Open Access
Add to BibSonomy
Abstract
The fluorescence spectra of dye solutions change their spectral signature with temperature. This effect is frequently used for temperature imaging in liquids and sprays based on two-color laser-induced fluorescence (2cLIF) measurements by simultaneously detecting the fluorescence intensity in two separate wavelength channels resulting in a temperature-sensitive ratio. In this work, we recorded temperature-dependent absorption and fluorescence spectra of solutions of five laser dyes (coumarin 152, coumarin 153, rhodamine B, pyrromethene 597, and DCM) dissolved in ethanol, a 35/65 vol.% mixture of ethanol/2-ethylhexanoic acid, ethanol/hexamethylsiloxane, o-xylene, and 1-butanol to investigate their potential as temperature tracers in evaporating and burning sprays. The dissolved tracers were excited at either 266, 355, and 532 nm (depending on the tracer) for temperatures between 296 and 393 K (depending on the solvent) and for concentrations ranging between 0.1 and 10 mg/l. Absorption and fluorescence spectra of the tracers were investigated for their temperature dependence, the magnitude of signal re-absorption, the impact of different solvents, and varying two-component solvent compositions. Based on the measured fluorescence spectra, the tracers were analyzed for their 2cLIF temperature sensitivity in the respective solvents. Coumarin 152 showed for single-component solvents the overall best spectroscopic properties for our specific measurement situation related to temperature imaging measurements in spray-flame synthesis of nanoparticles as demonstrated previously in ethanol spray flames [Exp. Fluids 61, 77 (2020) [CrossRef] ].
© 2021 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. INTRODUCTION
The non-intrusive spatially and temporally resolved investigation of spray formation and propagation is widely accepted as a necessary tool for a better understanding of these processes in various practical applications and for the development and validation of physical models. The work in this paper is motivated by the investigation of spray-flame synthesis of functional nanomaterials [1], where one or several metal-atom containing precursors are dissolved in combustible liquids that are processed towards oxidic nanoparticles in a flame. In these systems, the droplet sizes, droplet evaporation, and spatial and temporal distribution of the liquid phase as well as liquid-phase temperatures are of interest to provide data for analyzing, understanding, and ultimately simulating the process [2] using well-defined experimental conditions [3]. Nanoparticles can be formed through two competing mechanisms, the liquid-to-particle process, where precipitation occurs within the droplet, or the gas-to-particle process, where the droplet evaporates before homogeneous nucleation of particles occurs in the gas phase. In this context, the liquid-phase temperature is of particular interest because the temperature influences (i) the kinetics of liquid-phase reactions and subsequent precipitation, (ii) the evaporation [4], and (iii) the tendency towards droplet explosions [5,6], which are often considered to be beneficial because they accelerate spray break-up and thus evaporation, therefore supporting the often preferred particle formation through the gas-to-particle process.
Optical diagnostics techniques with imaging capabilities are often used to provide instantaneous information about the liquid phase, i.e., droplet velocities [7] and size distribution [8–10], temperature [11,12], species [13], and possibly particle characteristics [14,15] of materials such as soot [16,17] and metal-oxide particles, as well as the related molecular intermediates [18]. From the numerous optical techniques cited above, laser-induced fluorescence (LIF) has been widely applied for liquid-phase temperature imaging in mixing flows [12,19] and sprays [20–22] using molecular tracers, either naturally present (e.g., aromatic fuel components) or added to some fraction of the liquid. In case the fluorescence intensity should represent the liquid volume, the tracers require similar thermo-physical properties (vapor pressure, transport properties, etc.) as the hydrocarbon [23] or aqueous solvent [24] to ensure co-evaporation in the respective medium. For ratiometric measurements often used for thermometry, in contrast, tracers with especially high quantum yields are often preferred because they can be applied in minute quantities without influencing the system under study. Variations in local volumetric tracer concentrations due to evaporation of the solvent is canceled in ratiometric data analysis. Additionally, strong temperature sensitivity of the fluorescence in the related solvent after excitation with a suitable laser wavelength and chemical stability in the temperature range of interest are a prerequisite when selecting the best tracer for a specific experiment. Because of their high quantum efficiencies and excellent solubility in hydrocarbons, alcohols, and water, organic dyes are often a preferred choice. Two-color LIF (2cLIF) thermometry is one of the approaches for temperature imaging, where the different temperature dependence of the fluorescence signal in two selected wavelength regions of the fluorescence spectrum is exploited to determine the temperature from the intensity ratio [25–27]. Several investigations on tracers for LIF have been done recently [28–30].
In this paper, we compare various potential tracers for 2cLIF imaging [coumarin 152, coumarin 153, rhodamine B, DCM, and pyrromethene 597] in solvent systems frequently applied for spray-flame synthesis of nanomaterials, such as 1-butanol, ethanol (EtOH), o-xylene, and mixtures of EtOH and 2-ethylhexanoicacid (EHA) [31,32]. We also investigate one system containing a precursor for silica particle production (hexamethyldisiloxane, HMDSO) dissolved in EtOH. The present investigation is an extension of our earlier work devoted to the characterization of several tracers in pure EtOH [33]. Temperature-dependent absorption and fluorescence spectra (with excitation at 266, 355, and 532 nm) of the various dye solutions are presented and systematically analyzed for their temperature sensitivity in 2cLIF thermometry, also determining optimized choices of the selection of the two detection channels. Furthermore, the impact of fluorescence signal re-absorption was measured to validate our model assumptions (cf. Section 5) that were used to assess the impact of re-absorption on our measurements in the spray flame.
2. THEORETICAL BACKGROUND OF TWO-COLOR LIF THERMOMETRY
The principle of 2cLIF thermometry is detailed in many publications [27,34,35] and therefore will only be briefly summarized here. After laser excitation at a fixed wavelength within the absorption spectrum, the resulting fluorescence signal, ${I_{\rm{LIF}}}$, recorded at a wavelength $\lambda$ generally shows an exponential dependence on temperature as given by [27]
The temperature can be inferred from the fluorescence intensity ratio recorded on two selected spectral windows of the fluorescence spectrum by placing appropriate spectral filters [center wavelength ${\lambda _i}$, full width half-maximum (FWHM) bandwidth $\Delta {\lambda _i},\;i = {{1}},{{2}}$] in front of the detector [11]:
The $\beta$ factors at both detection wavelength bands can be evaluated from the measured fluorescence spectra. In practical applications, the ratio needs to be calibrated with a measurement at known temperature ${T_0}$ [11], from which the temperature can then be derived,
In measurements where high tracer concentrations are necessary and/or the detected fluorescence light travels through extended regions of the liquid towards the detector, fluorescence signal re-absorption (signal trapping) can be a problem with most rhodamine dyes, where absorption and emission spectra overlap significantly [36]. With the tracer concentration approaching a critical value, e.g., in the late stages of droplet evaporation, fluorescence self-quenching also can additionally affect the two-color ratio calibration [11].
3. EXPERIMENTAL DETAILS
A. Measurement of Absorption and Fluorescence Spectra
Absorption and fluorescence spectra of five laser dyes (coumarin 152, coumarin 153, DCM, pyrromethene 597, and rhodamine B) were recorded in various solvents (EtOH, EHA/EtOH, 1-butanol, o-xylene, water). Relevant photo-physical and thermo-physical properties of the laser dyes and relevant chemical properties of the solvents are provided in Tables S1 and S2 of Supplement 1. Fluorescence measurements were performed with a spectrofluorometer (Horiba, Fluorolog-3 Model FL3-22), which uses a xenon arc lamp as light source. To achieve the desired temperatures, the probe solution was placed with a UV-transmissive optical quartz cuvette (Hellma cuvette 111-QS, ${{10}}\;{\rm{mm}} \times {{10}}\;{\rm{mm}}$) inside a temperature-controlled measurement cell, which in turn was placed in the instrument. This measurement cell was designed to fit in the sample compartment of the fluorescence spectrometer with its UV-transmissive quartz windows in the beam paths of the instrument. The temperature of the tracer solution was measured with a thermocouple (type K) directly placed in the liquid inside the cuvette. From the results during the calibration process described in our previous publication [33], a homogeneous temperature distribution within the cuvette was shown to exist. Depending on their previously determined absorption spectra in the respective solvent, the tracers were excited with a bandwidth of 1 nm (FWHM) at harmonics of neodymium-doped yttrium aluminum garnet (Nd:YAG) laser wavelengths listed in Table 1. The fluorescence spectra were detected at 90° to the excitation light sheet with a 0.5 nm increment with a detection bandwidth of 1 nm. The integration time was fixed to 0.1 s per wavelength increment. The slightly fluctuating excitation intensity of the xenon lamp was corrected by an integrated reference detector. The wavelength-dependent detection sensitivity of the instrument was corrected by a software-integrated correction file provided and updated by the manufacturer. Furthermore, the fluorescence spectra were background corrected by prior performed fluorescence measurements of the pure solvent at all desired temperature settings, eliminating minor disturbing background artifacts. For each temperature, three measurements were conducted and averaged in the evaluation process to improve the signal-to-noise ratio.
