F s ðλem Þ ns 2 fx ðλex Þ ¼ 1 - 10 - Ax ðλex Þ ¼ 1 - 10 - εx ðλex Þcx l

ð2Þ

Depending on the spectral match between sample and standard, minimum data correction, i.e., usually only the knowledge of the spectral responsivity of the emission channel, s(λem) (see section I, Supporting Information), to obtain properly corrected emission spectra, is necessary, rendering this approach reliable, robust, and popular. The availability of reliable fluorescence quantum yield standards, however, is still an issue of major concern,14 especially in view of the renewed interest in all aspects of quality assurance in fluorescence measurements that was basically triggered by intensified activities in this area at the U.S. National Institute of Standards and Technology (NIST) and Germany’s BAM Federal Institute for Materials Research and Testing since the early 2000s.19 Today’s major trends in the analytical application of fluorescence concern the miniaturization of equipment for sensing and diagnostics applications, microscopy-based techniques like total internal reflection fluorescence (TIRF) microscopy or fluorescence recovery after photobleaching (FRAP) techniques, and scanner-based imaging.20-26 In many of these applications, especially in the areas of imaging and microscopy, the red visible and near-infrared (NIR) spectral range is of particular importance. Whenever biological or medical samples are assessed without any sophisticated cleanup, the most sensitive and unperturbed measurements of fluorescence signals are possible in the so-called “biological window” or “NIR window” between 650 and 900 nm.27 Accordingly, the development of NIR fluorophores for such applications receives strong attention.28-31 However, in sharp contrast to this importance, the number of reliable red and NIR emitting fluorescence quantum yield standards available today is negligible.14,32 The data often lack consistency, predominantly because of insufficient instrument correction. If one compares, for example, the Φf data published for rhodamine B in ethanol, values between 0.7333 and 0.9934 can be found (for a detailed discussion, see ref 14). Symptomatic for instance is also the case of the seminal work of Benson and Kues on fluorescence properties of widely used cyanine dyes who self-critically state in their publication that “the quantum yields should be considered for

general trends rather than for comparing individual dyes”.35 Nonetheless, such data are frequently cited as reference values in the determination of Φf of newly developed red and NIR emitting fluorophores in the literature. The aim of the present work is to provide the relative fluorescence quantum yields of a series of commercially available red and NIR emitting dyes which are potential fluorescence standards in the range of 600-1000 nm (Chart 1A). The standard dye that serves as the primary reference is NIST’s standard reference material SRM 936a (quinine sulfate dihydrate, QSSRM, Chart 2). Although SRM 936a has been certified for its relative molecular emission spectrum and its quantum yield has only been given as an informative value in that document, this value is perhaps the most reliable QY reported today, Φf = 0.544 ( 0.03 in 0.5 M H2SO4.36 The relative determination of Φf is carried out on a fully characterized and traceably spectrally calibrated fluorometer at BAM,37,38 which was also used for the certification of a set of spectral fluorescence standard dyes available from Fluka AG and BAM.39 Traceability here means that the fluorescence instrument is calibrated with radiometric transfer standards through an unbroken chain of comparisons with given uncertainties to primary radiometric standards, guaranteeing lowest possible uncertainties and hence the reliable measurement of relative fluorescence spectra and quantum yields.40 In addition to the commercial dyes listed in Chart 1A, the fluorescence quantum yield data of four complementary noncommercial dyes are reported. Two of these dyes are promising standard dye candidates (HEDITCP and HPDITCP, Chart 1B) because they show comparatively broad and nonstructured bands and a pronounced Stokes shift in conjunction with rather high Φf, a combination that is rarely found yet highly desirable in the red wavelength range (cf., e.g., p 9 of ref 16). The other two dyes (ONITCP and ODNITCP, Chart 1B) are very interesting because they absorb and emit between 850 and 1000 nm, in a wavelength range in which only few dyes with such favorable properties are available. Because the lowest-energy or first absorption bands, which should be used for excitation when determining the fluorescence quantum yield, of the reference dye QS and the shortest wavelength absorbing dye of the series shown in Chart 1, rhodamine 101 (Rh-101), are not overlapping, several other commercial and well-known fluorescent (laser) dyes (Chart 2) have been employed to construct a chain of internal reference standards to bridge that gap.

’ EXPERIMENTAL SECTION Materials. Commercially available dyes41 rhodamine 101

(batch no. 019502), cresyl violet (batch no. 019328), oxazine 170 (batch no. 069313), oxazine 1 (batch no. 090214), HITCI (batch no. 029006), IR-140 (batch no. 100209), IR-125 (batch no. 109707), coumarin 102 (batch no. 079508), coumarin 153 (batch no. 029303), and DCM (batch no. 029401) were purchased from Lambda Physik GmbH, G€ottingen; cryptocyanine (batch no. 13715JZ) was from Sigma-Aldrich, and quinine hemisulfate monohydrate (BioReagent, suitable for fluorescence; batch no. 1347313 V) were from Fluka AG. Quinine sulfate dihydrate (SRM 936a) was obtained from the SRM lot at NIST in 2003 and stored under the recommended conditions (at þ4 °C, in the dark). Repetitive measurements of freshly prepared solutions of SRM 936a and several other certified standard dyes39 on the calibrated instrument at BAM38 guaranteed the chemical integrity of the SRM. HEDITCP and 1233

dx.doi.org/10.1021/ac101329h |Anal. Chem. 2011, 83, 1232–1242


ARTICLE

Chart 1. Chemical Structures of the Red and NIR Dyes Investigateda

a

(A) Commercially available dyes; (B) complementary cyanine derivatives. Counterions: Rh-101, CV, Ox-170, Ox-1, IR-140, HEDITCP, HPDITCP, ONITCP, ODNITCP = perchlorate, Cc, HITCI = iodide, IR-125 = sodium. Synonyms: rhodamine 101 = rhodamine 640, cresyl violet = cresyl violet 670 = oxazine 9, oxazine 170 = oxazine 720, oxazine 1 = oxazine 725 = basic blue 3, cryptocyanine = DCI-4, HITCI = hexacyanine 3 = NK-125; IR-125 = indocyanine green = ICG.

Chart 2. Chemical Structures of the Commercially Available UV-Vis Emitting Dyes Investigateda

a

QS as dihydrate. Synonyms: coumarin 102 = coumarin 480, coumarin 153 = coumarin 540A = C6F.

