Influence of the Near Molecular Vicinity on the Temperature Regulated Fluorescence Response of Poly(N-vinylcaprolactam)

A series of new fluorescent dye bearing monomers, including glycomonomers, based on maleamide and maleic esteramide was synthesized. The dye monomers were incorporated by radical copolymerization into thermo-responsive poly(N‑vinyl-caprolactam) that displays a lower critical solution temperature (LCST) in aqueous solution. The effects of the local molecular environment on the polymers’ luminescence, in particular on the fluorescence intensity and the extent of solvatochromism, were investigated below as well as above the phase transition. By attaching substituents of varying size and polarity in the close vicinity of the fluorophore, and by varying the spacer groups connecting the dyes to the polymer backbone, we explored the underlying structure–property relationships, in order to establish rules for successful sensor designs, e.g., for molecular thermometers. Most importantly, spacer groups of sufficient length separating the fluorophore from the polymer backbone proved to be crucial for obtaining pronounced temperature regulated fluorescence responses.


Introduction
Many nonionic polymers exhibit thermo-responsive behavior in aqueous solution, i.e., they undergo a coil-to-globule collapse phase transition exhibiting a lower critical solution temperature (LCST) [1][2][3][4][5][6]. This phase transition is associated with the transformation of the polymer chains from a well hydrated, and thus expanded state to a shrunken collapsed one. Still in many cases, macroscopic phase separation is not observed, but phase separation is limited to the mesoscopic level due to the formation of stable colloidal aggregates in the 100-1000 nm range, the so-called "mesoglobules" [7].
Occasionally, such thermo-responsive polymers have been functionalized by incorporating fluorescence dyes. Beyond acting as tracers [8][9][10][11], the incorporated dyes have been used, on the one hand, to follow the phase transition more easily, or to elucidate the transition mechanism [12][13][14][15][16][17][18][19][20][21][22][23][24]. On the other hand, the combination of thermo-responsiveness and color in the same polymers has been explored for various sensor applications, e.g., for molecular thermometers [25][26][27][28][29][30][31][32][33][34][35][36][37][38][39][40][41][42][43][44] or for the detection of ions or biologically relevant molecules [45][46][47][48][49][50]. When passing through the phase transition, the environment of the polymer anchored dyes is changed so that the spectroscopic behavior may be altered in manifold ways. For instance, the extinction coefficient or the quantum yield may For labeling PNVCL by fluorophores, copolymerization of NVCL with dye-functionalized comonomers is advantageous [13,15], as both the monomer NVCL and the homopolymer are difficult to functionalize. However, NVCL is much less reactive in radical copolymerization than styrenic and (meth)-acrylic derivatives. We therefore designed a series of new solvatochromic naphthalimidebearing comonomers derived from maleic acid (Figure 1), in order to ensure the incorporation of isolated fluorophores into the polymers and thus, to exclude possible local clustering effects on the spectral properties. Although maleic acid derivatives do not homopolymerize, they are known to copolymerize reasonably well with N-vinylamides, with reactivity ratios close to 0 [61][62][63][64]. Moreover, as only one of the two carboxyl groups is needed to anchor the fluorophore to the polymerizable moiety, the second carboxyl group can be used for incorporating additional functional groups. We took advantage of this option by introducing substituents of diverse polarity, i.e., ranging from highly hydrophilic to strongly hydrophobic, in the near vicinity of the fluorophore within an identical thermo-sensitive polymer matrix. In this way, we aimed at exploring the influence of the local molecular environment on the spectral properties of the dye label, below as well as above the coil-toglobule phase transition.

2-(acetoxymethyl
Monomer (1e) is synthesized in two steps. In the first step, the synthetic procedure of monomers 1a-d is followed using propargylamine as functional amine to be coupled, yield 90%.