Table 1. Tracers Investigated in This Work with the Investigated Concentration Range (0.1–10 mg/l) and Their Molar Massa
Absorption measurements were performed with a UV/visible (VIS) absorption spectrometer (Varian Inc., Mod. Cary 400) using the same heatable sample cell and the same quartz cuvette, resulting in a probe path length of 10 mm. The absorption and transmission spectra were measured with an increment of 0.5 nm between 175 and 900 nm. The measurements were corrected for the wavelength-dependent intensity of the light source and detector sensitivity by the internal software ($I/{I_0}$). The absorption of the solvent and scattering effects from the cell and cuvette were corrected by subtracting pre-recorded spectra of the pure solvent in the probe containment (baseline correction).
With tracer concentrations of 0.1, 1, and 10 mg/l, a factor of 100 was covered in the measurements. The 10 mg/l samples were prepared by weighing (Mettler AT201, readability: 0.01 mg) 10 mg of laser dye in powder form and dissolving it in 1 L of solvent, while the smaller concentrations were generated by further dilution. Table 1 lists all tracers investigated in this work together with respective excitation wavelengths and the concentration ranges expressed in mg/l and mol/l.
In view of tracer dissociation, the irradiance in the used instrument (Horiba, Fluorolog-3 Model FL3-22) is significantly lower than commonly used laser beam irradiances. Additionally, the dye solution is only exposed to the light beam during the actual measurements, minimizing potential dye degradation by photo-dissociation. On test purposes, we recorded the dye spectra after the measurements and cooling down without any detectable impact on the spectra proving negligible photo and thermal degradation. Similar results were described recently by Koegle et al. [29], who measured the fluorescence spectra of nile red from low to high and from high to low temperatures, acquiring the same curves.
4. RESULTS
In this section, temperature-dependent absorption and fluorescence spectra of coumarin 152, coumarin 153, rhodamine B, DCM, and pyrromethene 597 dissolved (by volume) in a mixture of 65% EHA and 35% EtOH (65EHA35EtOH) are presented. An investigation of the impact of changing solvent compositions during the evaporation process on the fluorescence spectra of the tracers is presented in Section 6. This mixture is of interest in liquid-phase temperature measurements because it is used as fuel in spray-based gas-phase flame synthesis of metal-organic nanoparticles because of its beneficial stabilizing and droplet-explosion enhancing effect [3,4]. Similar results obtained for the investigated tracers dissolved in EtOH, 1–butanol, and o-xylene are presented in Supplement 1. The absorption spectra are presented in units of a molar attenuation coefficient $\varepsilon (\lambda)$, which is derived from the Beer–Lambert law with the absorbance $A(\lambda)$, molar concentration $c$, and absorption path length $l$:
In addition to temperature, the effect of tracer concentration and the type of solvent on the absorption and fluorescence spectra are investigated. The given values for the blue- and redshift were calculated at the respective half-maximum (HM) points of the fluorescence spectra. In the upper diagram of all figures presenting the temperature-dependent spectra, the absorption spectra are plotted in absolute units (${\rm{l}}\;{{\rm{mol}}^{- 1}}\;{{\rm{cm}}^{- 1}}$), and the fluorescence spectra are shown on an individual absolute intensity scale delivered by the spectrometer software in counts per second (CPS). The respective lower diagrams present the peak-normalized spectra, the $\beta$ factors, and the optimized 2cLIF detection bands as blue and red rectangles (cf. Section 8). Measurements were performed for temperatures up to 393 K. For the concentration dependence of the fluorescence spectra (for a signal path length of 5 mm in the cuvette resulting from an excitation with the light sheet passing through the center of the cuvette), these were normalized at a spectral position where no overlap of absorption and fluorescence spectra exists, i.e., re-absorption of signal radiation can be neglected. From these results, the magnitude of fluorescence signal re-absorption can be inferred, and model calculations to assess the impact of this effect on the detectable fluorescence spectra were validated (cf. Section 5). Finally, the effect of the solvent on the absorption and fluorescence spectra of each tracer is shown.
With respect of tracer solubility, rhodamine B was the only tracer with a high solubility in water, coumarin 152 showed a very low solubility, and coumarin 153, DCM, and pyrromethene 597 were not soluble in water. Among the other solvents, only rhodamine B was not soluble in o-xylene. Furthermore, for all tracers, no effect of the excitation wavelengths was observed on the fluorescence spectra.
Fig. 1. Temperature-dependent absorption (up to 373 K) and fluorescence spectra (up to 393 K) of 1 mg/l coumarin 152 in 65EHA35EtOH, excitation at 355 nm indicated with a purple line. Top: absolute values with the inset illustrating the peak intensities versus temperature; bottom: peak-normalized values, optimized 10 nm 2cLIF detection bands (blue and red rectangles), and $\beta$ factor for the spectra at 303 and 373 K.
A. Coumarin 152
1. Temperature Dependence of Absorption and Fluorescence Spectra
In Fig. 1, the absorption and fluorescence spectra of 1 mg/l coumarin 152 in 65EHA35EtOH are presented for a temperature range from room temperature to 393 K. Additionally, the $\beta$ factors calculated from the smoothed spectra for the temperature span of 303–373 K are plotted in the lower diagram. Below 300 nm, the absorption spectra are cut off because EHA starts to absorb strongly. Hence, coumarin 152 in 65EHA35EtOH can only be excited at 355 nm among the possible harmonics of the Nd:YAG laser, as 266 nm is strongly attenuated by the solvent, and 532 nm is not absorbed by the tracer.
In comparison to other tracers, coumarin 152 exhibits a large Stokes shift and thus a small overlap of the absorption and fluorescence spectra. As depicted in the inset of Fig. 1, the absolute fluorescence intensity decreases from 303 to 363 K by 17% moderately before the intensity starts to rise strongly by 78% from 363 to 393 K. The same behavior was described recently by Koegle et al. in 2020 [29] for nile red in a solvent mixture of 80 vol.% $n$-decane and 20 vol.% butanol. This phenomenon is also observable for the absorption spectra, which we measured up to 373 K. Here, the absorption also starts to rise beyond 363 K. With 0.14 nm/K between 303 and 343 K, the fluorescence spectra of coumarin 152 in 65EHA35EtOH exhibit a significant blueshift (Table 2), which becomes stronger with rising temperature. Also, the absorption spectra show a blueshift with increasing temperature—here the blueshift is 0.09 nm/K smaller. The $\beta$ factor is negative in the spectral range below 470 nm and starts to rise from ${-}{{1700}}\;{\rm{K}}$ at about 430 nm to about 500 K from 550 nm on. The variation of the $\beta$ factors over wavelength already indicates a high temperature sensitivity of this tracer for 2cLIF.
Table 2. Characterization of the Fluorescence Spectra of 1 mg/l Coumarin 152 Dissolved in Various Solvents, Concentration in Water due to Poor Solubility not Exactly Known, Excitation at 355 nma
2. Impact of Concentration: Signal Re-Absorption
Shown in Fig. 2 are the fluorescence spectra of coumarin 152 at concentrations of 0.1, 1, and 10 mg/l dissolved in (a) 65EHA35EtOH and (b) o-xylene. The respective absorption spectra are shown to assess the impact of the spectral overlap on the fluorescence spectra.
Fig. 2. Peak-normalized absorption and fluorescence spectra of coumarin 152 dissolved in (a) 65EHA35EtOH and (b) o-xylene at various concentrations. Excitation wavelength: 355 nm indicated with a purple line; temperature: 303 K. The dashed area shows the overlap $O$ (unit: nm) between normalized absorption and emission for the 0.1 mg/l case.