HPDITCP42 were synthesized in our laboratory as described recently.43 ONITCP and ODNITCP44 were a kind gift of Drs. Y. L. Slominskii and J. L. Bricks, National Academy of

Sciences of the Ukraine, Kiev and synthesized according to ref 45. The purity of all the dyes was checked by HPLC, and no impurities absorbing in the relevant excitation wavelength ranges of the respective dyes were found. Ethanol of UVspectroscopic grade was obtained from Fluka AG and not specifically dried or degassed before use. HClO4 (70% Suprapur) was purchased from Merck, and bidistilled water was provided by BAM’s Primary Calibration Substances; Elemental Trace Analysis working group. Whereas for all the dyes except QS ethanol was used as solvent, both QS samples were measured in 0.105 M HClO4 as recommended by Velapoldi and Mielenz because the quantum yield is independent of acid concentration in HClO4.16 Instruments and Methods: General. Steady-state absorption and fluorescence measurements were carried out on a Cary 5000 UV-vis-NIR spectrophotometer and a Spectronics Instrument SLM 8100 spectrofluorometer. The latter operates in T-optic design with an emission channel for the UV-vis (SLM Aminco MC 400 monochromator, Hamamatsu R928 photomultiplier tube; corrected wavelength range of 300-800 nm) and a second emission channel for the red-vis-NIR spectral range (Acton Research Spectra Pro-500 monochromator, Si avalanche photodiode OEC/EMM-Si-APD-0.5-TE2 from OEC GmbH, Zusmarshausen; corrected wavelength range of 5001100 nm), both traceably characterized with respect to the 1234


Analytical Chemistry primary radiometric standard blackbody radiator, and thus to the spectral radiance scale, hosted at the Physikalisch-Technische Bundesanstalt (PTB) in Berlin. The calibration procedures for both instruments regarding wavelength accuracy and for the fluorometer regarding the wavelength- and polarization-dependent spectral responsivity of the detection channels and the wavelength- and polarization-dependent relative spectral irradiance at sample position, characterizing the spectral source radiance and spectral throughput of the excitation monochromator and excitation optics, have been described in earlier publications, including also a description of all the primary and secondary, i.e., transfer standard calibration tools used.37,38,46 For excitation beyond 700 nm, the excitation monochromator of the SLM 8100 was placed at zeroth order and band-pass filters of 760, 780, and 820 nm (full width at half-maximum (fwhm) = 10 nm) from Thorlabs (FBxxx-10 series) were employed. The spectral transmission of the filters across the absorption spectrum of the respective dyes was additionally recorded with the Cary 5000. For all measurements, the temperature was kept constant at 298 ( 1 K. Fluorescence experiments were performed with a 90° standard geometry, with polarizers set at 54.7° for emission and 0° for excitation. All the fluorescence spectra presented here are corrected. Measurement of Molar Absorption Coefficient. The molar absorption coefficient ε(λex) was determined from three separately weighted and dissolved stock solutions in a repeat determination of each sample (N = 6 independent measurements). The dilution was chosen in such a way that the absorbances of the sample solutions equaled 0.15 ( 0.02 at the maximum of the first absorption band. The quartz cells (Hellma 110-QS, 110-50-40) were sealed with Teflon stops after each filling step. Measurement of Fluorescence Quantum Yield. For fluorescence measurements, only dilute solutions with an absorbance of less than 0.08 at the first absorption maximum were used. The solutions were prepared in 50 mm optical path length quartz cells, and the absorption spectrum was measured as described before. An amount of 3.5 mL of this solution was then transferred to a 10 mm optical path length quartz cell (Hellma 110-QS, 11010-40) and subjected to a fluorescence measurement. Subsequently, the absorption of the sample was measured again to exclude photobleaching effects (for a detailed overview of the photostability of the dyes, see section III, Supporting Information). A blank was recorded after rinsing, cleaning, and refilling the cell with solvent. The Φf values reported were obtained from at least N = 6 independent measurements (N = 12 for UV-vis dyes) and calculated according to eq 1. The absorbances Ax(λex) of both sample (x = i) and standard (x = s) at the respective excitation wavelength (see eq 2) have been assessed according to a formalism proposed by Demas in ref 47 that minimizes the effect of the width of the excitation slit itself on the determination of Φf (eq 3). Ax ðλex Þ ¼

1 ½Ax ðλex Þ þ Ax ðλex þ Δλ=2Þ þ Ax ðλex - Δλ=2Þ 3 ð3Þ

For the measurements employing the band-pass filters in the excitation channel, the filter transmission profile was applied on the absorption spectrum in the excitation region instead. In the case of different solvents used for both sample and standard, the refractive index correction term (nx2) becomes effective in eq 1. Refractive index data for ethanol have been taken from ref 48; the

ARTICLE

value for 0.105 M HClO4 has been determined to 1.3347 with an Abbe refractometer model G from Carl Zeiss, Jena. Temperature-Dependent Measurements. The temperature-dependent measurements were performed with a continuous flow cryostat CF 1204 from Oxford Instruments. Liquid nitrogen was pumped from a storage container via a transfer tube (GFS 300, Oxford Instruments) and flow control (VC 30, Oxford Instruments) by a gas flow pump (GF 2, Oxford Instruments) through the cryostat. The temperature was externally controlled by heating and flow adjustment with the temperature controller ITC 4 from Oxford Instruments. The temperature in the sample rod was monitored via the temperature-dependent resistance of a sensor which was calibrated with a Peltier element. Cooling was performed by gradually decreasing the temperature in steps of 20 K with equilibration times of 20 min at every point of measurement. Three independent runs were performed.

’ RESULTS AND DISCUSSION Three important issues have to be accounted for when attempting to measure an accurate fluorescence quantum yield of a liquid, optically transparent sample against a fluorescence quantum yield standard under similar sample conditions. First, the standard has to be properly chosen. Second, the sample solutions have to be carefully prepared. Third, the measured data have to be adequately corrected. With respect to the first requirement, one commonly has to distinguish between two cases. If correction curves for the emission and the excitation channel of the instrument to be employed are known, the absorption bands of standard and sample do not necessarily have to overlap because two different wavelengths can be used for excitation and later on the differences can be corrected for. If only the emission correction curve of the instrument is known, which is the much more common case for most users, standard and sample should be excited at the same wavelength, and thus, the absorption spectra of the two compounds have to overlap to some degree. (If an emission correction curve is also not available, success is most often questionable, and reliable data can perhaps only be generated for very closely absorbing and emitting compounds such as, for instance, anthracene and certain substituted anthracenes that also show the characteristic anthracene-type spectra.) With respect to sample preparation, the purity of compounds, solvents, and measurement cells used is naturally essential as are issues of sensitivity against oxygen (either degassed or oxygen-saturated solutions), the presence of trace amounts of water (in an organic solvent), acid, or base, and the susceptibility to photobleaching (under the required excitation conditions). Besides these intrinsic and environmentally important prerequisites, the nature of the measurement solutions is important, i.e., these solutions should be optically transparent and sufficiently dilute so that the active volume in the cell is homogeneously illuminated and no reabsorption effects can occur. This is usually achieved by adjusting the absorbance to e0.08 at the maximum of the first absorption band. Since these absorbances are rather low and can entail undesired uncertainties when measuring the absorption spectrum, cells with a longer path length (if possible) should be used for the absorption measurement. If this is not possible due to geometric restrictions of the sample chamber of the available spectrophotometer, one has to assess the potential uncertainties that are introduced by measuring either low absorbances or by diluting the solution between absorption and fluorescence measurement. Finally, concerning instrument calibration and 1235