Mass spectrum ( In the second step, the propargylamide intermediate (1 eq.), DIEA (3 eq.) and tetraacetyl-1ethoxy-2-azido-mannose (1 eq.) in 1 mL of dry CH 2 Cl 2 were purged for 10 min with argon. After adding bromo-tris(triphenylphosphine)copper(I) (0.1 eq.), the mixture was stirred for 24 h at 40˝C. The solvent was evaporated, the residue dissolved in little methanol, and the mixture precipitated by adding a large excess of diethylether. The product is filtered off and dried in vacuo. Yield 72%, yellow solid.

Synthesis of the Polymers
N-vinylcaprolactam (NVCL, 1 eq.), the selected dye-functionalized monomer (0.1 mol %) and azoinitiator (1 mol %) in dry methanol (25 w % solid content) were placed into a Schlenk flask equipped with a rubber septum. The mixture was purged with argon for 15 min and heated to 60˝C for 24 h in an oil bath. Then, the solution was cooled to ambient temperature, and dialyzed for 2 days against 50 vol % aqueous ethanol, and for 5 more days against water. The polymers containing dye comonomers 1e and 1f with the protected mannose or, respectively, glucose derived substituents were additionally dialyzed in semi-dilute aqueous K 2 CO 3 (12.5 g¨L´1) to split off the acetyl groups. The polymers were lyophilized to yield pale yellow powders. Table 1 lists the composition of the polymerization feeds and the yields of the dye-functionalized PNVCL samples prepared.

Methods
NMR-spectra were taken with a Bruker Avance 300 NMR spectrometer (Bruker, Billerica, MA, USA). 1 H NMR spectra were recorded at 300 MHz, and 13 C NMR spectra at 75 MHz (APT modus). The chemical shift δ was calibrated to the solvent signal. High resolution mass spectra (HRMS) were taken with a GC-MS Trace DSQ II (EI method), or an ESI-Q-TOFmicro mass spectrometer (ESI method) (both mass spectrometers: Thermo Scientific, Waltham, MA, USA). For samples taken with the EI method the relative mass m/z is given. For samples recorded with the ESI method, the mass is given plus the detected Ion (H + ). IR spectra (ATR modus) were recorded with a NEXUS-FT-IR spectrometer with an ATR Smart Endurance Element (Thermo Instrument Systems Inc., Waltham, MA, USA). Size exclusion chromatography (SEC) was performed in dimethylformamide (DMF) as eluent. The set-up consists of a single-channel degasser (WGE Dr. Bures, Dallgow-Doeberitz, Germany), an isocratic pump P 1000 (Spectra Physics, Santa Clara, CA, USA), a set of columns (Guard 7,5¨75 mm, PolarGel L 7,5¨300 mm; Polymer Laboratories-Agilent Technologies, Santa Clara, California, USA), a UV-VIS detector SEC 3010 and a refractometer SEC 3010 (WGE Dr. Bures, Dallgow-Döberitz, Germany). Measurements were taken at 50˝C with a flow rate of 1 mL¨min´1, and calibrated with linear polystyrene standards (Polymer Standards Service; PSS, Mainz, Germany). UV-VIS signals were recorded at 398 nm for naphthalimide-bearing polymers. Absorption spectra in solution were recorded with a UV/VIS/NIR two beam spectrometer (Lamda 19, Perkin Elmer, Waltham, MA, USA, 1 nm slit, scan speed of 120 nm¨min´1) in PMMA cuvettes (path length 10 mm). For turbidity studies, the transmittance of polymer solutions (3.00 g¨L´1) was monitored by a photometer (Varian, Cary 50 UV-VIS, Palo Alto, CA, USA) at 600 nm, as function of temperature with heating and cooling rates of 0.2 K¨min´1. Steady state fluorescence of polymer solutions was recorded a LS 50B Luminescence Spectrometer (Perkin Elmer, Waltham, MA, USA) in PMMA cuvettes (10 mm path length). Temperature dependent steady state fluorescence was measured with a Spectrofluorometer Jasco FP-8500 (Jasco International, Tokyo, Japan). The cuvette holder was equipped with a thermoelectric Peltier element for temperature control. Solutions were permanently stirred during the measurements with an external magnetic stirrer. Photoluminescence quantum yields were determined with a set-up Hamamatsu C9920-02 (Hamamatsu Photonics, Hamamatsu, Japan) that includes an integrating sphere with a photon multi-channel analyzer. The absorbance of the solutions was always kept below 0.1.

