1. Introduction

In recent years, fluorescence spectroscopy has emerged as an attractive technique for non-invasive, early diagnosis of cancer due to its sensitivity to subtle biochemical changes [1–5]. The static fluorescence spectra from tissue contain rich biochemical information. It is however difficult to extract this information from the measured spectra because it gets strongly modulated by the wavelength dependent absorption and scattering properties of tissue [6–7]. The bulk fluorescence spectra contain hidden biochemical information of tissue fluorophores as well as morphological information through the modulated wavelength dependent scattering and absorption properties of the tissue. Extraction of intrinsic fluorescence, by removing the distorting effects of scattering and absorption, allows one to separate the biochemical information from the morphological one and to, hence, develop optimal diagnostic algorithms for tissue diagnosis. Considerable efforts have therefore been made in the recent past to remove these distorting effects and to extract intrinsic line shape and intensity of fluorescence [7–14]. Attempts to eliminate interplay of absorption and scattering properties of tissue on the measured fluorescence have been based on a theoretical description of light propagation in a turbid medium, producing an analytical relationship between measured and intrinsic fluorescence. The Kubelka-Munk theory [7], Monte Carlo simulation of propagation of excitation and fluorescence light [10], diffusion theory [12] and photon migration theory [8,9,13,14], have been used to relate the intrinsic fluorescence of tissue fluorophores to the measured fluorescence in terms of other measurable parameters like the optical transport parameters of tissue (namely, scattering coefficient µ_s, reduced scattering coefficient ${{\mu }_{\text{s}}}^{/}$=µ_s(1-g), absorption coefficient µ_a, and anisotropy parameter g [15]) and the diffuse reflectance. These methods have been demonstrated to extract intrinsic fluorescence from the measured tissue fluorescence with reasonable accuracy. However, owing to the different levels of approximations involved in the theoretical frame works, the applicability range of many of these methods have been limited to the range of the visible and NIR wavelengths (500 nm–800 nm) for the optical parameters [µ_a≪${{\mu }_{\text{s}}}^{/}$] [8–12]. More recently, Muller et al [13] and Zhang et al [14], extending the earlier work of Wu et al [8], proposed a method based on photon migration approach that is applicable in the entire range of optical transport parameters of tissue, covering the ultra-violet (µ_a≥${{\mu }_{\text{s}}}^{/}$), visible and NIR (µ_a≪${{\mu }_{\text{s}}}^{/}$) wavelength regions. This method is based on concomitant measurement of fluorescence and diffuse reflectance from tissue and derivation of an analytical relationship between the intrinsic fluorescence, the measured fluorescence and the diffuse reflectance spectra of tissue by using a photon migration model. The motivation behind developing a technique for extracting intrinsic fluorescence from the measured tissue fluorescence in the range of optical transport parameters where µ_a~${{\mu }_{\text{s}}}^{/}$, is that many of the reported auto-fluorescence studies in tissues have been conducted with ultra-violet excited UV-visible fluorescence emission range (350 nm–700 nm) where the values for µ_a of tissues are considerably higher (due to strong absorption of different forms of hemoglobin at around 420 nm) and are comparable to the values of ${{\mu }_{\text{s}}}^{/}$ [13,14].

In this paper, we present an experimental approach to remove the distorting effect of scattering and absorption on intrinsic fluorescence of fluorophores embedded in a turbid medium such as biological tissue. The method is based on simultaneous measurement of polarized fluorescence and polarized elastic scattering spectra from a turbid medium. The polarized fluorescence normalized by the polarized elastic scattering spectra (in the wavelength range of fluorescence emission) is found to be free from the distorting effects of absorption and scattering properties of the medium. The applicability range of this technique to recover intensity and line shape information of intrinsic fluorescence has been investigated by carrying out studies on a series of tissue phantoms having different absorption and scattering properties. The results obtained in these studies show that this technique has the potential to recover intrinsic line shape and intensity information of fluorescence from fluorophores embedded in a scattering medium for the range of optical transport parameters typically found in biological tissue (even for the range of optical transport parameters where the value for µ_a is ~${{\mu }_{\text{s}}}^{/}$). We offer plausible explanation for the striking success of this method to recover intrinsic fluorescence from the measured fluorescence of a scattering medium even for the range of optical transport parameters where the values for ${{\mu }_{\text{s}}}^{/}$ and µ_a are comparable. The main advantage of this method would be the fact that since it is based on a simple normalization of the measured polarized fluorescence by the measured polarized elastic scattering spectra, it does not require any additional theoretical calculations to recover intrinsic fluorescence and can also be used in the range of optical transport parameters, µ_a ~${{\mu }_{\text{s}}}^{/}$, where most of the other methods fail to extract intrinsic fluorescence with reasonable accuracy.