There is no change of the fluorescence spectra and hence no re-absorption measurable for coumarin 152 dissolved in 65EHA35EtOH (same with EtOH and 1-butanol—refer to Supplement 1). A minimal re-absorption effect is measurable in the case of coumarin 152 dissolved in o-xylene, apparently because of the substantially smaller Stokes shift (right panel in Fig. 2). In contrast to the other solvents (cf. Fig. 3), this smaller Stokes shift leads to a larger spectral overlap (the overlapping area $O$ for the normalized spectra for 0.1 mg/l solutions $O = {3.34}\;{\rm{nm}}$, which is a factor of 2.37 larger than the overlap in 65EHA35EtOH with $O = {1.41}\;{\rm{nm}}$), which increases re-absorption of fluorescence light. For the concentration range (factor 100) investigated in this work and for path lengths of 5 mm, it can be stated that coumarin 152 as tracer for 2cLIF in the named solvents is not measurably affected by fluorescence signal re-absorption; this is shown in some more detail in Section 5 by simple model calculations using the Beer–Lambert law.
Fig. 3. Peak-normalized absorption (dashed lines) and fluorescence (solid lines) spectra of 1 mg/l coumarin 152 dissolved in various solvents, where the concentration in water due to poor solubility not exactly known; excitation: 355 nm indicated with a purple line; temperature: 303 K.
3. Impact of Various Solvents
Figure 3 shows the peak-normalized absorption and fluorescence spectra of coumarin 152 dissolved in various solvents. The absorption spectra are plotted as dashed lines, and the fluorescence spectra are plotted as solid lines. In all cases, the excitation wavelength was 355 nm, which is indicated as a purple line.
As expected, the solvents have a strong influence on the spectral location and half-width of all spectra, particularly the fluorescence spectra, e.g., the fluorescence spectrum in o-xylene at 303 K features its peak at 458.5 nm with a FWHM of 72.1 nm, the peak in water is at 534 nm with a FWHM of 99.6 nm. An overview of the respective values of coumarin 152 dissolved in various solvents is given in Table 2.
The peak positions and FWHM of this tracer in the solvents 65EHA35EtOH, EtOH, and 1–butanol are similar, while the peak position for o-xylene and water shifts to shorter and longer wavelengths, respectively. The FWHM for o-xylene and water is narrower and broader, respectively, than in the other three mentioned solvents. Except for o-xylene, where the fluorescence intensity increases with temperature, the intensity for the other solvents decreases within the considered temperature range (see Supplement 1). In water, the blueshift and the spectral broadening of coumarin 152 is the strongest, while it is the smallest in o-xylene.
Fig. 4. Temperature-dependent absorption (up to 373 K) and fluorescence spectra (up to 373 K) of 1 mg/l coumarin 153 in 65EHA35EtOH, with excitation at 355 nm indicated with a purple line. Top: absolute values with the inset illustrating the peak intensities versus temperature; bottom: peak-normalized values, optimized 10 nm 2cLIF detection bands (blue and red rectangles), and $\beta$ factor for the spectra at 303 and 373 K.
B. Coumarin 153
1. Temperature Dependence of Absorption and Fluorescence Spectra
In Fig. 4, the absorption and fluorescence spectra of coumarin 153 dissolved in 65EHA35EtOH and excited at 355 nm are shown up to the temperature of 373 K.
As for coumarin 152, it can be observed that coumarin 153 in 65EHA35EtOH has a large Stokes shift resulting in a corresponding small overlap of absorption and fluorescence spectra. In contrast to coumarin 152, the absolute fluorescence intensity of coumarin 153 increases with increasing temperature (see inset in Fig. 4), while the temperature-dependent absorption spectra behave very similar to the ones of coumarin 152. The absorption decreases with temperature up to 363 K, from where it starts to strongly increase again. Absorption and fluorescence spectra exhibit a significant blueshift and a low spectral broadening (Table 3). The $\beta$ factors start to rise from ${-}{{1250}}\;{\rm{K}}$ at 470 nm to about ${-}{{200}}\;{\rm{K}}$ in the range between 580 and 660 nm.
2. Impact of Concentration: Signal Re-Absorption
The concentration dependence of the fluorescence spectra of coumarin 153 dissolved in 65EHA35EtOH are presented in Fig. 5. It can be seen that the spectral overlap ($O = {6.03}\;{\rm{nm}}$) causes a slight spectral change of the shape in the range of the overlap. In regions where no overlap exists, no difference in the shape or position of the spectra is observable.
Table 3. Characterization of the Fluorescence Spectra of 1 mg/l Coumarin 153 Dissolved in Various Solvents, Excitation at 355 nma
Fig. 5. Peak-normalized absorption and fluorescence spectra of coumarin 153 dissolved in 65EHA35EtOH at various concentrations. Excitation wavelength: 355 nm indicated with a purple line; temperature: 303 K. The dashed area shows the overlap $O$ (unit: nm) between normalized absorption and emission for the 0.1 mg/l case.
3. Impact of Various Solvents
Figure 6 shows the absorption and fluorescence spectra of coumarin 153 dissolved in o–xylene, 65EHA35EtOH, and EtOH. As for coumarin 152, the spectra are significantly affected by the kind of solvent, e.g., a significant blueshift of the spectrum is observed for coumarin 153 dissolved in o-xylene.
Fig. 6. Peak-normalized absorption (dashed lines) and fluorescence (solid lines) spectra of 1 mg/l coumarin 153 dissolved in various solvents; excitation: 355 nm indicated with a purple line; temperature: 303 K.
The FWHM of the fluorescence of coumarin 153 in o-xylene is with 76.5 nm, much smaller than for the other two solvents presented in Table 3. The absolute fluorescence intensity increases in EtOH with 0.52 %/K, significantly stronger than in the other solvents. The blueshift of coumarin 153 is smaller in comparison to coumarin 152, and, in o-xylene, almost no spectral broadening is observable.
C. Rhodamine B
1. Temperature Dependence of Absorption and Fluorescence Spectra
Figure 7 presents the absorption and fluorescence spectra of rhodamine B dissolved in 65EHA35EtOH. Absorption spectra were measured up to 373 K and the fluorescence spectra up to 393 K. The excitation wavelength of 532 nm is indicated as a green line.
Fig. 7. Temperature-dependent absorption (up to 373 K) and fluorescence spectra (up to 393 K) of 1 mg/l rhodamine B in 65EHA35EtOH, with the excitation at 532 nm indicated with a green line. Top: absolute values with the inset illustrating the peak intensities versus temperature; bottom: peak-normalized values, optimized 10 nm 2cLIF detection bands (blue and red rectangles), and $\beta$ factor for the spectra at 303 and 373 K.
Rhodamine B exhibits a small Stokes shift and hence a large overlap of absorption and fluorescence spectra. In comparison to the coumarin tracers, the overlap is substantially larger. The fluorescence intensity decreases strongly with increasing temperature, and the absorption decreases moderately while the slope corresponds (see inset). From the lower plot in Fig. 7, it can be observed that in the range of shorter wavelengths the fluorescence spectra show a blueshift, which intensifies with temperature, while for the range of longer wavelengths a slight redshift is observable. Hence, with increasing temperature, the spectra exhibit a spectral broadening, which magnifies at higher temperatures. The $\beta$ factors range between 0 and 1750 K, with the minimum located at the short wavelength edge of the measured spectral region, and the maximum at 580 nm.
2. Impact of Concentration: Signal Re-Absorption
The normalized absorption and fluorescence spectra of rhodamine B at different concentrations and a temperature of 303 K are presented in Fig. 8. Different from the coumarin tracers, absorption and fluorescence spectra significantly overlap, causing an apparent peak shift due to an increase of re-absorption (5 mm path length). Therefore, in this case, the fluorescence spectra were normalized at a wavelength where this overlap is negligible, i.e., at 630 nm.
Fig. 8. Peak-normalized absorption spectrum and at 630 nm normalized fluorescence spectra of rhodamine B dissolved in 65EHA35EtOH at various concentrations. Excitation wavelength: 532 nm indicated with a green line; temperature: 303 K. The dashed area shows the overlap $O$ (unit: nm) between normalized absorption and emission for the 0.1 mg/l case.