Synthesis of the Fluorophore-Labeled Polymers
The synthesis of the monomers followed classical strategies, using the dye-functionalized maleic acid monoester or monoamide, respectively, as intermediates. This enabled the facile and versatile variation of the second carboxyl substituent of the polymerizable maleic acid motif in the final step. The various new dye-functionalized maleamides or maleic esteramides were engaged as minority component in the radical copolymerization with NVCL in methanol. The relative amount of the dye comonomers was kept as low as 0.1 mol % in the reaction mixtures, to obtain polymers that incorporate the fluorophores in the midst of the chains, but do not contain more than one fluorophore per chain. Synthetic procedures and the analytical data of the copolymers obtained are summarized in Table 1. As the amounts of dye comonomer incorporated were too low to be analyzed by 1 H NMR spectroscopy (see Figure 2), the amounts were derived from the visual absorbance band of the chromophore, assuming that the extinction coefficients of the monomers do not change upon incorporation into the polymers. Within the precision of the analysis, the copolymers contain the same amount of dye monomers as engaged in the reaction mixture, in agreement with the high copolymer yields. The 1 H NMR spectra of the various copolymers correspond to the spectrum of PNVCL homopolymer, as exemplified for PNVCL-g in Figure 2. The apparent molar masses of the various copolymers were all similar, being within the range of 70,000 to 120,000. path length). Temperature dependent steady state fluorescence was measured with a Spectrofluorometer Jasco FP-8500 (Jasco International, Tokyo, Japan). The cuvette holder was equipped with a thermoelectric Peltier element for temperature control. Solutions were permanently stirred during the measurements with an external magnetic stirrer. Photoluminescence quantum yields were determined with a set-up Hamamatsu C9920-02 (Hamamatsu Photonics, Hamamatsu, Japan) that includes an integrating sphere with a photon multi-channel analyzer. The absorbance of the solutions was always kept below 0.1.

Synthesis of the Fluorophore-Labeled Polymers
The synthesis of the monomers followed classical strategies, using the dye-functionalized maleic acid monoester or monoamide, respectively, as intermediates. This enabled the facile and versatile variation of the second carboxyl substituent of the polymerizable maleic acid motif in the final step. The various new dye-functionalized maleamides or maleic esteramides were engaged as minority component in the radical copolymerization with NVCL in methanol. The relative amount of the dye comonomers was kept as low as 0.1 mol % in the reaction mixtures, to obtain polymers that incorporate the fluorophores in the midst of the chains, but do not contain more than one fluorophore per chain. Synthetic procedures and the analytical data of the copolymers obtained are summarized in Table 1. As the amounts of dye comonomer incorporated were too low to be analyzed by 1 H NMR spectroscopy (see Figure 2), the amounts were derived from the visual absorbance band of the chromophore, assuming that the extinction coefficients of the monomers do not change upon incorporation into the polymers. Within the precision of the analysis, the copolymers contain the same amount of dye monomers as engaged in the reaction mixture, in agreement with the high copolymer yields. The 1 H NMR spectra of the various copolymers correspond to the spectrum of PNVCL homopolymer, as exemplified for PNVCL-g in Figure 2. The apparent molar masses of the various copolymers were all similar, being within the range of 70,000 to 120,000.