ARTICLE

Figure 1. Three approaches to determine Φf(Rh-101) with QS as the primary standard. Path A: two different excitation wavelengths (348 and 525 nm) in the first absorption bands of both dyes are employed, taking into account the instrument’s excitation correction curve. Path B: identical excitation conditions at 348 nm are used after verifying that the fluorescence of Rh-101 is independent of the excitation wavelength in the range of 340-570 nm. Path C: a chain of transfer standard dyes is employed with identical excitation conditions for each pair QS/C-102, C-102/C-153, C-153/DCM, and DCM/Rh-101. Color code: spectra of QS = black, C-102 = green, C-153 = violet, DCM = brown, Rh-101 = blue; red arrows denote suitable excitation wavelengths; dashed arrows denote the number of steps between the first pair of dye and final result; graph on the top, right shows the relative spectral absorption and emission features of QS and Rh-101.

correction, this point is essentially important to exclude all instrument-related sources of errors. Since these aspects have been discussed in detail in various recent publications, the reader is referred to the literature for further reading.14-16,37-39,46 Strategy. The strategy followed in this work relied on QS as the primary fluorescence quantum yield standard, and Φf = 0.60 (QSSRM in 0.105 M HClO4)16 was used in all the calculations. In a first set of experiments, the fluorescence quantum yield of Rh101 was measured relative to QSSRM by employing two different excitation wavelengths in the lowest-energy absorption bands of the two dyes with subsequent emission and excitation correction (Figure 1A, 525 nm for Rh-101, 348 nm as recommended for QSSRM).16,36 In a second series of measurements, the independence of Φf of Rh-101 from the excitation wavelength in the range of 340-570 nm was verified (Figure 1B). This allowed for a direct determination of Φf(Rh-101) with QSSRM employing identical excitation conditions with λex = 348 nm. Third, a reference chain using C-102, C-153, and DCM was constructed between QSSRM and Rh-101 and the Φf were determined pairwise: QSSRM was used as standard for C-102, then C-102 as standard for C-153, C-153 for DCM, and finally DCM as standard for Rh-101 (Figure 1C). Since all the pairs show partly overlapping first absorption bands, identical excitation conditions

avoiding excitation into higher states than S1 could be employed. This approach allowed to countercheck the other two measurement series and provided a valuable test for the measurement of the red and NIR dyes of interest. Because of the spectral limitations of the excitation channel of the spectrofluorometer employed (usable only up to λexc ∼ 700 nm);IR-140, IR-125, ONITCP, and ODNITCP had to be excited with the monochromator in zeroth order and a band-pass filter;these dyes could not be referenced directly to Rh-101, but a reference chain had to be constructed. To gain a consistent picture, the fluorescence quantum yields of C-102, C-153, and DCM were also determined against QSSRM according to the first and second procedures.49 Finally, because the fluorescence quantum yields for Rh-101 in ethanol published in the literature vary (at least) between 0.89 and 1.00 at room temperature,50-55 irrespective of absolute or relative methods of determination, the temperature dependence of Φf(Rh-101) was recorded in the range of 298-153 K. From QS to Rh-101. Table 1 contains the data obtained for the first and the third approach. It is apparent that the results agree well, i.e., fluorescence quantum yield determinations relying either on proper excitation correction or on a chain of chemical transfer standards yield similar Φf within experimental error. 1236



ARTICLE

Table 1. Fluorescence Quantum Yields of C-102, C-153, DCM, and Rh-101 in Ethanol at 298 K Φf (λex) (nm)a dye

dye

Φf (λex) (nm)b standard

dye

standard

C-102

0.766 ( 0.041 (365)

QS

, 0.60 (348)

0.762 ( 0.039 (365)

QS

C-153

0.546 ( 0.029 (402)

QSSRM, 0.60 (348)

0.542 ( 0.028 (402)

C-102, 0.762 (402)

DCM

0.437 ( 0.024 (443)

QSSRM, 0.60 (348)

0.433 ( 0.023 (443)

C-153, 0.542 (443)

Rh-101

0.916 ( 0.049 (525)

QSSRM, 0.60 (348)

0.910 ( 0.048 (520)

DCM, 0.433 (520)

SRM

SRM

, 0.60 (365)

Obtained relative to QSSRM (in 0.105 M HClO4, λex = 348 nm, Φf = 0.60 (ref 16)) with different excitation wavelengths as indicated in parentheses and taking into account excitation correction (path A in Figure 1). b Reference chain constructed from QSSRM to Rh-101 with C-102, C-153, and DCM as transfer standards using identical excitation conditions for each pair (path C in Figure 1). a

Table 2. Absorption and Fluorescence Maxima of QSSRM in 0.105 M HClO4, C-102, C-153, DCM, and Rh-101 in Ethanol at 298 K

a

dye

λabs nma

λem nma

QSSRM

347.5

454.9

C-102

388.8

466.0

C-153

422.2

533.0

DCM

467.9

622.9

Rh-101

564.4

587.6

(0.5 nm.

The average Φf of C-102, C-153, DCM, and Rh-101 in ethanol at 298 K amount to 0.764, 0.544, 0.435, and 0.913. The corresponding maxima of the first absorption and the fluorescence bands of the dyes are collected in Table 2. The second series of experiments revealed that (at least) above 340 nm, the absorption and corrected fluorescence excitation spectra of Rh-101 are identical within experimental uncertainty. These findings agree well with earlier results on the chemically closely related dye rhodamine B which is historically used as a quantum counter in fluorometers and which shows matching absorption and fluorescence excitation spectra down to (at least) ca. 360 nm.56,57 Accordingly, the determination of Φf(Rh-101) with QSSRM as the standard and an excitation wavelength of 348 nm yields a value of 0.915 ( 0.046 that compares well with the data reported for the other two methods in Table 1. All three approaches have shown that the fluorescence quantum yield of Rh-101 does not reach unity at room temperature as published by IUPAC and others.50-52,55,58,59 A value of 0.92 has been previously reported by Sauer et al.,53 utilizing an independent absolute method for determination (thermal lensing) and updating the earlier report of Drexhage.50 Moreover, a comparison of literature values for the other UV-vis-emitting dyes measured here yields data that do not agree especially for the two coumarins where the user currently has to choose from very diverse values, e.g., 0.58,60 0.74,61 0.93,62 or 0.9963 for C-102 and 0.26,61 0.38,60 0.55,62 or 0.5863 for C-153 in ethanol at room temperature. Influence of Oxygen. Although intersystem crossing is negligible for dyes such as Rh-101,64 quenching by dissolved oxygen has to be considered as a potential nonradiative process that can compete with fluorescence and reduce Φf under such ideal conditions as commonly used for fluorescence quantum yield measurements, i.e., a pure dye in a neat, spectroscopic-grade solvent. To assess the role of possible oxygen quenching of the fluorescence of Rh-101, six repeat determinations were thus carried out