2. Experimental methods

The tissue phantoms used in this study were prepared using known concentrations of Flavin Adenine Dinucleotide (FAD) (Sigma Chemicals, U.S.A.) as fluorophore, suspensions of monodisperse polystyrene microspheres having mean diameter of 1.07µm (Bangs Lab., U.S.A.) as scatterer and Protoporphyrin IX (Sigma Chemicals, U.S.A.) as absorber. Studies were carried out on samples prepared using different concentrations of fluorophores, scatterers and absorbers. The values for the reduced scattering coefficient (${{\mu }_{\text{s}}}^{/}$) of the samples at different wavelengths were calculated using Mie theory [16]. The absorption coefficients (µ_a) of the dilute solutions of FAD and Protoporphyrin IX were measured separately using a spectrophotometer before adding these to the microsphere suspension. The values for ${{\mu }_{\text{s}}}^{/}$ in the phantoms used were varied from 0.35–2.065 mm^-1, 0.32–1.9 mm^-1 and 0.32–1.88 mm^-1 at 440, 520 and 540 nm respectively. The values for µ_a in the phantoms used were varied from 0.11–1.06 mm^-1, 0.07–0.76mm^-1 and 0.09–0.92 mm^-1 at 440, 520 and 540 nm were varied from respectively. The value for the anisotropy parameter (g) of 1.07 µm diameter polystyrene microspheres suspension in water was comparable to the typical value of g found in tissue (typical values of g in tissue are in the range 0.8–0.99 [15]) and was in the range 0.91 to 0.93 for the wavelength range investigated in this study. The Porphyrins used in this study did not show notable fluorescence in the wavelength range 470–600 nm with 440 nm excitation, though it was observed to produce a weak red fluorescence, with peaks centered on 630 nm and 690 nm. Therefore, for the purpose of our study, this could be treated as pure absorber in the wavelength range of interest i.e. in the characteristic fluorescence band of FAD (470–600 nm).

A commercial spectrofluorometer (SPEX, Fluorolog 3, Model FL3-22) was used to record both polarized fluorescence and polarized elastic scattering spectra from the samples. The excitation light, from a 450W Xenon lamp having a spot size dimension of ~1mm×12mm was incident on the samples kept in a quartz cuvette with a path length of 10mm. The polarized fluorescence spectra (λ_EM=470 nm–700 nm) from the samples were recorded with 440 nm excitation. All the polarized fluorescence and polarized elastic scattering spectra were collected at 22.5° from the direction of incident excitation light to minimize the specular reflection from the surface. The band-pass for both the excitation and emission monochromators was 2 nm. The integration time was kept 0.2 second and the emission monochromator stepped through 1 nm while recording spectra. The polarized elastic scattering spectra were recorded by exciting the samples with white light (zero order grating) over the wavelength range 470–700 nm. The excitation polarizer was oriented vertically and both the polarized fluorescence and polarized elastic scattering spectra were recorded with emission polarizer oriented in horizontal [I_⊥(λ)] and vertical [I_‖(λ)] positions respectively.

The polarized fluorescence and polarized elastic scattering spectra were generated as

Here G(=I_HV/I_HH) is the ratio of the sensitivity of the instrument to the vertically and horizontally polarized light [17]. For measuring spectral dependence of G, very dilute solution (10 ppm) of glycogen in distilled water was taken in a quartz cuvette and synchronous scan over the wavelength range 440–700 nm was recorded with zero offset between excitation and emission monochromator. The excitation polarizer was kept horizontal and emission polarizer was placed at horizontal (I_HH) and vertical (I_HV) orientations for measurement of G. From the measured polarized fluorescence and polarized elastic scattering spectra, the intrinsic fluorescence (IF) was extracted following the proposed scheme.

Here the subscripts FL and ESS represent fluorescence and elastic scattering respectively.