Spectroscopic Properties of the Monomers and Polymers In Ethanol
As dye monomers, 1a-1g were virtually insoluble in water, spectroscopic properties of the markedly solvatochromic monomers and polymers were first characterized in ethanol, as illustrated

Spectroscopic Properties of the Monomers and Polymers In Ethanol
As dye monomers, 1a-1g were virtually insoluble in water, spectroscopic properties of the markedly solvatochromic monomers and polymers were first characterized in ethanol, as illustrated in Figure 3. Table 2 summarizes the key values for the monomers derived therefrom. Typically, they show absorbance maxima in the range of 395-398 nm, and emission maxima in the range of 530-534 nm (Figure 3a,b). Accordingly, the fluorescence spectra exhibit pronounced Stoke shifts. The quantum yields are low in the polar protic solvent ethanol. The latter findings correspond well to a recent report on the fluorescence of a methacrylate dye monomer bearing the closely related fluorophore 4-(N,N-dimethylamino)-1,8-naphthalimide [21]. in Figure 3. Table 2 summarizes the key values for the monomers derived therefrom. Typically, they show absorbance maxima in the range of 395-398 nm, and emission maxima in the range of 530-534 nm (Figure 3a,b). Accordingly, the fluorescence spectra exhibit pronounced Stoke shifts. The quantum yields are low in the polar protic solvent ethanol. The latter findings correspond well to a recent report on the fluorescence of a methacrylate dye monomer bearing the closely related fluorophore 4-(N,N-dimethylamino)-1,8-naphthalimide [21].    Still, despite the chemically very similar structure of the 4-(morpholino)-1,8-naphthalimide fluorophore used here, all monomers 1a-1g exhibit a blue shift of about 20-25 nm for the absorbance band in comparison to the previously studied system, suggesting a notably weaker electron donor effect of the morpholino compared to the dimethylamino substituent, while the emission band is hypsochromically shifted by only 5 nm.
The homopolymer PNVCL-0 as well as all copolymers of NVCL with the various dye monomers, i.e, PNVCL-a-PNVCL-g, were soluble in both ethanol and water at ambient temperature. For comparison with the monomers, their spectroscopic properties were first characterized in ethanol (Figure 3c,d). Similar to the monomers, all absorbance maxima λ max Abs are within the range of 395-398 nm, while emission maxima λ max Em are all 530˘1 nm, as summarized in Table 3. Compared to the behavior of the monomers (see Table 2), the absorbance maxima of the chromophores seem hardly changed by incorporation into PNVCL, while for the emission maxima, a small but systematic hypsochromic shift of ca. 2 nm seems to occur. The spectroscopic properties of the polymers were additionally investigated in water (Figure 3e,f), in which the monomers are insoluble. All polymers display very similar absorbance spectra, in which the maxima are shifted bathochromically by about 10 nm in comparison to ethanol solutions. In addition, concerning the fluorescence spectra, we note a general bathochromic shift of the emission maxima in water compared to ethanol, which is in the order of 10 nm. The general red shift encountered of both the absorbance and emission spectra in water compared to the less polar solvent ethanol, is in agreement with the established solvatochromic behavior of the 4-amino-1,8-naphthalimide chromophore [21,68]. The emission maxima of the various dye-labeled polymers in water are all located closely around 540 nm, with the exception of PNVCL-a for which λ max Em is blue-shifted to about 534 nm. This copolymer bears the most hydrophobic substituent, namely an octadecyl chain, in the direct neighborhood of the fluorophore, thus apparently reducing the local polarity somewhat. Noteworthy, a close analysis of the shape of the emission bands of the polymers in aqueous solution reveals small shoulders, and in the case of PNVCL-e even a small secondary maximum on the low wavelength slope of the bands. This points to the formation of H-aggregates, presumably due to the hydrophobic character of the fluorophore, despite of the low dye content. As these shoulders are very small, the role of the presumably underlying aggregates was supposed to be negligible for the studies of the thermo-responsive behavior (see below). Still, the finding underlines that the potential risk of hydrophobic aggregation of environmentally sensitive dyes has always to be kept in mind when designing and analyzing polymeric dye based molecular thermometers.