Figure 2. Fluorescence excitation (left) and emission (right) spectra of Rh-101 in ethanol at 298 K (black) and 153 K (red). The inset shows the integrals of the bands corresponding to the S1 r S0 transition. The blue bar indicates the excitation wavelength used in the experiments. The bathochromic shift and narrowing of the band upon cooling result in a relative decrease in absorbance at the excitation wavelength.

for samples first prepared and measured under normal (air) conditions (see Table 1) and subsequently degassed by flowing argon through a septum-sealed cell for 20 min (determination along path B with λex = 348 nm).65 No significant effect was observed, i.e., Φf(Rh-101, degassed) = 0.912 ( 0.046 at 298 K, excluding the influence of oxygen quenching for this dye. Thus, ca. 8% of the absorbed photons are not converted into emitted photons, but their energy is funneled into internal conversion and vibrational and/or rotational motions. Temperature Dependence of Φf(Rh-101). If the last statement in the previous section is true, Φf(Rh-101) should increase and eventually reach unity upon reducing the temperature of the sample. We therefore measured the fluorescence quantum yield of Rh-101 as a function of temperature down to 153 K, i.e., slightly below the melting point of ethanol. In addition to the sources of error as described above, two additional aspects have to be considered when measuring Φf(T).14 First, both the refractive index n and the density F of a given solvent show a dependence on temperature66 and have to be accounted for by correcting the measured data for n(T) and F(T). Second, for most dyes, the shape of the absorption spectrum also changes as a function of temperature. The bands are often bathochromically shifted because of an increase of the solvent’s dielectric constant and narrowed because of the freezing of vibronic modes. Knowledge of the actual absorbance at every point of the temperature run is thus required. The latter was obtained as follows. After preparation, the initial solution was measured on a spectrophotometer and the absorbance at the excitation wavelength was 1237



ARTICLE

determined. Then, the sample was placed in the cryostat and fluorescence emission and excitation spectra were measured at the same temperature. After adjusting the temperature of the sample to the next desired value, fluorescence excitation and emission spectra were again recorded. To exclude the possible influence of shifting or narrowing absorption bands, the ratio of the integrals of the fluorescence excitation spectrum and the intensity reading at the excitation wavelength were calculated for both temperatures and compared. Because the oscillator strength of an electronic transition does not change as a function of temperature, such a comparison allows assessment of the change in absorbance at the excitation wavelength (Figure 2).67 Table 3 compiles the data obtained here for three independent temperature runs of three different samples. When starting with a reference value of 0.913 at 298 K, the fluorescence quantum yield of Rh-101 reaches unity below 180 K, where the solvent is highly viscous and vibrational modes are quasi-frozen. Table 3 also lists both absorption and emission maximum during the temperature run. Between 298 and 233 K, the Stokes shift remains virtually constant, suggesting that the parallel shifts are basically due to Table 3. Steady-State Fluorescence Properties and Stokes Shifts of Rh-101 in Ethanol as a Function of Temperature (λex = 520 nm) TK

a b

λex nma

λem nma

Δν~abs-em cm-1

Φf 0.913

b

298

564.4

587.6

700

273

565.7

589.0

699

0.941 ((0.072)

253

566.5

590.0

703

0.971 ((0.075)

233

567.5

591.1

703

0.980 ((0.076)

213

568.8

592.0

689

0.984 ((0.076)

193

570.0

593.1

683

0.985 ((0.076)

173 153

571.8 573.0

592.7 590.9

617 529

0.996 ((0.077) 1.000 ((0.077)

Maximum of fluorescence excitation and emission spectra, ( 0.5 nm. Used as a starting reference value.

changes in the solvent’s refractive index and dispersive interactions. Below 233 K, the reduction in Stokes shift indicates that both bands shift toward the 0,0-transition energy. These diverse trends reveal that the correction procedure employed for the absorbance at the excitation wavelength is essential. Moreover, the temperature-dependent measurements support our room temperature data that Rh-101 is highly fluorescent under normal laboratory conditions, yet its quantum yield is less than one. Red and NIR Dyes. Having established the fluorescence quantum yield of Rh-101 used as the standard for the red and NIR title dyes, the Φf of the latter were obtained according to the third approach outlined in the Strategy section, i.e., by constructing a chain of internal references from Rh-101 to the red (in analogy to Figure 1C). Table 4 compiles the data for every pair of compounds employed, along with the molar absorption coefficients and fluorescence anisotropies in ethanol at 298 K. Figure 3 shows the corresponding absorption and fluorescence emission spectra of the commercial dyes, and Figure 4 shows the respective spectra of the noncommercial candidate dyes. The data in Table 4 reveal that the xanthene and oxazine dyes Rh-101, CV, Ox-170, and Ox-1 fulfill best the requirements on simple-to-use fluorescence standards such as a moderate to high fluorescence quantum yield Φf > 0.1 and an almost isotropic emission. Commonly, polarization effects can be neglected at anisotropy values r e 0.05.38,69 Whereas Cc is not recommended as a fluorescence standard for the inexperienced user because of its considerably low Φf and high r, which is basically due to the fact that the excited chromophore decays very rapidly (τf = 83 ps in ethanol according to ref 70) and does not have sufficient time to fully rotate or reorientate during its excited state lifetime, HITCI is a borderline case with respect to polarization effects. IR-125, a popular NIR dye, already shows a considerable anisotropy of ca. 0.2. Our findings show that it is strongly recommended to use defined polarization conditions when attempting to measure fluorescence quantum yields of elongated fluorophores such as cyanine dyes, but also red-emitting styryl dyes such as LDS-821 (= styryl 9M), styryl 15, or styryl 2071

Table 4. Absorption and Fluorescence Maxima, Molar Absorption Coefficients, Fluorescence Quantum Yields, and Fluorescence Anisotropies of Rh-101, CV, Ox-170, Ox-1, Cc, HITCI, IR-140, IR-125, HEDITCP, HPDITCP, ONITCP, and ODNITCP in Ethanol at 298 K Φf dye