3. Results and discussion

In Fig. 1, we show 440 nm excited unpolarized fluorescence spectra recorded from different tissue phantoms (shown as phantoms i to iv in the legend). The inset of the figure shows the fluorescence spectra normalized by intensity at 520 nm (peak intensity of the spectra), for the same samples. In these phantoms, the value for FAD concentration was kept fixed at 20 µM and the concentration of scatterers and absorbers were varied. The values of the reduced scattering coefficients (${{\mu }_{\text{s}}}^{/}$) and absorption coefficients (µ_a) in these phantoms were varied in the range 1.34 mm^-1 to 2.06 mm^-1 and 0.07 mm^-1 to 0.21 mm^-1 respectively. These values for ${{\mu }_{\text{s}}}^{/}$ and µ_a at 440 nm (the excitation wavelength), 520 nm and 540 nm are listed in Table 1. It can be seen from the figure that, for the same fluorophore concentration, with increasing values of absorption coefficient (µ_a), the fluorescence intensity (F.I.) decreased drastically. The maximum variation in intensity at 520 nm was found to be ~50% (the variation is calculated as [(${{\text{I}}^{520}}_{\text{max}}$-${{\text{I}}^{520}}_{\text{min}}$)/${{\text{I}}^{520}}_{\text{max}}$], ${{\text{I}}^{520}}_{\text{max}}$ and ${{\text{I}}^{520}}_{\text{min}}$ being the maximum and the minimum value of intensity at 520 nm for the samples having the same fluorophore concentration). The characteristics dips of porphyrin absorption at 540 nm and 580 nm are also clearly visible in the fluorescence spectral profile (see inset of the figure). This demonstrates how the fluorescence from fluorophores embedded in a turbid medium can lose intensity and line shape information of intrinsic fluorescence due to modulation by absorption and scattering properties of the medium.

 

Fig. 1. 440 nm excited unpolarized fluorescence spectra recorded from four different tissue phantoms having a fixed FAD concentration of 20 µM. The inset of the figure shows the fluorescence spectra normalized by intensity at 520 nm. The values of the reduced scattering coefficient ${{\mu }_{\text{s}}}^{/}$ and absorption coefficient (µ_a) at 440 nm and 540 nm for the four samples are listed in table 1. (F.I. stands for Fluorescence Intensity in all the subsequent figures).

Download Full Size | PDF

Table 1. Values of absorption and reduced scattering coefficients at different wavelengths

View Table | View all tables in this article

In Fig. 2(a), we show the 440 nm excited polarized fluorescence spectra (I_pol (λ)=[I_‖(λ)-G×I_⊥(λ)]) recorded from the same samples used in Fig. 1. The inset of the figure shows the peak intensity normalized fluorescence spectral profiles (fluorescence spectra normalized by intensity at 520 nm) for the samples. It is noticed that the characteristic dips of porphyrin absorption are not as enhanced in the polarized spectral profiles as compared to the unpolarized spectral profiles. The intensity at 520 nm also shows slightly lesser variation with changing absorber and scatterer concentration (~34%) as compared to unpolarized fluorescence. The reduced effect of absorption and scattering on the polarized fluorescence intensity and line shape arises due to the fact that the polarized fraction of total fluorescence (I_pol(λ)) suffers lesser number of scattering events, and has therefore traveled lesser path inside the turbid medium as compared to unpolarized fluorescence. Hence, absorption affects polarized fluorescence spectra to a lesser degree compared to unpolarized fluorescence. However, the effect of absorption and scattering on fluorescence line shape and intensity is not completely removed on using polarized fluorescence. It has been shown earlier that in a medium comprised of smaller sized scatterers (size a≪λ, g≤0.3), incident polarized light preserves it initial state of polarization up to ~ one transport mean free path (l_t=1/${{\mu }_{\text{s}}}^{/}$). In contrast in a medium comprised of larger sized scatterers (a≥λ, g≥0.7), owing to the forward scattering nature of the medium, polarization can be preserved up to several transport mean free paths [18,19]. This is the reason why, for the phantoms used in this study (g ~0.92 at 540 nm), polarized component of total fluorescence still bears some signature of the absorption properties of the medium and use of polarized fluorescence could not completely recover the intensity and line shape information of intrinsic fluorescence. Indeed, for samples prepared using 0.3 µm diameter polystyrene micro spheres suspension, for the similar range of ${{\mu }_{\text{s}}}^{/}$ and µ_a, use of polarized fluorescence yielded better results (data not shown here) than the ones shown in Fig. 2(a).