Thermoresponsive Behavior of the Dye-Labeled Polymers
The cloud point, i.e., the phase transition temperature, of the reference homopolymer PNVCL-0 in aqueous solution was found by turbidimetry as 33.7˝C at a concentration of 3 g¨L´1 (Figure 4). The hysteresis observed for heating and cooling curves were about 1˝C or less, and thus, virtually negligible. The cloud point value agrees well with literature data on comparably concentrated aqueous solutions of PNVCL samples having similar molar mass [54,56,58,59].
shoulders are very small, the role of the presumably underlying aggregates was supposed to be negligible for the studies of the thermo-responsive behavior (see below). Still, the finding underlines that the potential risk of hydrophobic aggregation of environmentally sensitive dyes has always to be kept in mind when designing and analyzing polymeric dye based molecular thermometers.

Thermoresponsive Behavior of the Dye-Labeled Polymers
The cloud point, i.e., the phase transition temperature, of the reference homopolymer PNVCL-0 in aqueous solution was found by turbidimetry as 33.7 °C at a concentration of 3 g·L −1 (Figure 4). The hysteresis observed for heating and cooling curves were about 1 °C or less, and thus, virtually negligible. The cloud point value agrees well with literature data on comparably concentrated aqueous solutions of PNVCL samples having similar molar mass [54,56,58,59]  Similarly for polymers PNVCL-a-PNVCL-g, the cloud points were all in the range of 31-33.5 °C ( Figure 4). As expected, the incorporation of small amounts of a functional comonomer did not interfere much with the thermo-responsiveness of the parent homopolymer PNVCL. Still, we find that all copolymers, which contain in average less than 1 comonomer unit per polymer chain, exhibit a small, but nevertheless notable decrease of the cloud point of 0.5 to 2.5 °C compared to PNVCL-0, the homopolymer reference. One might be tempted to correlate the reduced cloud points directly to the hydrophobicity of the dye monomer units incorporated, in particular to the relative differences in the distinct hydrophobic character of the substituents attached to the second carboxyl group of the maleic acid moiety. Indeed, PNVCL-a incorporating the most hydrophobic comonomer 1a, shows the lowest cloud point within the series. However, the situation is more complex, as the most hydrophilic substituents, as found in monomers 1d and 1f, do not provoke the highest cloud points of the copolymer series. In fact, it seems that the small differences seen additionally reflect the differences in molar mass of the samples (see Table 1). Accordingly, the slightly lower cloud points of the dye-labeled copolymers result from combined effects of the somewhat increased molar masses and the somewhat increased overall hydrophobicity. Anyhow, the phase transitions of the various polymers prepared occur within a very narrow temperature window, and thus enable direct comparative studies of the temperature induced spectroscopic changes.
The corresponding evolution of the fluorescence spectra of the copolymer series with increasing temperature is illustrated in Figures 5-11. We find for all dye-labeled polymers marked changes of their fluorescence behavior with the temperature. Qualitatively, we observe a marked increase of the fluorescence intensity when passing through the phase transition. In addition, we note a marked Similarly for polymers PNVCL-a-PNVCL-g, the cloud points were all in the range of 31-33.5˝C ( Figure 4). As expected, the incorporation of small amounts of a functional comonomer did not interfere much with the thermo-responsiveness of the parent homopolymer PNVCL. Still, we find that all copolymers, which contain in average less than 1 comonomer unit per polymer chain, exhibit a small, but nevertheless notable decrease of the cloud point of 0.5 to 2.5˝C compared to PNVCL-0, the homopolymer reference. One might be tempted to correlate the reduced cloud points directly to the hydrophobicity of the dye monomer units incorporated, in particular to the relative differences in the distinct hydrophobic character of the substituents attached to the second carboxyl group of the maleic acid moiety. Indeed, PNVCL-a incorporating the most hydrophobic comonomer 1a, shows the lowest cloud point within the series. However, the situation is more complex, as the most hydrophilic substituents, as found in monomers 1d and 1f, do not provoke the highest cloud points of the copolymer series. In fact, it seems that the small differences seen additionally reflect the differences in molar mass of the samples (see Table 1). Accordingly, the slightly lower cloud points of the dye-labeled copolymers result from combined effects of the somewhat increased molar masses and the somewhat increased overall hydrophobicity. Anyhow, the phase transitions of the various polymers prepared occur within a very narrow temperature window, and thus enable direct comparative studies of the temperature induced spectroscopic changes.