λabs nma

ελabs M-1 cm-1 b

λem nma

dye

no. of transfer standards used relative to Rh-101

r

Rh-101

564.4

118604

587.6

0.913 ( 0.046

CV

611.6

63574

625.7

0.578 ( 0.031

1

0.034

Ox-170

627.3

55450

644.7

0.579 ( 0.032

2

0.035

Ox-1

646.2

118820

662.7

0.141 ( 0.008

3

0.051

Cc

710.0

222632

720.8

0.012 ( 0.001

5

0.260

HITCI

743.0

251929

772.7

0.283 ( 0.017

4

0.084

IR-125

787.3

194171

818.2

0.132 ( 0.008

5

0.208

IR-140 HEDITCP

803.9 654.3

173919 129095

843.5 780.4

0.167 ( 0.010 0.195 ( 0.011

5 2

0.136 0.104

HPDITCP

658.3

71687

782.6

0.161 ( 0.009

3

0.111

ONITCP

860.5

193969

887.6

0.023 ( 0.002

6

0.283

ODNITCP

902.1

229857

929.8

0.014 ( 0.002

7

0.336

c

0.024

(0.5 nm. b uεrel = 2.53%, see section II.1, Supporting Information. Although the solutions could be kept dilute due to the use of 50 mm optical path length quartz cells, a possible negative bias due to fluorescence from the sample has to be considered. Such bias can amount to ca. -1.0% for rather resonant dyes such as fluorescein (ref 68). Corrections for such a bias were not made here; see additional comment in the Supporting Information. c Taking path B with the lowest uncertainty for the first chain link in this series Rh-101, urel = 5.08% (see the Supporting Information). a

1238



ARTICLE

Figure 4. Absorption and emission spectra of HEDITCP, HPDITCP, ONITCP, and ODNITCP in ethanol at 298 K.

Figure 3. Absorption and emission spectra of the commercial red and NIR dyes in ethanol at 298 K.

in rather low-viscosity liquids such as ethanol. The value found by us for Φf(CV) = 0.578 agrees well with several published values that lie between 0.56 and 0.6.72-74 Values such as 0.49 as given in ref 75 seem to be too low. Whereas an earlier value for Ox-170 (Φf = 0.5) published in ref 50 also seems to underrate that dye’s fluorescence quantum yield, 0.6 given in ref 72 agrees better with our value of 0.579. For Ox-1, data from the literature also deviate, ranging from less than 0.1 reported in ref 50 over 0.11 in ref 72 and 0.15 in ref 75 to 0.16 in ref 74, compared with 0.141 found here. The mismatch is, however, more pronounced when going further to the red as exemplified by Φf = 0.007 reported for Cc in ethanol76 or Φf = 0.05 reported for IR-125 in ethanol;77 here, Φf(Cc) = 0.012 and Φf(IR-125) = 0.132, see Table 4. A comparison of the spectra in Figures 3 and 4 reveals that the two midchain dimethylamino-substituted and rigidized cyanines HEDITCP and HPDITCP are promising candidates for fluorescent NIR standards because of their considerable Stokes shift and rather broad absorption bands, allowing for excitation across a broad wavelength range from ca. 530 to ca. 690 nm. These dyes possess Φf that lie between those of the IR-x dyes and HITCI and r values that are considerably small for dyes emitting in this spectral region. The two dyes with emission farthest into the NIR, ONITCP and ODNITCP, show smaller fluorescence quantum yield values of 0.023 and 0.014, which is most likely (at least in part) due to the increasing influence of the energy gap rule, i.e., enhanced internal conversion as the gap between ground and exited state is reduced.78 However, both dyes complement well popular dyes such as IR-125 and IR-140, which can only be used up to λex ∼ 800 nm, in covering the long-wavelength edge of the so-called “biological window” up to 950 nm for fluorescence applications.

Uncertainty of Measurement. The assessment of the relative uncertainties of measurement is given in detail in section II, Supporting Information. The propagation of uncertainty has been treated according to ref 79 for uncorrelated quantities. Whereas the relative uncertainties of the absorbances and the integrated fluorescence intensities amount to ca. 0.6% and e0.06%, respectively, the major contribution to the overall uncertainty stems from the reported value of the QY of QSSRM, i.e., 5%. This rather large uncertainty for the first chain link leads to a best achievable relative uncertainty of ca. 5.1% for Rh-101 in the case of direct determination along path B. For the other dyes and a determination along path C, depending on the amount of chain links necessary to bridge the gap to QSSRM (UV-vis range) or Rh-101 (vis-NIR range) and the fluorescence ability of a sample,80 the relative uncertainties lie between 5.08 and 5.31% (UV-vis range) and 5.35 and 11.62% (vis-NIR range). Because two calibration procedures have to be considered, the uncertainty for the approach that relies on different excitation wavelengths (path A) is higher than that of the transfer chain method with five UV-vis dyes as chain links. In our case, the use of two different λex would unfold an advantage only for the sixth and higher chain links on in the UV-vis range. However, as can be seen in subsection II.2.8, Supporting Information, improvements in the relative uncertainty of the first chain link;i.e., QSSRM here;would have a dramatic effect on the overall uncertainties along a transfer chain. This shows the importance of directing future work toward the determination of absolute QYs with lower uncertainties. Availability of QSSRM. Currently, the primary reference standard QSSRM is not available from NIST. We thus tested a commercially available QS sample that is, to our knowledge, the QS of highest purity presently on the market (see the Experimental Section for details). This sample was measured against QSSRM according to the procedure described in Figure 1B (λex = 348 nm). Absorption and fluorescence maxima were found to be identical within experimental error to those of QSSRM, and the fluorescence quantum yield also agrees well with that of QSSRM, i.e., amounts to 0.609 ( 0.031. The quinine sulfate from this source can thus serve as a suitable alternative. Guidelines for Use of Dyes and QY Values. The data provided here are perhaps most relevant for users who want to determine relative fluorescence quantum yields according to the 1239



ARTICLE

Table 5. Φf Values and Recommended Wavelength Ranges of Use for the Dyes in Ethanol at Room Temperature (298 K)a dye

Φf (urel)

QS

0.609 ((0.031)

C-102 C-153

0.764 ((0.039) 0.544 ((0.028)

excitation range nmb emission range nmc

340-420 360-470

430-530 485-635

DCM

0.435 ((0.022)

400-530

560-700

Rh-101

0.913 ((0.046)

515-565

585-665

CV

0.578 ((0.031)

520-610

625-710

Ox-170

0.579 ((0.032)

490-625

645-740

Ox-1

0.141 ((0.008)

570-645

665-775

HITCI

0.283 ((0.017)