As a result of the small Stokes shift and hence the large overlap of the absorption and fluorescence spectra, rhodamine B is strongly affected by signal re-absorption. It can be seen from the diagram that the disparities of the fluorescence spectra correspond to the spectral overlap area ($O = {{16}}\;{\rm{nm}}$). Regions where no spectral overlap exists have no measurable divergence to each other. In practical applications where re-absorption comes into effect either because of strong concentration changes or sizeable fluorescence signal path length variations, the 2cLIF measurement accuracy of rhodamine B significantly declines. The effect can be illustrated by a less application-related extreme case of a concentration change from 0.1 to 10 mg/l at a path length of 5 mm: the related spectral shift would cause a ratio change by a factor of 13.8, which corresponds to a spectral change caused by a temperature variation of ${\sim}{{850}}\;{\rm{K}}$ (at the optimized color band wavelengths presented in Fig. 7). A concentration change from 0.1 to 1 mg/l under the same conditions cause a ratio change by a factor of 1.27, which corresponds to a temperature difference of ${\sim}{{18}}\;{\rm{K}}$.
Fig. 9. Peak-normalized absorption (dashed lines) and fluorescence (solid lines) spectra of 1 mg/l rhodamine B dissolved in various solvents; excitation: 532 nm indicated with a green line; temperature: 303 K.
3. Impact of Various Solvents
Figure 9 shows the peak-normalized absorption and fluorescence spectra of rhodamine B dissolved in various solvents at a temperature of 303 K.
In comparison to the coumarin tracers, the rhodamine B absorption and fluorescence spectra are less affected by the kind of solvent. This can also be seen in Table 4 from the much smaller variation in peak position and FWHM than was measured for the coumarin tracers. Additionally, in all considered solvents, the fluorescence peak intensity decreases substantially with temperature. Rhodamine B exhibits the strongest and smallest blueshift when dissolved in EtOH and water, respectively. Except for 1-butanol, the spectral broadening moves in the same range across the investigated solvents.c
Table 4. Characterization of the Fluorescence Spectra of 1 mg/l Rhodamine B Dissolved in Various Solvents, Excitation at 532 nma
Table 5. Characterization of the Fluorescence Spectra of 1 mg/l DCM Dissolved in Various Solvents, Excitation at 532 nm, for o-Xylene Excitation at 355 nma
D. DCM
1. Temperature Dependence of Absorption and Fluorescence Spectra
The absorption and fluorescence spectra of DCM dissolved in 65EHA35EtOH at different temperatures are depicted in Fig. 10. DCM dissolved in pure EtOH and 65EHA35EtOH can be excited in the UV at 355 nm as well as in the VIS at 532 nm, whereas it is only excitable at 355 nm when dissolved in o-xylene because the absorption spectra are strongly shifted to shorter wavelengths.
Fig. 10. Temperature-dependent absorption (up to 373 K) and fluorescence spectra (up to 373 K) of 1 mg/l DCM in 65EHA35EtOH, with the excitation at 532 nm indicated with a green line. Top: absolute values with the inset illustrating the peak intensities versus temperature; bottom: peak-normalized values, optimized 10 nm 2cLIF detection bands (blue and red rectangles), and $\beta$ factors for the spectra at 303 and 373 K.
For a tracer excitable at 532 nm, DCM shows a comparatively large Stokes shift. As depicted in the inset of Fig. 10 the fluorescence intensity decreases strongly with increasing temperature, while the absorption spectra decrease from room temperature to 333 K and then stay almost constant. The normalized spectra show a distinct blueshift, which is stronger for the absorption than for the fluorescence spectra. In comparison to the other investigated tracers, DCM exhibits a notably stronger blueshift of the absorption spectra. As a result of the strong decrease in fluorescence intensity with increasing temperature, the $\beta$ factor features comparatively high values (between 2000 and 4500 K). The lower $\beta$ factor are located at shorter wavelengths, while the highest can be found at around 630 nm.
2. Impact of Concentration: Signal Re-Absorption
Figure 11 shows the fluorescence spectra of DCM at 303 K for various concentrations. Since the spectral overlap between absorption and fluorescence spectra is not reaching out to the fluorescence peak, the fluorescence spectra were normalized at their respective peak positions.
Fig. 11. Peak-normalized absorption and fluorescence spectra of DCM dissolved in 65EHA35EtOH at various concentrations. Excitation wavelength: 532 nm indicated with a green line; temperature: 303 K. The dashed area shows the overlap $O$ (unit: nm) between normalized absorption and emission for the 0.1 mg/l case.
Fig. 12. Peak-normalized absorption (dashed lines) and fluorescence (solid lines) spectra of 1 mg/l DCM dissolved in various solvents; excitation for 65EHA35EtOH and EtOH: 532 nm indicated with a green line; excitation for o-xylene: 355 nm indicated with a purple line; temperature: 303 K.
It can be seen from the diagram that the variation of concentration only slightly affects the fluorescence spectra in the region of the spectral overlap, which is the smallest of the investigated tracers excitable at 532 nm. The same behavior is observed for the other investigated solvents—refer to Supplement 1. Since the overlap area coincides with the excitation wavelength, only the half-overlap area up to the intersection of absorption and fluorescence spectrum can be calculated (${O_{\rm intersec.}} = {1.86}\;{\rm{nm}}$). Absorption and fluorescence spectra can be considered as mirror images, which approximately allows an estimation of the spectral overlap area by doubling the calculated value resulting in $O = {3.72}\;{\rm{nm}}$.
3. Impact of Various Solvents
The absorption and fluorescence spectra of DCM dissolved in EtOH, 65EHA35EtOH, and o–xylene are presented in Fig. 12. Two excitation wavelengths are indicated because DCM dissolved in o-xylene can only be excited at 355 nm among the harmonics of the Nd:YAG laser.
Also, for this tracer, the fluorescence spectrum is strongly affected by the solvent (the shift of the fluorescence color with different solvents is visible by the naked eye). The trends in spectral shift of the absorption spectra with solvent are the same as for the fluorescence spectra.
From Table 5 it can be learned that, among the investigated solvents, the decrease of the fluorescence peak intensity of DCM with temperature is the smallest in EtOH. In o-xylene, DCM has a significantly larger FWHM than the others, exhibits the largest spectral broadening, and has the smallest blueshift.
E. Pyrromethene 597
1. Temperature Dependence of Absorption and Fluorescence Spectra
Figure 13 presents the absorption and fluorescence spectra of pyrromethene 597 dissolved in 65EHA35EtOH. From the tracers investigated in this work, pyrromethene 597 has the strongest absorption at 532 nm.
Fig. 13. Temperature-dependent absorption (up to 373 K) and fluorescence spectra (up to 373 K) of 1 mg/l pyrromethene 597 in 65EHA35EtOH, with excitation at 532 nm indicated with a green line. Top: absolute values with the inset illustrating the peak intensities versus temperature; bottom: peak-normalized values, optimized 10 nm 2cLIF detection bands (blue and red rectangles), and $\beta$ factor for the spectra at 303 and 373 K.
In comparison to DCM, pyrromethene 597 exhibits a smaller Stokes shift, resulting in a larger spectral overlap. With increasing temperature, the fluorescence intensity of pyrromethene 597 in 65EHA35EtOH decreases steadily. From the peak-normalized spectra, it can be seen that the spectra exhibit almost no blueshift in the wavelength range left of the peak, while in the range of longer wavelengths a redshift is observable, which leads to a distinct spectral broadening. The absorption spectra show no significant change with rising temperature. The peak of the absorption spectrum first decreases with temperature and then approximately stays constant (see inset in Fig. 13). From the short wavelength edge of the fluorescence spectra, the $\beta$ factor rises to a peak value of 2350 K at about 560 nm and decreases to 1600 K at the long wavelength edge.
2. Impact of Concentration: Signal Re-Absorption
Fluorescence spectra of pyrromethene 597 at concentrations between 0.1 and 10 mg/l are presented in Fig. 14. They were normalized at 600 nm, where no overlap of absorption and fluorescence spectra exists.
Fig. 14. Peak-normalized absorption and at 600 nm normalized fluorescence spectra of pyrromethene 597 dissolved in 65EHA35EtOH at various concentrations. Excitation wavelength: 532 nm indicated with a green line; temperature: 303 K. The dashed area shows the overlap $O$ (unit: nm) between normalized absorption and emission for the 0.1 mg/l case.