The corresponding evolution of the fluorescence spectra of the copolymer series with increasing temperature is illustrated in Figures 5-11. We find for all dye-labeled polymers marked changes of their fluorescence behavior with the temperature. Qualitatively, we observe a marked increase of the fluorescence intensity when passing through the phase transition. In addition, we note a marked hypsochromic shift of the emission maximum, indicating a less polar environment of the chromophores above the phase transition. Accordingly, all dye-labeled polymers may act a priori as molecular thermometers. hypsochromic shift of the emission maximum, indicating a less polar environment of the chromophores above the phase transition. Accordingly, all dye-labeled polymers may act a priori as molecular thermometers.    hypsochromic shift of the emission maximum, indicating a less polar environment of the chromophores above the phase transition. Accordingly, all dye-labeled polymers may act a priori as molecular thermometers.    hypsochromic shift of the emission maximum, indicating a less polar environment of the chromophores above the phase transition. Accordingly, all dye-labeled polymers may act a priori as molecular thermometers.              Nevertheless, the quantitative fluorescence responses to temperature changes differ substantially between the individual polymers. Clearly, the thermally-induced spectral shift as well as the induced fluorescence enhancement are by far the smallest for PNVCL-g ( Figure 11). More pronounced, but still relatively weak effects are observed in the case of PNVCL-a ( Figure 5). The strongest fluorescence enhancement is seen for samples PNVCL-d and PNVCL-e, while the highest blue shifts are encountered for samples PNVCL-d and PNVCL-f (Figures 8-10).
The key data extracted from the various plots are summarized in Table 4. When correlating the data with the chemical structure of the dye comonomers incorporated, we see that the by far weakest effects of the phase transition are seen when incorporating comonomer 1g. In this comonomer, the chromophore is attached to the polymer backbone by a short spacer only. As the caprolactam side chains are amphiphilic, and in particular, present a rather large hydrophobic fragment in the vicinity of the naphthalimide fluorophore, one may attribute this finding to an inherent partial hydrophobic shielding of the fluorophore. Such a shielding is effective also at temperatures well below the phase transition. A similar explanation was proposed for explaining the weak fluorescence enhancement of incorporated naphthalimide dyes in poly(oligoethyleneglycol acrylate)s above the phase transition of the LCST-type. Alternatively, the short spacer group possibly precludes the efficient transfer of the fluorophore from an aqueous environment to the hydrophobic pockets formed by the collapsed polymer coils above the phase transition, thus exposing the label to a partially hydrated environment even at high temperatures. The latter explanation is supported by the spectral data (see Table 3). While the value of λmax Em of 541 nm for PNVCL-g below the cloud point is indicative of a wellhydrated environment, λmax Em reaches only a value of 532 nm above the phase transition temperature, which indicates a still rather polar environment.
The situation becomes much more favorable, when a longer spacer is incorporated between label and polymer backbone (as in copolymers PNVCL-a-PNVCL-f), which enables the fluorophore "to stick out" beyond the side chains of the polymer matrix, and simultaneously renders it more mobile. However, the still moderate effects found for PNVCL-a exemplify, that overly big hydrophobic fragments attached to the polymer reduce the sensitivity of the label to the conformational changes nevertheless, and thus, obscure notably the subsequent changes of the local environment. This explanation would fit with the relatively low value for λmax Em of 536 nm at 10 °C, pointing to a partially dehydrated environment already at low temperatures.