 

Fig. 2. (a) 440 nm excited polarized fluorescence spectra recorded from four different tissue phantoms having a fixed FAD concentration of 20 µM. The inset 2a shows the fluorescence spectra normalized by intensity at 520nm.

Download Full Size | PDF

 

Fig. 2. (b) Polarized elastic scattering spectra recorded from the same samples.

Download Full Size | PDF

In Fig. 2(b), the polarized elastic scattering spectra (I_pol (λ)=[I_‖(λ)-G×I_⊥(λ)]) recorded from the same samples are shown. The influence of changing absorber and scatterer concentrations on the spectral line shape and intensity of polarized elastic scattering and polarized fluorescence can be seen to be strikingly similar (see Figs. 2(a) and 2(b) for changing values of ${{\mu }_{\text{s}}}^{/}$ and µ_a). Based on this observation, we normalized the polarized fluorescence spectra by the polarized elastic scattering spectra for all the samples. The results are presented in Fig. 2(c). The corresponding peak normalized fluorescence spectral profiles and the spectral profile for a dilute solution of FAD (20µM) are shown in Fig. 2(d). The fluorescence spectral shapes and intensities show remarkable resemblance for all the samples. The intensity at 520 nm is seen to vary ≤10% only, for the four solutions with changing ${{\mu }_{\text{s}}}^{/}$ and µ_a but having the same FAD concentration (20µM). Similar results were obtained in phantoms with other concentrations of FAD for the range of values of µ_a/${{\mu }_{\text{s}}}^{/}$~0-0.14. These results indicate that polarized fluorescence spectra normalized by polarized elastic scattering spectra can recover the intrinsic line shape and intensity information of intrinsic fluorescence from a turbid medium for the range of optical transport parameters where the value for the ratio of µ_a/${{\mu }_{\text{s}}}^{/}$ are between 0 and 0.14, which is typical for tissue in the visible and NIR wavelength regions [15,20].

 

Fig. 2. (c) Polarized fluorescence spectra normalized by polarized elastic scattering spectra and 2 (d) the peak-normalized spectral line shapes (normalized by intensity at 520 nm) for the same spectra from the four samples. The values of the reduced scattering coefficients (${{\mu }_{\text{s}}}^{/}$) and absorption coefficients (µ_a) at 440 nm and 540 nm for the four samples are listed in Table 1.

Download Full Size | PDF

In the next step, we investigated the efficacy of this method to recover intrinsic fluorescence, where the concentration of absorber and scatterers were adjusted to yield comparable values of ${{\mu }_{\text{s}}}^{/}$ and µ_a in the phantoms. In Fig. 3(a), we show 440 nm excited unpolarized fluorescence spectra recorded from different samples with varying concentrations of FAD, scatterers and absorbers (shown as phantoms A to I in the legend). The values for ${{\mu }_{\text{s}}}^{/}$ and µ_a at 440 nm, 520nm and 540 nm and the concentrations of FAD used in these phantoms are listed in Table 2. Here, too, the modulation of spectral line shape by the characteristic absorption of porphyrins can be seen with the effect being more pronounced as compared to the results shown in Fig. 1. The intensity does not vary linearly with increasing concentration of fluorophores, but get strongly modulated by the absorption and scattering properties of the medium. For the set of phantoms with similar concentrations of fluorophores, the intensity of fluorescence can be seen to be reduced drastically with increasing value of µ_a (see, for example, for the two phantoms G and H having the same fluorophore concentration of 40 µM). In Fig. 3(b), the polarized fluorescence spectra from the same phantoms are shown (shown phantoms A to I in the legend). The inset of the figure shows the corresponding polarized elastic scattering spectra from the same samples. As has been observed earlier, though the polarized fluorescence spectral line shape shows slightly lesser modulation of the absorption properties of the medium as compared to unpolarized fluorescence, polarized fluorescence solely cannot be used to recover intensity and line shape information of intrinsic fluorescence from the fluorophores. In Fig. 3(c), we show the polarized fluorescence spectra normalized by polarized elastic scattering spectra, recorded from the same samples. The inset of the figure shows the corresponding peak intensity normalized fluorescence spectral profiles. The recovery of spectral line shape of intrinsic fluorescence of FAD can be seen to be quite satisfactory. Further, the intensity of fluorescence also shows consistent variation with increasing fluorophore concentrations used in these samples.