It can be observed that due to fluorescence re-absorption, especially the spectrum with 10 mg/l, an apparent peak shift with reduced intensity in the overlap region is experienced. Similar effects regarding signal re-absorption were measured for the other investigated solvents. For the same reason as for DCM, only the overlap area up to the intersection of absorption and fluorescence spectrum can be calculated (${O_{\rm intersec.}} = {4.32}\;{\rm{nm}}$), resulting in an estimation of $O = {8.64}\;{\rm{nm}}$.
3. Impact of Various Solvents
Figure 15 presents the peak-normalized absorption and fluorescence spectra of pyrromethene 597 at 303 K dissolved in various solvents with a concentration of 1 mg/l.
Fig. 15. Peak-normalized absorption (dashed lines) and fluorescence (solid lines) spectra of 1 mg/l pyrromethene 597 dissolved in various solvents; excitation: 532 nm indicated with a green line; temperature: 303 K.
As a unique property among the tracers investigated here, the fluorescence spectra of pyrromethene 597 do not change significantly in shape and position with the solvents investigated. The absorption spectra in EtOH and its mixture with EHA are almost indistinguishable, while the one in o-xylene is only slightly displaced and somewhat narrower (Table 6).
Table 6. Characterization of the Fluorescence Spectra of 1 mg/l Pyrromethene 597 Dissolved in Various Solvents, Excitation at 532 nma
While the spectral shape and position of the fluorescence spectra are nearly the same in all solvents, the peak intensity decreases for o-xylene somewhat less with temperature. The blueshift and the spectral broadening stay approximately constant among the investigated solvents.
5. FLUORESCENCE SIGNAL RE-ABSORPTION IN EVAPORATING DROPLETS: TEMPERATURE MEASUREMENT ERROR
To assess the impact of fluorescence signal re-absorption for 2cLIF temperature measurements for applications in, e.g., spray flames, we calculated the measurement error caused by re-absorption in spherical evaporating EtOH droplets of different initial diameter and tracer concentration. The calculations are an extension of our prior model calculations presented in [33] and are based on the Lambert–Beer law, as given in Eq. (5):
with the detected fluorescence intensity, ${I_{\rm{fluo}}}$, the initial fluorescence intensity emitted from the probe volume, ${I_{{\rm fluo},0}}$, the molar attenuation coefficient, ${\varepsilon _\lambda}$, concentration, $c$, of the tracer, and the path length, $d$, the fluorescence signal travels through the tracer-doped liquid to the detector. For the initial fluorescence intensity and the molar attenuation coefficients, measured spectra were used, which were verified by data given by Brackmann [44]. Neglecting reflections and stratifications within the droplet, the average fluorescence signal path length through the liquid was assumed as the droplet radius. Considering a tracer-doped EtOH droplet in a spray flame, with progressing evaporation, its radius decreases, and hence the tracer concentration increases. Originating from the fact that the concentration increases significantly stronger than the radius of the droplet decreases, the fluorescence signal re-absorption increases with progressing droplet evaporation (cf. Eq. (5)). This would cause an increasingly lower detected fluorescence intensity in chosen 2cLIF color bands, which are located in the spectral range of the overlap region of absorption and fluorescence spectrum (see, e.g., Fig. 7 for the case of rhodamine B). With progressing droplet evaporation, this consequently leads to increasingly changing color band intensity ratios and hence to increased temperature measurement error.To achieve a sufficient signal-to-noise ratio in measurements, different tracer concentrations are necessary in dependence of its properties. For this evaluation, adjustment of tracer concentrations to compensate for different FQYs was not taken into consideration; the temperature dependence of FQY and possible color band widths (trade-off with good temperature sensitivity) were also not included. Furthermore, reflections within the droplet and additional droplets on the path between the considered droplet and the detector were not taken into account. The calculation results of rhodamine B (large spectral overlap) and coumarin 152 (small spectral overlap) are presented in Fig. 16. The measurement error caused by fluorescence signal re-absorption is plotted against the droplet diameter. Various application-related initial droplet diameters (15–100 µm) [45] and tracer concentrations (1–20 mg/l) were considered. As an indication, the dashed lines represent tracer concentrations exceeding 50 g/l.
Fig. 16. Measurement error caused by fluorescence signal re-absorption versus droplet diameter, with the initial droplet diameter and the initial tracer concentration given in the legend, for (a) rhodamine B and (b) coumarin 152 dissolved in EtOH. Assumed detection bandwidths: 10 nm; dashed lines: tracer concentration exceeding 50 g/l.
It can be seen from the diagrams that for fixed start parameters of droplet size and tracer concentration the ratio-based temperature error increases exponentially with decreasing droplet diameter. Larger initial droplet diameters and higher initial tracer concentrations cause larger measurement errors during the droplet evaporation process. The impact of fluorescence signal re-absorption on the measurement accuracy is significantly larger for rhodamine B in comparison to coumarin 152 originating from the larger spectral overlap of absorption and fluorescence spectra, which includes regions where the optimized color detection channels are located. Three-color ratio techniques [36] have been developed to circumvent this diagnostic deficiency, but are experimentally more involved. In order to minimize the measurement error caused by fluorescence singal re-absoprtion within evaporating droplets in general, a low tracer concentration (e.g., achievable by high FQYs) and tracers with small spectral overlaps are beneficial.
Figure 17 shows the developing temperature error for the tracers investigated in this work for fixed initial droplet conditions, where the tracers are dissolved with a concentration of 10 mg/l in an EtOH droplet with a diameter of 50 µm. It can be seen that the coumarin tracers exhibit the smallest measurement error due to re-absorption and rhodamine B the highest. Among the tracers excitable at 532 nm, pyrromethene 597 shows the smallest error.
Fig. 17. Measurement error caused by fluorescence signal re-absorption versus droplet diameter; initial droplet diameter: 50 µm; initial tracer concentration: 10 mg/l, with the chosen tracers dissolved in EtOH. Optimized detection bandwidth: 10 nm; dashed lines: tracer concentration exceeding 50 g/l.
6. VARIATION OF SOLVENT COMPOSITION
In this section, we investigate the effect on the spectra when chosen tracers are dissolved in a solvent mixture composed of two components of varying amounts typically present in spray-based nanoparticle synthesis flames, where 2cLIF liquid-phase temperature measurements may be applied. Related research to this topic was published by [46].
A. Coumarin 152 and Pyrromethene 597 in EHA/Ethanol
The mixture of 65EHA35EtOH is commonly being used as a solvent for metal-nitrates precursors in spray-flame-based nanoparticle synthesis [4,32]. To measure the liquid-phase temperature of this evaporating fuel composition, a tracer is desired, whose fluorescence spectrum is not affected by preferential evaporation processes inside this two-component mixture. Since the vapor pressure of EtOH is significantly higher than that of EHA (EHA: ${\lt}{0.01}\;{\rm{hPa}}$, EtOH: 59 hPa at 20°C), a strong increase in the volume fraction of EHA is expected during evaporation. Figure 18 shows the fluorescence spectra of coumarin 152 and pyrromethene 597 dissolved in mixtures of various EHA/EtOH fractions.
Fig. 18. Normalized fluorescence spectra of (a) 10 mg/l coumarin 152 and (b) 1 mg/l pyrromethene 597 dissolved in solvent mixtures of EtOH and EHA at various fractions, and the fractions are displayed by volume; excitation: (a) 355 nm, (b) 532 nm; temperature: 298 K.
As can be seen from the diagrams, a change in solvent composition of EHA/EtOH causes a strong blueshift for coumarin 152 with increasing fractions of EHA, whereas pyrromethene 597 is not detectably affected (see inset). From these results, it can be concluded that coumarin 152 is not a suitable temperature tracer for this two-component solvent, since the described spectral shift is clearly predominant compared to the spectral shift caused by temperature changes. In contrast, pyrromethene 597 is well-suited for this application case. Unfortunately, this tracer exhibits a comparatively low temperature sensitivity (cf., Section 8.A) and is more affected by fluorescence signal re-absorption than the coumarin tracers.