In reverse, hydrophilic moieties attached in the vicinity of the dye label improve the fluorescence response, as seen by the very strong spectral as well as intensity changes encountered for polymers PNVCL-d, PNVCL-e, and PNVCL-f. They bear the most hydrophilic side groups fixed to the maleamide comonomers. Note, that for the copolymers with the long spacer separating the label from the backbone, even moderately hydrophobic substituents, as introduced by comonomers 1b and 1c, still allow for strong thermally-induced fluorescence effects.  Nevertheless, the quantitative fluorescence responses to temperature changes differ substantially between the individual polymers. Clearly, the thermally-induced spectral shift as well as the induced fluorescence enhancement are by far the smallest for PNVCL-g ( Figure 11). More pronounced, but still relatively weak effects are observed in the case of PNVCL-a ( Figure 5). The strongest fluorescence enhancement is seen for samples PNVCL-d and PNVCL-e, while the highest blue shifts are encountered for samples PNVCL-d and PNVCL-f (Figures 8-10).
The key data extracted from the various plots are summarized in Table 4. When correlating the data with the chemical structure of the dye comonomers incorporated, we see that the by far weakest effects of the phase transition are seen when incorporating comonomer 1g. In this comonomer, the chromophore is attached to the polymer backbone by a short spacer only. As the caprolactam side chains are amphiphilic, and in particular, present a rather large hydrophobic fragment in the vicinity of the naphthalimide fluorophore, one may attribute this finding to an inherent partial hydrophobic shielding of the fluorophore. Such a shielding is effective also at temperatures well below the phase transition. A similar explanation was proposed for explaining the weak fluorescence enhancement of incorporated naphthalimide dyes in poly(oligoethyleneglycol acrylate)s above the phase transition of the LCST-type. Alternatively, the short spacer group possibly precludes the efficient transfer of the fluorophore from an aqueous environment to the hydrophobic pockets formed by the collapsed polymer coils above the phase transition, thus exposing the label to a partially hydrated environment even at high temperatures. The latter explanation is supported by the spectral data (see Table 3). While the value of λ max Em of 541 nm for PNVCL-g below the cloud point is indicative of a well-hydrated environment, λ max Em reaches only a value of 532 nm above the phase transition temperature, which indicates a still rather polar environment. The situation becomes much more favorable, when a longer spacer is incorporated between label and polymer backbone (as in copolymers PNVCL-a-PNVCL-f), which enables the fluorophore "to stick out" beyond the side chains of the polymer matrix, and simultaneously renders it more mobile. However, the still moderate effects found for PNVCL-a exemplify, that overly big hydrophobic fragments attached to the polymer reduce the sensitivity of the label to the conformational changes nevertheless, and thus, obscure notably the subsequent changes of the local environment. This explanation would fit with the relatively low value for λ max Em of 536 nm at 10˝C, pointing to a partially dehydrated environment already at low temperatures.
In reverse, hydrophilic moieties attached in the vicinity of the dye label improve the fluorescence response, as seen by the very strong spectral as well as intensity changes encountered for polymers PNVCL-d, PNVCL-e, and PNVCL-f. They bear the most hydrophilic side groups fixed to the maleamide comonomers. Note, that for the copolymers with the long spacer separating the label from the backbone, even moderately hydrophobic substituents, as introduced by comonomers 1b and 1c, still allow for strong thermally-induced fluorescence effects. Interestingly, the closer analysis of the data in Table 4 also reveals, that the extents of fluorescence enhancement and of induced spectral shifts do not strictly coincide. Copolymer PNVCL-e, which shows the highest increase of fluorescence when crossing the phase transition, provides a spectral shift of 12 nm only, which is clearly below the shifts of 16-18 nm achieved by PNVCL-b-PNVCL-d and PNVCL-f. At present, we can only speculate about this at the first sight somewhat puzzling observation. Possibly, the particularly long spacer group in comonomer 1e, due to the build up by the azide-alkyne "click reaction", places the fluorophore and the hydrophilic fragment very far apart. This enables the fluorophore to avoid the neighborhood of the hydrophilic group effectively, and to stay close to the more hydrophobic lactam side chains of the polymer matrix. This explanation would be consistent with the relatively low value observed for λ max Em of 538 nm at 10˝C, pointing to a partially dehydrated environment already at low temperatures, as discussed above for the case of PNVCL-a.