650-745

770-870

IR-125 IR-140

0.132 ((0.008) 0.167 ((0.010)

690-785 700-800

820-935 840-915

HEDITCPd

0.195 ((0.011)

570-690

700-890

HPDITCPd

0.161 ((0.009)

530-670

740-880

ONITCPd

0.023 ((0.002)

710-860

890-1000

ODNITCPd 0.014 ((0.002)

760-900

930-1025

Refs 82 and 83. b The recommended excitation wavelength range of use is defined by a band height of ca. 20% of the maximum height at the longer and shorter wavelength side of the first absorption band for the dyes with large Stokes shifts (e.g., C-102) and by a band height of ca. 20% of the maximum height at the shorter wavelength side and the maximum of the absorption band for the dyes with rather resonant bands (e.g., HITCI). c Emission range refers to the usable range when no emission correction curve is available. d Noncommercial dyes.

good agreement. Whereas the approach that relies on matching absorption and fluorescence excitation spectra of an unknown dye (Figure 1B) is limited to those dyes that fulfill this requirement, the strategy utilizing both excitation and emission correction curves supposedly is not very relevant for most users because of the inherent difficulty and considerable effort of measuring excitation correction curves (Figure 1A). The last approach (Figure 1C), using a similarly absorbing standard under identical excitation conditions and/or constructing a chain of transfer standards from a reliable reference fluorophore is thus the most straightforward one since it only requires the generation of an emission correction curve. Several protocols exist for this task,38,69,84 and certified standards are commercially available.39,85 On the basis of the protocol sketched above and the data reported here, we believe that the fluorescence quantum yield data generated by more or less experienced users of fluorescence instrumentation in the future will benefit from the considerably simple yet robust approach proposed, involving fluorescence quantum yields determined on a traceably characterized instrument.

a

popular Parker-Rees method, especially if a reliable excitation correction curve is not available. The following synopsis illustrates the procedure: (a) choice of an appropriate standard dye taking into account the recommended ranges in Table 5; (b) preparation of (dilute) solutions and measurement of the spectra (at best using longer optical path length quartz cells for absorption) at least in a repeat determination;81 (c) correction of the spectra for blank (absorption and fluorescence) and spectral signatures of the instrument (fluorescence, commonly with an emission correction curve); (d) extraction of Ai(λex) and As(λex) from the absorption spectra, preferably taking into account eq 3, and calculation of fi(λex) and fs(λex); R(e) integration Rof the corrected fluorescence spectra to obtain λemFi(λem) and λemFs(λem); (f) if necessary, consultation of a reference work on refractive indices or measurement of the refractive indices of the different solvents employed for both sample and standard; (g) inclusion of values from steps d-f and the respective Φsf from Table 5 into eq 1 and calculation of Φif; (h) assessment of the uncertainty of measurement (e.g., consulting section II.2, Supporting Information, for details and analogies) using urel for the standard from Table 5.

’ CONCLUSION The fluorescence quantum yield is a parameter that plays a significant role in many fluorescence-based applications. Many laboratories carry out such measurements, and a large number of scientific publications report values for commercial and wellinvestigated as well as exotic or uniquely synthesized fluorophores. Despite this popularity, values on particular dyes found in the literature often differ largely. In this work, we have presented three different approaches to obtain relative fluorescence quantum yields for dyes, all three of them yielding results that are in

’ ASSOCIATED CONTENT

bS

Supporting Information. Additional information as noted in text. This material is available free of charge via the Internet at http://pubs.acs.org.

’ AUTHOR INFORMATION Corresponding Author

*Fax: þþ49(0)30-81041157. E-mail: [email protected].

’ ACKNOWLEDGMENT The authors thank Drs. Y. L. Slominskii and J. L. Bricks, National Academy of Sciences of the Ukraine, Kiev, for kindly providing samples of ONITCP and ODNITCP, Drs. D. Pfeifer, J. Pauli, and U. Resch-Genger, BAM Div. 1.5, for calibration support with the SLM 8100, BAM Div. 1.2 Organic Chemical Analysis; Reference Materials for HPLC purity checks of the dyes investigated, and Drs. W. Bremser, BAM Div 1.4, and C. Monte, AG 7.31 Temperaturstrahlung, PTB, Berlin, for critical discussions. ’ REFERENCES (1) Topics in Fluorescence Spectroscopy Series; Lakowicz, J. R., Ed.; Plenum: New York and Springer: Berlin, Germany, 1992-2006; Vols. 1-11. (2) Springer Series on Fluorescence; Wolfbeis, O. S., Ed.; Springer: Berlin, Germany, 2001-2010; Vols. 1-6, 8, 9. (3) Handbook of Biological Confocal Microscopy, 3rd ed.; Pawley, J. B., Ed.; Springer: New York, 2006. (4) Probes and Tags to Study Biomolecular Function; Miller, L. W., Ed.; Wiley-VCH: Weinheim, Germany, 2008. (5) Demchenko, A. Introduction to Fluorescence Sensing; Springer: Berlin, Germany, 2009. (6) Lakowicz, J. R. Principles of Fluorescence Spectroscopy, 3rd ed.; Springer: Berlin, Germany, 2006. (7) Valeur, B. Molecular Fluorescence; Wiley-VCH: Weinheim, Germany, 2001. (8) Gonc-alves, M. S. T. Chem. Rev. 2009, 109, 190–212. (9) Borisov, S. M.; Klimant, I. Analyst 2008, 133, 1302–1307. (10) Zheng, J.; Nicovich, P. R.; Dickson, R. M. Annu. Rev. Phys. Chem. 2007, 58, 409–431. 1240