 

Fig. 3. (a) 440 nm excited unpolarized fluorescence spectra recorded from different tissue phantoms with varying concentration of FAD. The inset of the figure shows the fluorescence spectra normalized by intensity at 520 nm.

Download Full Size | PDF

Table 2. Values of absorption and reduced scattering coefficients at different wavelengths for different set of phantoms

View Table | View all tables in this article

All values shown here have an error ~2%

 

Fig. 3. (b) 440 nm excited polarized fluorescence spectra recorded from the same samples. The inset of the figure shows the polarized elastic scattering spectra recorded from the same phantoms A, C, D, E, F, G, H, I.

Download Full Size | PDF

 

Fig. 3. (c) Polarized fluorescence spectra normalized by polarized elastic scattering spectra for the same samples. The inset of the figure shows peak normalized spectral line shape (normalized by intensity at 520 nm). The values of the reduced scattering coefficients (${{\mu }_{\text{s}}}^{/}$) and absorption coefficients (µ_a) at 440 nm and 540 nm for the samples and the concentrations of FAD used in the samples are listed in Table 2.

Download Full Size | PDF

Figure 4(a) shows the normalized fluorescence spectra of pure FAD solution and the extracted intrinsic fluorescence spectrum of phantom that shows the emission peak of porphyrin at 630 nm. Figure 4(b) shows the variations in the measured fluorescence intensity and the recovered intrinsic fluorescence intensity at 520 nm with varying FAD concentration of the same set of phantoms for which the spectra are shown in Fig. 3. The linear fit shown in the figure clearly demonstrates that the recovered intrinsic fluorescence varies linearly with increasing concentration of fluorophores (the value of correlation factor, R², is 0.98 for this linear fit).

 

Fig. 4. (a) The normalized fluorescence spectrum of pure FAD with extract ed intrinsic fluorescence from phantom. Fig. 4 (b) Variation of measured and intrinsic fluorescence intensities at 520 nm with varying FAD concentration

Download Full Size | PDF

The intensities of fluorescence for the samples having the same fluorophore concentration but different absorber and scatterer concentrations are observed to be similar (see for phantoms C and D, E and F of Table 2). The maximum variation in intensity at 520 nm is found to be ~15% for the samples G and H, having the same fluorophore concentration of 40 µM. For these two samples the value for the ratio of µ_a/${{\mu }_{\text{s}}}^{/}$ was in the range 1.2–1.8. For the rest of the samples, with the ratio of µ_a/${{\mu }_{\text{s}}}^{/}$ in the range 0.32–1.2, the variation in intensity of fluorescence for same fluorophore concentration is ≤10%. It is pertinent to mention here that the unpolarized fluorescence normalized by unpolarized elastic scattering spectra for these samples could not recover the line shape and intensity information of intrinsic fluorescence (not shown here). It has earlier been demonstrated that a simple normalization of unpolarized fluorescence spectra by the unpolarized elastic scattering spectra cannot completely recover the line shape and, in particular, the intensity information of intrinsic fluorescence for the range of optical transport parameters where the value for µ_a is ~${{\mu }_{\text{s}}}^{/}$ [8,13,14]. This is due to the fact that owing to the isotropic nature of fluorescence emission and a difference in the values for optical transport parameters (µ_s, µ_a, and g) at the excitation and any emission wavelength, the scattering path of the fluorescence photons (at any emission wavelength λ_EM) in the medium would be different from that of the elastically scattered photons (at wavelength λ_EM). The results obtained in this study show that in contrast to this, the polarized fluorescence normalized by polarized elastic scattering could recover both the line shape and intensity information of intrinsic fluorescence with reasonable accuracy even for the range of optical transport parameters where the value for the ratio of µ_a/${{\mu }_{\text{s}}}^{/}$ were ~1.8. However, for even higher values of µ_a, i.e., for µ_a/${{\mu }_{\text{s}}}^{/}$>1.8, this proposed method could not recover intrinsic fluorescence from turbid medium with reasonable accuracy.