In contrast, the local polarity sensed above the phase transition is comparable to the ones found for the other copolymers (see Table 4). Concerning temperature-regulated fluorescence effects, these findings suggest that one has to distinguish between effects due to a coil-to-globule collapse induced change of the average dielectric constant of the fluorophore's direct vicinity (governing the positions of the absorbance and emission bands, and thus, the extent of the solvatochromic shifts), and of the number of solvent molecules in direct contact to the fluorophore (that act as quenchers and govern the quantum yield). Accordingly, an appropriate design for efficient polymeric molecular thermometers asks for the separation of the fluorophore from the backbone by a spacer group of a specific length, in addition to the appropriate selection of the environmentally sensitive fluorophore. Moreover, it seems that hydrophilic groups fixed in the vicinity of the fluorophores are advantageous, as they increase the difference felt between the well-hydrated state below the phase transition temperature, and the-more or less-dehydrated state above, thus increasing the thermometer's sensitivity.
Another instructive feature evident from Figures 5-11 are the differing temperature profiles of the fluorescence intensity changes and of the induced solvatochromic shifts. The latter extend systematically over a considerably broader temperature range. This implies that for monitoring precisely very small temperature changes, fluorescence intensity will be the more sensitive parameter to follow, whereas for monitoring a larger temperature window, the solvatochromic shift will be the more suited parameter. In any case, the simultaneous monitoring of both parameters enables to crosscheck the temperature value deduced, thus enhancing the reliability of the method.

Conclusions
Naphthalimide functionalized maleic acid diamides or esteramides bearing various hydrophilic or hydrophobic substituents are useful functional comonomers for the copolymerization with vinylamides such as N-vinylcaprolactam. The resulting dye-labeled copolymers are thermo-sensitive in aqueous solution, and behave as molecular thermometers. On the one hand, the fluorescence intensity of the polymers undergoes a dramatic increase when passing from below to above the phase transition temperature. On the other hand, the emission band maximum of the solvatochromic fluorophores incurs a marked hypsochromic shift. Both parameters enable independently to follow the temperature of the system in a window of about 10 or 40 K, respectively, around the phase transition temperature. Interestingly, the temperature regulated responses of the two effects, i.e., the extent of the modulation of the fluorescence intensity and of the solvatochromic shift encountered, do not coincide exactly. In addition, they cover temperature windows of differing widths. Still, they may be exploited jointly, to improve the reliability of such spectroscopic temperature measurements.
The broad variation of the substituents undertaken on the maleic comonomers provides some guidelines for the molecular design of effective polymeric molecular thermometers. First of all, the studies demonstrate the importance of a spacer group of a specific length that separates the dye label from the polymer backbone. An appropriate spacer seems essential to obtain sensitive temperature regulated fluorescence responses. Accordingly, the successful realization of effective polymeric molecular thermometers based on PNIPAM in the past seems serendipitous, as the side chains of PNIPAM are small, and therefore, the spacer group can be very short. In addition to the need for a spacer group, it seems that hydrophilic groups fixed in the vicinity of the fluorophore label help to increase the difference felt between the well hydrated state below, and, respectively, the less hydrated state above the phase transition temperature, thus increasing the thermometer's sensitivity.