Analytical Chemistry (11) Wolff, M.; Kredel, S.; Wiedenmann, J.; Nienhaus, G. U.; Heilker, R. Comb. Chem. High Throughput Screening 2008, 11, 602–609. (12) Resch-Genger, U.; Grabolle, M.; Cavaliere-Jaricot, S.; Nitschke, R.; Nann, T. Nat. Methods 2008, 5, 763–775. (13) Thomas, S. W.; Joly, G. D.; Swager, T. M. Chem. Rev. 2007, 107, 1339–1386. (14) Rurack, K. Springer Ser. Fluoresc. 2008, 5, 101–145. (15) Demas, J. N. In Measurement of Photolouminescence, Optical Radiation Measurements; Mielenz, K. D., Ed.; Academic: New York, 1982; Vol. 3, pp 195-248. (16) Velapoldi, R. A.; Mielenz, K. D. NBS Special Publication 260-64; National Bureau of Standards: Washington, DC, 1980. (17) Parker, C. A.; Rees, W. T. Analyst 1960, 85, 587–600. (18) fx(λex) is equivalent to the absorptance Rx(λex), the use of which, however, is not recommended anymore; see: Braslavsky, S. E. Pure Appl. Chem. 2007, 79, 293–465. (19) DeRose, P. C.; Wang, L.; Gaigalas, A. L.; Kramer, G. W.; ReschGenger, U.; Panne, U. Springer Ser. Fluoresc. 2008, 5, 33–62. (20) Dittrich, P. S.; Manz, A. Anal. Bioanal. Chem. 2005, 382, 1771– 1782. (21) Seidel, M.; Niessner, R. Anal. Bioanal. Chem. 2008, 391, 1521– 1544. (22) Petty, H. R. Microsc. Res. Tech. 2007, 70, 687–709. (23) Sako, Y.; Uyemura, T. Cell Struct. Funct. 2002, 27, 357–365. (24) Reits, E. A. J.; Neefjes, J. J. Nat. Cell Biol. 2001, 3, E145–E147. (25) Erickson, S. J.; Godavarty, A. Med. Eng. Phys. 2009, 31, 495–509. (26) Wessels, J. T.; Busse, A. C.; Mahrt, J.; Dullin, C.; Grabbe, E.; Mueller, G. A. Cytometry 2007, 71A, 542–549. (27) Weissleder, R. Nat. Biotechnol. 2001, 19, 316–317. (28) Achilefu, S. Technol. Cancer Res. Treat. 2004, 3, 393–409. (29) Kiyose, K.; Kojima, H.; Nagano, T. Chem. Asian J. 2008, 3, 506–515. (30) Descalzo, A. B.; Xu, H.-J.; Shen, Z.; Rurack, K. Ann. N.Y. Acad. Sci. 2008, 1130, 164–171. (31) Amiot, C. L.; Xu, S.; Liang, S.; Pan, L.; Zhao, J. X. Sensors 2008, 8, 3082–3105. (32) Miller, J. N. Springer Ser. Fluoresc. 2008, 5, 147–162. (33) Galanin, M. D.; Kut’enkov, A. A.; Smorchkov, V. N.; Timofeev, Y. P.; Chizhikov, Z. A. Opt. Spectrosc. 1982, 53, 405–409. (34) Bindhu, C. V.; Harilal, S. S. Anal. Sci. 2001, 17, 141–144. (35) Benson, R. C.; Kues, H. A. J. Chem. Eng. Data 1977, 22, 379– 383. Benson and Kues name fluctuations and wavelength inaccuracies of the excitation channel as well as dye purity issues as major potential sources of errors. (36) Certificate of Analysis, Standard Reference Material SRM 936a, Quinine Sulfate Dihydrate. National Institute of Standards and Technology: Gaithersburg, MD, 1994. https://www-s.nist.gov/srmors/view_ detail.cfm?srm=936A (accessed Dec 1, 2010). Note that, according to NIST’s website, this SRM is currently out of stock. (37) Hollandt, J.; Taubert, R. D.; Seidel, J.; Resch-Genger, U.; GuggHelminger, A.; Pfeifer, D.; Monte, C.; Pilz, W. J. Fluoresc. 2005, 15, 301–313. (38) Resch-Genger, U.; Pfeifer, D.; Monte, C.; Pilz, W.; Hoffmann, A.; Spieles, M.; Rurack, K.; Hollandt, J.; Taubert, D.; Sch€onenberger, B.; Nording, P. J Fluoresc. 2005, 15, 315–336. (39) (a) Pfeifer, D.; Hoffmann, K.; Hoffmann, A.; Monte, C.; ReschGenger, U. J. Fluoresc. 2006, 16, 581–587. (b) Updated, see: BAM-F001, -F002a, -F003-F005, Calibration Kit, Spectral Fluorescence Standards; BAM Federal Institute for Materials Research and Testing: Berlin, 2009. http://www.comar.bam.de/home/search_crm.php (accessed Dec 1, 2010). (c) Resch-Genger, U.; Pfeifer, D.; Monte, C.; Hoffmann, A.; Nording, P.; Sch€onenberger, B.; Hoffmann, K.; Spieles, M.; Rurack, K. German Patent DE 10 2004 044 717, 2010; U.S. Patent US 2010/ 0219333, 2010. (40) Monte, C.; Resch-Genger, U.; Pfeifer, D. R.; Taubert, D. R.; Hollandt, J. Metrologia 2006, 43, S89–S93. (41) Rh-101, 8-(2-carboxyphenyl)-2,3,5,6,11,12,14,15-octahydro1H,4H,10H,13H-diquinolizino[9,9a,1-bc:90 ,9a0 ,1-hi]xanthylium perchlorate;