The results presented above clearly show that the proposed scheme of normalizing polarized fluorescence by polarized elastic scattering spectra recorded from a turbid medium can recover the intrinsic fluorescence of fluorophores embedded in a turbid medium with reasonable accuracy for the range of optical transport parameters where the ratio of µ_a/${{\mu }_{\text{s}}}^{/}$ lie between 0 and 1.8. The remarkable success of this method to recover intrinsic fluorescence even for comparable values of µ_a and ${{\mu }_{\text{s}}}^{/}$ of a turbid medium lies in the fact that the wavelength dependent absorption and scattering of the medium affects the polarized component of the fluorescence and the polarized component of elastically scattered light almost similarly. This may arise due to the fact that on recording the polarized component of fluorescence and elastically scattered light one would eliminate, to a great extent, the long path photons (both fluorescence and elastically scattered photons) i.e. the photons that have undergone significantly larger number of scattering events. It has previously been shown that scattering of the excitation light (at wavelength λ_EX) and the fluorescence (at wavelength λ_EM) depolarizes light in a similar way [17,21]. It would thus follow that the average number of scattering events suffered by the excitation photons (at wavelength λEX) and the polarization preserving fluorescence photons (at wavelength λ_EM), would be similar to that of the polarization preserving elastically scattered photons (at wavelength λ_EM), provided the values for the anisotropy parameter (g) and scattering coefficient (µ_s) at the excitation wavelength are similar to those at emission wavelength. Since depolarization of fluorescence in a scattering medium is strongly dependent upon the value of g and µ_s of the medium, this hypothesis would breakdown for very dissimilar values of g and µ_s at the excitation and the emission wavelengths [18,19,21]. However, for a medium containing larger sized scatterers (a≥λ, g≥0.7) like biological tissue or the kind of phantoms used in this study (g~0.92), this approximation holds good, because, for such media the wavelength dependence of g and µ_s is rather weak. In that case, the absorption and the scattering properties of the medium would affect the intensity and line shape of the polarized component of fluorescence and the polarized component of elastic scattering quite similarly. It is also important to note here that, since the value for the scattering albedo [a=µ_s/(µ_s+µ_a)] can be considered as the effective photon weight reduction factor at individual scattering events [8,13,14], the above approximation would be valid if the values for the albedo of the medium at the excitation and the emission wavelengths were reasonably close. This indeed is the case for the phantoms used in this study (even for the maximum value of µ_a used in the phantoms, a_EX=0.88 and a_EM=0.89) or for biological tissue in the wavelength range 350 nm–700 nm [15]. The fact that this method could not recover intrinsic fluorescence from a turbid medium for extremely high absorption where the value for the ratio of µ_a/${{\mu }_{\text{s}}}^{/}$ were > 1.8, arises because the above-mentioned hypothesis does not work with reasonable accuracy for such high values of absorption coefficient of the medium.

It should be emphasized here, that in this study, we have carried out relative measurements. The proposed method of normalizing polarized fluorescence spectra by concomitantly measured polarized elastic scattering spectra have been observed to recover the intrinsic line shape of fluorescence in a turbid medium with reasonable accuracy. The method has also been employed successfully to probe changes in concentration of fluorophores present in a turbid medium. This might come out to be useful in probing intrinsic variation of fluorescence intensity and line shapes in biological tissues. However, since the propagation path of polarized fluorescence and elastically scattered light in tissue might also be affected by several other parameters such as a distribution in size, shape and refractive index of scatterers present in tissue and the complex heterogeneity of tissue micro-structure, a direct conclusion regarding the applicability of this technique to tissue is not possible at this moment. Experiments on tissue samples are underway to explore this possibility.

4. Conclusions

To conclude, we have presented an experimental approach to recover intrinsic fluorescence from a turbid medium by removing the distorting effects of the absorption and scattering properties of the medium. The method is based on simultaneous measurements of polarized fluorescence and polarized elastic scattering spectra from a turbid medium. The polarized fluorescence normalized by the polarized elastic scattering spectra (in the wavelength range of fluorescence emission) was found to be free from the distorting effects of absorption and scattering properties of the medium. The results obtained from studies on different tissue phantoms show that this technique can recover line shape and intensity information of intrinsic fluorescence from fluorophores embedded in a scattering medium with reasonable accuracy for the range of optical transport parameters where the ratio of µ_a/${{\mu }_{\text{s}}}^{/}$ lies between 0 and 1.8.