B. Coumarin 152 in Ethanol/HMDSO
For the spray-flame synthesis of silica nanoparticles, the precursor HMDSO is dissolved in EtOH and provides the liquid mixture in the spray combustion process [47]. Because of the slightly different vapor pressures (HMDSO: 44 hPa, EtOH: 59 hPa at 2°C), a variation of the fractional composition during the combustion process cannot be excluded. A comparison of the resulting absorption and fluorescence spectra for 10% and 50% mixtures in EtOH is shown in Fig. 19.
Fig. 19. Absorption and fluorescence spectra of 10 mg/l coumarin 152 dissolved in solvent mixtures of EtOH and HMDSO at various fractions, and the fractions are displayed by volume; excitation: 355 nm; temperature: 303 K.
With an increasing amount of HMDSO, the absorption and fluorescence spectra show a blueshift. In consequence of partial evaporation of one of the components during the heat-up phase of the solution in a spray flame, the spectral shift with solvent composition will affect the accuracy of the 2cLIF temperature measurements. For case considered here, a change in the solvent fraction from 90% EtOH and 10% HMDSO to 50% EtOH and 50% HMDSO would cause a ratio change by a factor of 1.8 (using the optimized color band positions for the first mentioned case). This would correspond to a temperature bias of 43.7 K.
7. PEAK AND CENTROID SHIFT OF FLUORESCENCE SPECTRA
A useful parameter that characterizes the behavior of the fluorescence spectra of the various investigated tracers with temperature is the spectral shift of the spectrum peak or its centroid. In Fig. 20, we show the spectral shift $\Delta {\lambda _i} = {\lambda _i}(T) - {\lambda _i}({T_0})$ of the positions of the fluorescence peaks and centroids ($i$ stands for peak and centroid, respectively) with respect to the reference temperature ${T_0}$ of 303 K. In diagrams (a) and (b) of Fig. 20, these results are shown for all investigated tracers dissolved in 65EHA35EtOH and o-xylene, respectively, while in diagrams (c) and (d) a comparison is made for coumarin 152 and rhodamine B dissolved in various solvents. To guide the eye through the data points, the centroid and peak positions were fitted with polynomials and plotted as solid and dashed lines, respectively.
Fig. 20. Shift of peak position and centroid [$\Delta {\lambda _i} = {\lambda _i}(T) - {\lambda _i}({T_0} = {{303}}\;{\rm{K}})$] of all investigated tracer fluorescence spectra (1 mg/l) dissolved in (a) 65EHA35EtOH, (b) o-xylene, (c) coumarin 152, and (d) rhodamine B in various solvents; dashed lines: fit of peak positions; solid lines: fit of centroid positions.
For a specific tracer, the peak and centroid positions vary with temperature similarly, indicating that the shape of the spectra does not change much with temperature during a blue- or redshift (negative and positive sign, respectively). An exception is DCM in o-xylene, where the peak shifts to longer wavelengths, whereas the centroid shifts to shorter wavelengths, indicating a change in the spectral shape. While coumarin 152 exhibits the strongest blueshift, pyrromethene 597 shows a redshift in all solvents. There is no general influence of the solvents on the shift direction or determinable magnitude. While pyrromethene 597 shifts in the investigated solvents to longer wavelengths, coumarin 152 shifts to shorter wavelengths independently of the used solvent [cf. Fig. 20(c), lower left]. Furthermore, for coumarin 152, the solvent has only a slight influence on the trends in peak and centroid shifts, whereas this is not the case for rhodamine B (lower graph); for instance, EtOH causes a blueshift and water a redshift, respectively.
8. TEMPERATURE SENSITIVITY OF 2cLIF INTENSITY RATIOS
Based on the measured fluorescence spectra, a systematic evaluation towards the best temperature sensitivity for liquid-phase 2cLIF thermometry based on intensity ratios was conducted using a Matlab [48] routine under the constraint that the recorded intensity in each of the two-color channels exceeds 10% of the fluorescence peak intensity of the respective tracer/solvent combination. For two fixed filter bandwidths (10 and 40 nm), all possible combinations of color band positions were considered to maximize the calculated intensity ratio. For more detailed information, refer to [33]. The results are presented in Arrhenius-type plots as logarithmic intensity ratios normalized for the respective ratio at 303 K against the difference of inverse temperatures (Eq. 3), where the steepness of the slope is a measure for the temperature sensitivity of the respective temperature tracer. Exponential fits of the ratios are plotted as solid lines. In this section, results are shown for the chosen tracers dissolved in 65EHA35EtOH and o-xylene. The temperature sensitivities of most of the tracers dissolved in pure EtOH were investigated in our previous work [33]. Overall, coumarin 152 exhibits the strongest slope of the intensity ratio, i.e., temperature sensitivity for all investigated solvents.
A. 65EHA35EtOH
For the assumed filter transmission bandwidths of 10 and 40 nm, the respective band positions resulting in the highest temperature sensitivity of the intensity ratio are shown in Table 7. In view of the fact that the ratios do not change linearly, the ratio change represents the linear interpolated change for the respective temperature intervals averaged over the temperature range from 303 to 373 K in percent per Kelvin.
Table 7. Optimized Detection Band Positions for 2cLIF Thermometry, Concentration: 1 mg/l in 65EHA35EtOH, Excitation Wavelength, Resulting Best Spectral Band Position Combinationsa
With regard to the calculated optimized color band positions, it is observed that these do not change significantly with the choice of the pre-defined bandwidth. There is a general range for the optimal color band positions (usually the outer region of a spectrum), which is limited to the inner spectral regions due to the pre-defined minimal color band intensity.
As it can be seen from Fig. 21, coumarin 152 shows the highest temperature sensitivity for both color bandwidths. Because of the much narrower fluorescence spectrum, the temperature sensitivity of rhodamine B is significantly higher for bandwidths of 10 nm. In the case of smaller bandwidths, the best spectral ranges can be selected more specifically, whereas in the case of broader bandwidths the color band inevitably includes regions that are less temperature sensitive. Pyrromethene 597 shows the overall lowest temperature sensitivity. Coumarin 153, DCM, and Pyrromethene 597 show an almost linear behavior in the temperature range investigated here, whereas in the cases of coumarin 152 and rhodamine B clearly a strong non-linear behavior can be observed.
Fig. 21. Calculated fluorescence intensity ratios (symbols) for various tracers dissolved in 65EHA35EtOH, 1 mg/l, plotted as logarithmic ratios normalized at ${T_0} = {{303}}\;{\rm{K}}$ versus inverse temperature difference; solid line: exponential fit of intensity ratios; assumed spectral bandwidths: (a) 40 nm and (b) 10 nm; the chosen center wavelength positions of the “blue” and “red” filter combinations are listed in Table 7; depicted temperature range: 298–393 K.
B. o-Xylene
For all tracers, the fluorescence spectra in o-xylene were measured from room temperature up to 393 K. Rhodamine B is excluded in this investigation because of its poor solubility in o-xylene. As already mentioned with Fig. 12, DCM was excited with 355 nm due to its small absorption at 532 nm. The results are presented in Table 8.
Table 8. Optimized Detection Band Positions for 2cLIF Thermometry, Concentration: 1 mg/l in o-Xylene, Excitation Wavelength, Resulting Best Spectral Band Position Combinationsa
As Fig. 22 shows, coumarin 153 (for a bandwidth of 40 nm) and coumarin 152 (for a bandwidth of 10 nm) show the overall highest temperature sensitivity. It is also seen in the Arrhenius-type plots that for o-xylene (in contrast to 65EHA35EtOH) all investigated tracers show a close to linear increase of the normalized intensity ratio with temperature, i.e., constant temperature sensitivity. The temperature sensitivity of coumarin 153 is significantly higher than in 65EHA35EtOH, and pyrromethene 597 again shows the lowest temperature sensitivity of all tracers investigated here.
Fig. 22. Calculated fluorescence intensity ratios (symbols) for various tracers dissolved in o-xylene, 1 mg/l, plotted as logarithmic ratios normalized at ${T_0} = {{303}}\;{\rm{K}}$ versus inverse temperature difference; solid line: exponential fit of intensity ratios; assumed spectral bandwidths: (a) 40 nm and (b) 10 nm; the chosen center wavelength positions of the “blue” and “red” filter combinations are listed in Table 8; depicted temperature range: 298–393 K.