ARTICLE

CV, 5,9-diaminobenzo[a]phenoxazonium perchlorate; Ox-170, 9-ethylamino-5-ethylimino-10-methyl-5H-benzo[a]phenoxazonium perchlorate; Ox-1, 3-diethylamino-7-diethyliminophenoxazonium perchlorate; Cc, 1,10 -diethyl-4,40 -carbocyanine iodide; HITCI, 1,10 ,3,3,30 ,30 -hexamethylindotricarbocyanine iodide; IR-140, 3,30 -diethyl-5,5-dichloro-3, 5-ethylene-4-(diphenylamino)-2,20 -indotricarbocyanine perchlorate; IR-125, 1,10 -di(butylenesulfonate)-3,3,30 ,30 -tetramethyl-4,40 ,5,50 -dibenzo-2,20 indotricarbocyanine sodium; C-102, 2,3,5,6,-1H,4H-tetrahydro-8-methylquinolizino-(9,9a,1-gh)coumarin; C-153, 2,3,5,6,-1H,4H-tetrahydro-8trifluormethylquinolizino-(9,9a,1-gh)coumarin; DCM, 4-dicyanomethylene2-methyl-6-(p-dimethylaminostyryl)-4H-pyran. (42) HEDITCP, 1,10 ,3,3,30 ,30 -hexamethyl-3,5-ethylene-4-(dimethylamino)-2,20 -indotricarbocyanine perchlorate; HPDITCP, 1,10 ,3,3,30 ,30 hexamethyl-3,5-propylene-4-(dimethylamino)-2,20 -indotricarbocyanine perchlorate. (43) Descalzo, A. B.; Rurack, K. Chem.;Eur. J. 2009, 15, 3173– 3185. (44) ONITCP, 1,10 ,3,3,30 ,30 (3,30 )-octamethyl-3,5-neopentylene-2,20 indotetracarbocyanine perchlorate; ODNITCP, 1,10 ,3,3,30 ,30 (3,30 )octamethyl-4,40 ,5,50 -dibenzo-3,5-neopentylene-2,20 -indotetracarbocyanine perchlorate. (45) Slominskii, Y. L.; Smirnova, A. L.; Popov, S. V. Ukr. Khim. Zh. 1984, 50, 638–640. (46) Grabolle, M.; Spieles, M.; Lesnyak, V.; Gaponik, N.; Eychm€uller, A.; Resch-Genger, U. Anal. Chem. 2009, 81, 6285–6294. (47) Demas, J. N. Anal. Chem. 1973, 45, 992–994. (48) Landolt-B€ornstein compendium. http://www.springermaterials. com/navigation/ (accessed Dec 1, 2010). (49) The absence of a dependence of Φf on λex was verified for these dyes across the wavelength range of their first absorption bands. (50) Drexhage, K. H. J. Res. Natl. Bur. Stand. 1976, 80A, 421–428. (51) Karstens, T.; Kobs, K. J. Phys. Chem. 1980, 84, 1871–1872. (52) Galanin, M. D.; Kut’enkov, A. A.; Smorchkov, V. N.; Timofeev, Y. P.; Chizhikov, Z. A. Opt. Spectrosc. 1982, 53, 405–409. (53) Sauer, M.; Han, K.-T.; M€uller, R.; Nord, S.; Schulz, A.; Seeger, S.; Wolfrum, J.; Arden-Jacob, J.; Deltau, G.; Marx, N. J.; Zander, C.; Drexhage, K. H. J Fluoresc. 1995, 5, 247–261. (54) Nunes Pereira, E. J.; Berberan-Santos, M. N.; Fedorov, A.; Vincent, M.; Gallay, J.; Martinho, J. M. G. J. Chem. Phys. 1999, 110, 1600–1610. (55) Porres, L.; Holland, A.; Palsson, L. O.; Monkman, A. P.; Kemp, C.; Beeby, A. J. Fluoresc. 2006, 16, 267–272. Note that these authors used acidic methanol as solvent. (56) Melhuish, W. H. J. Opt. Soc. Am. 1962, 52, 1256–1258. (57) Taylor, D. G.; Demas, J. N. Anal. Chem. 1979, 51, 712–717. (58) Eaton, D. F. J. Photochem. Photobiol., B 1988, 2, 523–531. (59) Eaton, D. F. Pure Appl. Chem. 1988, 60, 1107–1114. (60) Reynolds, G. A.; Drexhage, K. H. Opt. Commun. 1975, 13, 222–225. (61) Jones, G., II; Jackson, W. R.; Halpern, A. M. Chem. Phys. Lett. 1980, 72, 391–395. (62) Fletcher, A. N.; Bliss, D. E. Appl. Phys. 1978, 16, 289–295. (63) Kubin, R. F.; Fletcher, A. N. Chem. Phys. Lett. 1983, 99, 49–52. (64) Tomasini, E. P.; San Roman, E.; Braslavsky, S. E. Langmuir 2009, 25, 5861–5868. (65) With regard to the method of degassing, the performance of Ar deoxygenation proved to be similarly effective to the frequently employed freeze-pump-thaw technique while offering more straightforward handling. (66) For dependences, see the respective entries for ethanol in ref 48. (67) When lowering the temperature and especially when freezing the solvent, significant changes in viscosity occur and give rise to polarization effects, rendering the use of defined polarization conditions in excitation and emission mandatory. Furthermore, it is important to use optically dilute solutions because the absorption and emission bands are not only narrowed upon reduction of the temperature, but the Stokes shift is commonly also decreased, rendering reabsorption errors more significant at low temperatures. 1241



ARTICLE

(68) DeRose, P. C.; Kramer, G. W. J. Lumin. 2005, 113, 314–320. (69) Gardecki, J. A.; Maroncelli, M. Appl. Spectrosc. 1998, 52, 1179– 1189. (70) Bruce, D.; Miners, J. Photochem. Photobiol. 1993, 58, 464–468. (71) Brackmann, U. Lambdachrome Laser Dyes, 2nd ed.; Lambda Physik: G€ottingen, Germany, 1997. (72) Sens, R.; Drexhage, K.-H. J. Lumin. 1981, 24-25, 709–712. (73) Sabol, J. E.; Rockley, M. G. J. Photochem. Photobiol., A 1987, 40, 245–257. (74) Petukhov, V. A.; Popov, M. B.; Krymova, A. I. Sov. J. Quantum Electron. 1986, 16, 503–509. (75) Wang, T.; Zhang, B. Q.; Pan, J. L.; Gu, P.; Xu, Q.; Sun, Q. Chin. Sci. Bull. 1989, 34, 1756–1757. (76) Dempster, D. N.; Morrow, T.; Rankin, R.; Thompson, G. F. Chem. Phys. Lett. 1973, 18, 488–492. (77) Soper, S. A.; Mattingly, Q. L. J. Am. Chem. Soc. 1994, 116, 3744– 3752. (78) Siebrand, W. J. Chem. Phys. 1967, 46, 440–447. (79) Evaluation of Measurement Data;Guide to the Expression of Uncertainty in Measurement. Joint Committee for Guides in Metrology JCGM: Paris, 2008 (first edition), 2010 (corrected version). (80) The weaker the fluorescence of a compound, the higher is usually the relative uncertainty introduced by the experimental standard deviation; see section II.2.3e, Supporting Information. (81) It is also recommended to choose a standard dye that has a comparable fluorescence quantum yield as the sample, preferably Φfs e 10 Φfi. If the standard is, for instance, 100 times higher in fluorescence, instrument settings that avoid detector saturation might have to be chosen which can introduce a higher uncertainty for the dye displaying weak signals. (82) Due to its considerably low fluorescence quantum yield and high anisotropy, Cc is not included here. (83) The fluorescence emission of C-102, C-153, DCM, Rh-101, CV, Ox-170, Ox-1, and HEDITCP does not depend on the excitation wavelength within the specified excitation range. For HITCI, IR-125, IR140, ONITCP, and ODNITCP, the match of absorption and fluorescence excitation spectra could not be verified because of the spectral limitations of the excitation channel of the instrument employed. As reported in ref 43, absorption and fluorescence excitation spectra show a slight mismatch above ca. 680 nm for HPDITCP. Below 670 nm, the fluorescence quantum yield is independent of excitation wavelength in the specified range. (84) Lippert, E.; N€agele, W.; Seibold-Blankenstein, I.; Staiger, U.; Voss, W. Z. Anal. Chem. 1959, 170, 1–18. (85) SRMs 2940-2943, Relative Intensity Correction Standards for Fluorescence Spectroscopy (Orange, Green, Ultraviolet, Blue Emission). National Institute of Standards and Technology: Gaithersburg, MD, 1994. http://www.nist.gov/srm/ (accessed Dec 1, 2010).

1242


Fluorescence Quantum Yields of a Series of Red and Near-Infrared

Recommend Documents