9. DISCUSSION
The present work shows that the best suitable tracer for 2cLIF in liquids will depend on the specific application. For measurements with single-component liquids, all investigated tracers can be used (coumarin 152, coumarin 153, DCM, pyrromethene 597, and rhodamine B). They exhibit different temperature sensitivities depending on the solvent. In general, coumarin 152 exhibits a comparatively high temperature sensitivity, rhodamine B only features a high temperature sensitivity when using a narrow blue color band, and pyrromethene 597 shows the lowest temperature sensitivity compared to the tracers presented in this work. Another issue for tracer suitability is the need for large absorption cross sections at available laser excitation wavelengths to provide sufficient fluorescence signals with a low needed concentration. With a view to the desired temperature measurements, DCM, pyrromethene 597, and rhodamine B have the disadvantage that their FQY strongly decreases with rising temperatures, which will necessitate higher tracer concentrations for measurements at higher temperatures. In contrast to that, the FQY of the coumarin tracers tendentially rises towards higher temperatures, although from a lower level compared with rhodamine B. The extent of fluorescence signal re-absorption is another factor that can affect the tracer performance for 2cLIF thermometry. Choosing tracers with a small overlap between absorption and fluorescence spectra reduces errors resulting from variations in tracer concentration and/or path length through tracer-doped liquid. We have shown in this work that the investigated coumarin tracers exhibit very low re-absorption, while rhodamine B is strongly affected due to the large overlap of its absorption and fluorescence spectra. To reduce fluorescence signal re-absorption, a preferably low tracer concentration is essential, which favors tracers with high FQY. Finally, in systems where the solvent consists of more than one component with different evaporation characteristics, a tracer needs to be unaffected by changing solvent composition. For the specific cases aimed for spray-flame synthesis of metal and silicon oxides, the results show that for the EHA/EtOH solvent mixture, coumarin 152 is not suitable since its spectral shift with a change in composition is dominant over that caused by the temperature variation. In this respect, pyrromethene 597 turned out to be much less affected by a compositional change and might be well-suited for this thermometry method; although, the comparatively low temperature sensitivity and sizeable signal re-absorption limits its applicability. It therefore is obvious that a global estimation of tracer performance is essential when choosing a suitable tracer for this 2cLIF temperature diagnostics method, taking into account all mentioned aspects under the specific limitations of the application case, before a meaningful temperature measurement can be performed.
10. CONCLUSIONS
In this work, the absorption and fluorescence spectra of the tracers coumarin 152, coumarin 153, DCM, pyrromethene 597, and rhodamine B were investigated for the impact of changes in temperature, concentration, excitation wavelength, type, and composition of the solvent. The tracers were chosen with respect to their excitability by harmonics of a Nd:YAG laser to enable single-shot measurements based on pulsed lasers. This study was made to assess the applicability of the laser dyes as tracer for liquid-phase 2cLIF temperature measurements. In this process, the focus was on temperature sensitivity, the effect of fluorescence signal re-absorption, and changes of solvent compositions on the measurement accuracy. The results show that the temperature sensitivity and the signal re-absorption effect strongly depend on the tracer and solvent. Composition variations of solvent mixtures can lead to spectral changes and hence to significantly decreasing measurement accuracies. On the other hand—provided that the temperature stays constant—the spectral changes due to different solvent compositions provide the potential for concentration measurements.
Among the investigated tracers, coumarin 152 was found to be best suited as tracer for 2cLIF in the examined single-component liquids. For the two-component solvent EHA/EtOH, pyrromethene 597 turned out to be best suited because its fluorescence spectra are independent towards composition changes. Coumarin 152 is soluble in all examined solvents (in water poor solubility) and exhibits the highest temperature sensitivity for all considered solvents. With respect to the transmissivity of the used solvents, it is excitable by Nd:YAG lasers at 266 or 355 nm, and just like the other tracers proved stable against temperatures up to at least 393 K. Compared to the other tracers, the detectable fluorescence spectra of coumarin 152 are significantly less affected by concentration changes or fluorescence path length variations in the liquid-phase (lowest fluorescence signal re-absorption). This leads to a significantly higher measurement accuracy for our desired application in spray flames.
Funding
Deutsche Forschungsgemeinschaft (DFG) SPP1980 (374463258).
Disclosures
The authors declare no conflicts of interest.
Data Availability
Data underlying the results presented in this paper are available in Ref. [49].
Supplemental document
See Supplement 1 for supporting content.
REFERENCES AND NOTE
1. W. Y. Teoh, R. Amal, and L. Mädler, “Flame spray pyrolysis: an enabling technology for nanoparticles design and fabrication,” Nanoscale 2, 1324–1347 (2010). [CrossRef]
2. A. Rittler, L. Deng, I. Wlokas, and A. Kempf, “Large eddy simulations of nanoparticle synthesis from flame spray pyrolysis,” Proc. Combust. Inst. 36, 1077–1087 (2017). [CrossRef]
3. F. Schneider, S. Suleiman, J. Menser, E. Borukhovich, I. Wlokas, A. Kempf, H. Wiggers, and C. Schulz, “SpraySyn—a standardized burner configuration for nanoparticle synthesis in spray flames,” Rev. Sci. Instrum. 90, 085108 (2019). [CrossRef]
4. S. Angel, J. Neises, M. Dreyer, K. Friedel Ortega, M. Behrens, Y. Wang, H. Arandiyan, C. Schulz, and H. Wiggers, “Spray-flame synthesis of La (Fe, Co) O3 nano-perovskites from metal nitrates,” AIChE J. 66, e16748 (2020). [CrossRef]
5. C. D. Rosebrock, T. Wriedt, L. Mädler, and K. Wegner, “The role of microexplosions in flame spray synthesis for homogeneous nanopowders from low-cost metal precursors,” AIChE J. 62, 381–391 (2016). [CrossRef]
6. H. Li, C. D. Rosebrock, Y. Wu, T. Wriedt, and L. Mädler, “Single droplet combustion of precursor/solvent solutions for nanoparticle production: optical diagnostics on single isolated burning droplets with micro-explosions,” Proc. Combust. Inst. 37, 1203–1211 (2019). [CrossRef]
7. C. Edwards and R. Rudoff, “Structure of a swirl-stabilized spray flame by imaging, laser doppler velocimetry, and phase doppler anemometry,” in Symposium (International) on Combustion (Elsevier, 1991), pp. 1353–1359.
8. H.-E. Albrecht, M. Borys, N. Damaschke, and C. Tropea, Laser Doppler and Phase Doppler Measurement Techniques (Springer, 2003).
9. P. Le Gal, N. Farrugia, and D. Greenhalgh, “Laser sheet dropsizing of dense sprays,” Opt. Laser Technol. 31, 75–83 (1999). [CrossRef]
10. M. Storch, Y. N. Mishra, M. Koegl, E. Kristensson, S. Will, L. Zigan, and E. Berrocal, “Two-phase SLIPI for instantaneous LIF and Mie imaging of transient fuel sprays,” Opt. Lett. 41, 5422–5425 (2016). [CrossRef]
11. P. Lavieille, F. Lemoine, G. Lavergne, and M. Lebouché, “Evaporating and combusting droplet temperature measurements using two-color laser-induced fluorescence,” Exp. Fluids 31, 45–55 (2001). [CrossRef]
12. P. Wieske, S. Wissel, G. Grünefeld, and S. Pischinger, “Improvement of LIEF by wavelength-resolved acquisition of multiple images using a single CCD detector–simultaneous 2D measurement of air/fuel ratio, temperature distribution of the liquid phase and qualitative distribution of the liquid phase with the multi-2D technique,” Appl. Phys. B 83, 323–329 (2006). [CrossRef]
13. K. Kohse-Höinghaus, R. S. Barlow, M. Aldén, and J. Wolfrum, “Combustion at the focus: laser diagnostics and control,” Proc. Combust. Inst. 30, 89–123 (2005). [CrossRef]
14. E. M. Adkins and J. H. Miller, “Extinction measurements for optical band gap determination of soot in a series of nitrogen-diluted ethylene/air non-premixed flames,” Phys. Chem. Chem. Phys. 17, 2686–2695 (2015). [CrossRef]
15. X. Liu, M. E. Smith, and S. D. Tse, “In situ Raman characterization of nanoparticle aerosols during flame synthesis,” Appl. Phys. B 100, 643–653 (2010). [CrossRef]
16. D. R. Snelling, K. A. Thomson, G. J. Smallwood, and Ö. L. Gülder, “Two-dimensional imaging of soot volume fraction in laminar diffusion flames,” Appl. Opt. 38, 2478–2485 (1999). [CrossRef]
17. H. A. Michelsen, C. Schulz, G. J. Smallwood, and S. Will, “Laser-induced incandescence: particulate diagnostics for combustion, atmospheric, and industrial applications,” Prog. Energy Combust. Sci. 51, 2–48 (2015). [CrossRef]
18. T. Dreier and C. Schulz, “Laser-based diagnostics in the gas-phase synthesis of inorganic nanoparticles,” Powder Technol. 287, 226–238 (2016). [CrossRef]
19. H. Rochlitz and P. Scholz, “Application of laser-induced fluorescence technique in a duct flow with one heated wall,” Exp. Fluids 59, 54 (2018). [CrossRef]
20. I. Düwel, J. Schorr, J. Wolfrum, and C. Schulz, “Laser-induced fluorescence of tracers dissolved in evaporating droplets,” Appl. Phys. B 78, 127–131 (2004). [CrossRef]
21. E. Berrocal, E. Kristensson, and L. Zigan, “Light sheet fluorescence microscopic imaging for high-resolution visualization of spray dynamics,” Int. J. Spray Combust. Dyn. 10, 86–98 (2018). [CrossRef]
22. M. Koegl, Y. N. Mishra, M. Storch, C. Conrad, E. Berrocal, S. Will, and L. Zigan, “Analysis of ethanol and butanol direct-injection spark-ignition sprays using two-phase structured laser illumination planar imaging droplet sizing,” Int. J. Spray Combust. Dyn. 11, 1–16 (2018). [CrossRef]
23. T. D. Fansler and S. E. Parrish, “Spray measurement technology: a review,” Meas. Sci. Technol. 26, 012002 (2015). [CrossRef]
24. D. Greszik, H. Yang, T. Dreier, and C. Schulz, “Laser-based diagnostics for the measurement of liquid water film thickness,” Appl. Opt. 50, A60–A67 (2011). [CrossRef]
25. Y. N. Mishra, F. Abou Nada, S. Polster, E. Kristensson, and E. Berrocal, “Thermometry in aqueous solutions and sprays using two-color LIF and structured illumination,” Opt. Express 24, 4949–4963 (2016). [CrossRef]
26. J. Palmer, M. A. Reddemann, V. Kirsch, and R. Kneer, “Applying 2D-2cLIF-EET thermometry for micro-droplet internal temperature imaging,” Exp. Fluids 59, 51 (2018). [CrossRef]
27. P. Lavieille, F. Lemoine, G. Lavergne, J. F. Virepinte, and M. Lebouché, “Temperature measurements on droplets in monodisperse stream using laser-induced fluorescence,” Exp. Fluids 29, 429–437 (2000). [CrossRef]
28. Y. N. Mishra, A. Yoganantham, M. Koegl, and L. Zigan, “Investigation of five organic dyes in ethanol and butanol for two-color laser-induced fluorescence ratio thermometry,” Optics 1, 1–17 (2020). [CrossRef]
29. M. Koegl, M. Pahlevani, and L. Zigan, “A novel approach for measurement of composition and temperature of N-decane/butanol blends using two-color laser-induced fluorescence of nile red,” Sensors 20, 5721 (2020). [CrossRef]
30. M. Koegl, C. Mull, Y. N. Mishra, S. Will, and L. Zigan, “Characterization of fuel/water mixtures and emulsions with ethanol using laser-induced fluorescence,” Appl. Opt. 59, 1136–1144 (2020). [CrossRef]
31. R. Strobel and S. E. Pratsinis, “Effect of solvent composition on oxide morphology during flame spray pyrolysis of metal nitrates,” Phys. Chem. Chem. Phys. 13, 9246–9252 (2011). [CrossRef]
32. M. F. B. Stodt, M. Gonchikzhapov, T. Kasper, U. Fritsching, and J. Kiefer, “Chemistry of iron nitrate-based precursor solutions for spray-flame synthesis,” Phys. Chem. Chem. Phys. 21, 24793–24801 (2019). [CrossRef]
33. M. M. Prenting, M. I. Bin Dzulfida, T. Dreier, and C. Schulz, “Characterization of tracers for two-color laser-induced fluorescence liquid-phase temperature imaging in sprays,” Exp. Fluids 61, 77 (2020). [CrossRef]
34. H. W. Ge, I. Düwel, H. Kronemayer, R. W. Dibble, E. Gutheil, C. Schulz, and J. Wolfrum, “Laser-based experimental and Monte Carlo PDF numerical investigation of an ethanol/air spray flame,” Combust. Sci. Technol. 180, 1529–1547 (2008). [CrossRef]
35. P. Dunand, G. Castanet, and F. Lemoine, “A two-color planar LIF technique to map the temperature of droplets impinging onto a heated wall,” Exp. Fluids 52, 843–856 (2012). [CrossRef]
36. P. Lavieille, A. Delconte, D. Blondel, M. Lebouché, and F. Lemoine, “Non-intrusive temperature measurements using three-color laser-induced fluorescence,” Exp. Fluids 36, 706–716 (2004). [CrossRef]
37. P. Dahiya, M. Kumbhakar, T. Mukherjee, and H. Pal, “Effect of protic solvents on twisted intramolecular charge transfer state formation in coumarin-152 and coumarin-481 dyes,” Chem. Phys. Lett. 414, 148–154 (2005). [CrossRef]
38. S. Nad, M. Kumbhakar, and H. Pal, “Photophysical properties of coumarin-152 and coumarin-481 dyes: unusual behavior in nonpolar and in higher polarity solvents,” J. Phys. Chem. A 107, 4808–4816 (2003). [CrossRef]
39. C. Würth, M. Grabolle, J. Pauli, M. Spieles, and U. Resch-Genger, “Relative and absolute determination of fluorescence quantum yields of transparent samples,” Nat. Protoc. 8, 1535–1550 (2013). [CrossRef]
40. K. Rurack and M. Spieles, “Fluorescence quantum yields of a series of red and near-infrared dyes emitting at 600−1000 nm,” Anal. Chem. 83, 1232–1242 (2011). [CrossRef]
41. J. Bañuelos Prieto, F. López Arbeloa, V. Martínez Martínez, T. Arbeloa López, and I. López Arbeloa, “Photophysical properties of the pyrromethene 597 dye: solvent effect,” J. Phys. Chem. A 108, 5503–5508 (2004). [CrossRef]
42. R. F. Kubin and A. N. Fletcher, “Fluorescence quantum yields of some rhodamine dyes,” J. Lumin. 27, 455–462 (1982). [CrossRef]
43. M. Galanin, A. Kutyonkov, V. Smorchkov, Y. P. Timofeev, and Z. Chizhikova, “Measurement of photoluminescence quantum yield of dye solutions by the Vavilov and integrating-sphere methods,” Opt. Spectrosc. 53, 405–409 (1982).
44. U. Brackmann, Lambdachrome Laser Dyes (Lambda Physik AG, 2000).
45. M. F. B. Stodt, J. Kiefer, and U. Fritsching, “Ethanol droplet formation, dynamics and combustion mode in the flame of the SpraySyn-nozzle,” Exp. Fluids 60, 125 (2019). [CrossRef]
46. C. Maqua, G. Castanet, F. Lemoine, N. Doué, and G. Lavergne, “Temperature measurements of binary droplets using three-color laser-induced fluorescence,” Exp. Fluids 40, 786 (2006). [CrossRef]
47. L. Mädler, H. K. Kammler, R. Mueller, and S. E. Pratsinis, “Controlled synthesis of nanostructured particles by flame spray pyrolysis,” J. Aerosol Sci. 33, 369–389 (2002). [CrossRef]
48. “MATLAB, version 7.10.0 (R2019b)” (The MathWorks, Inc., 2020).
49. M. Prenting, M. Shilikhin, T. Dreier, C. Schulz, and T. Endres,“Characterization of tracers for two-color laser-induced fluorescence thermometry of liquid-phase temperature in ethanol, 2–ethylhexanoic-acid/ethanol mixtures, 1-butanol, and o-xylene,” Zenodo, 5 May 2021, https://doi.org/10.5281/zenodo.4613143.
