Introduction

Data transmission along optical fibers has progressed with increasing speed, data quality and decreasing cost. Planar waveguide devices offer the advantages of a cost-effective fabrication process, easy integration and compactness.1, 2 These waveguide devices potentially offer high reliability because unlike micro-opto-electromechanical systems, there are no moving parts, and the devices are driven by a change in the refractive index.3 Polymer waveguide materials have attracted increasing interest in the design of thermally operating devices, such as optical switches, optical splitters, optical attenuators and optical filters.4, 5 These devices should be designed and fabricated to achieve better performance in terms of low power consumption, low switching voltage and rapid response. The device performance is based on the thermo-optic effect of the composing material. A large thermo-optic coefficient (TOC) of a material means that a small temperature change is needed for the necessary change in refractive index, which reduces the power consumption for device operation.6 Many thermo-optic waveguide devices have been fabricated using polymer materials because of their large TOCs, which are much greater than those of inorganic materials, such as silica (1.1 × 10−5 K−1) or lithium niobate (4 × 10−5 K−1).6, 7, 8 Conventional poly(methyl methacrylate) has a TOC of −1.2 × 10−4 K−1. The polymer TOC provides a negative sign, which in contrast to inorganic materials, means a decrease in refractive index with increasing temperature. The TOC is largely dependent on the polymer structure. Polyimides have been reported to have a smaller TOC than poly(methyl methacrylate), and fluorinated acrylates increased the TOC up to −2.8 × 10−4 K−1.6 The thermal expansion effect of rubbery polyurethanes has attracted considerable attention despite their low thermal stability. The thermal deformation of urethane polymers has limited their use in waveguide device applications. Recently, Qiu et al.9, 10 reported on a thermal waveguide switch using azo-polyurethanes. In this study, thermally-stable linear polyurethanes were developed and assessed as thermo-optic waveguide material. UV-polymerization was adapted for stable film formation using acrylate monomers involving urethane-groups for a high thermo-optic effect.

Experimental Procedure

Materials and instruments

All reagents were purchased from Sigma-Aldrich Chemical Co. (Yongin, Korea), and the reagent-grade solvents were dried when necessary and purified by vacuum distillation. Column chromatography was performed using silica gel (Merck Korea, Seoul, Korea, 250–430 mesh). 1H-NMR spectroscopy experiments (Bruker AM-300 spectrometer, Gottingen, Germany) were performed using tetramethylsilane (TMS; δ=0 p.p.m.) as an internal standard. A MAGNA-IR 750 spectrometer (Nicolet Instrument Co., Madison, WI, USA) was used to record the FT-IR spectra. The mass spectra were recorded on an Agilent 1200LC/1100 MSD SL mass spectrometer (Agilent, Marlborough, MA, USA). The thermal analysis was performed on a Mettler Toledo DSC 822e (Mettler Toledo Inc., Columbus, OH, USA) and on a TAG/SDTA 851e (Mettler Toledo Inc.) using a heating rate of 10 °C min−1 under a nitrogen atmosphere. The molecular weights of the polymers were determined using Agilent PL-20 (Agilent, Santa Clara, CA, USA) with a polystyrene standard. The optical properties of the polymer films such as refractive index, optical loss and dn/dT were determined using a Sairon prism coupler SPA-4000 (Kwangju, Korea).

Synthesis of 1, 2, 3a

The molecules 1, 2, 3a were prepared using the method described in the literature11

Compound 1: 1H NMR (CDCl3, 300 MHz): δ 1.41–1.63 (m, 2H), 2.78 (s, 2H), 3.02 (s, 2H). 4.15–4.28 (m, 4H), 6.30 (s, 2H), 7.40–7.56 (m, 6H), 8.01–8.08 (m, 4H); 13C NMR (CDCl3, 75 MHz): δ 41.0, 45.9, 49.4, 65.4, 128.6, 129.8, 131.9, 133.2, 135.8, 166.7.

Compound 2: 1H NMR (CDCl3, 300 MHz): δ 2.19–2.40 (m, 2H), 2.82–3.02 (m, 4H), 4.21–4.38 (m, 2H), 4.42–4.59 (m, 2H), 7.28–7.40 (m, 5H), 7.44–7.58 (m, 2H), 7.78–7.99 (m, 3H), 9.71 (s, 2H); 13C NMR (CDCl3, 75 MHz): δ 44.0, 52.4, 61.8, 77.0, 128.7, 129.3, 129.8, 133.5, 166.1, 201.7; IR (KBr): νmax 3069, 2968, 2830, 2728, 1714, 1272 cm−1.

Compound 3a: 1H NMR (CDCl3, 300 MHz): δ 1.38–1.51 (m, 2H), 2.00–2.22 (m, 4H), 3.00–3.25 (m, 8H), 3.61–3.80 (m, 4H), 3.99 (s, 2H, OH), 4.65 (d, 2H); 13C NMR (CDCl3, 75 MHz): δ 34.9, 35.3, 43.4, 46.1, 59.4, 61.7; LC-MS, m/z: 339.07 (M+1)+.

Synthesis of 3b, 3c

2-Mercaptoethanol (0.79 ml, 11.2 mmol), p-toluene sulfonic acid (0.10 g, 0.50 mmol) and freshly prepared dialdehyde 2 (2.0 g, 5.1 mmol) were dissolved in benzene and refluxed for 8 h using a Dean-Stark apparatus (Yongin, Korea). The cooled reaction mixture was condensed and purified by column chromatographic separation to yield an intermediate product (2.1 g, 80%). 1H NMR (CDCl3, 300 MHz): δ 1.61 (t, 2H), 2.53 (m, 2H), 2.60 (m, 2H), 3.01 (m, 4H), 3.75 (m, 2H), 4.38 (m, 2H), 4.44 (d, 4H), 5.24 (d, 2H), 7.41 (m, 4H), 7.53 (m, 4H), 8.01 (d, 2H).

The intermediate (2.1 g, 4.1 mmol) was dissolved in tetrahydrofuran (THF) (6 ml) and mixed with KOH (0.7 g) in methanol (6 ml). After being stirred for 1 h, the reaction mixture was treated with saturated aqeous NH4Cl solution (10 ml) and extracted with ethyl acetate. The organic solution was washed with water (20 ml) and dried over anhydrous MgSO4. The volatile organics were removed using rotary evaporation. A column chromatography on silica gel using hexane and ethyl acetate (2/1, v/v) yielded 3b (1.1 g, 85%). 1H NMR (CDCl3, 300 MHz): δ 1.20–1.50 (m, 2H), 1.99–2.30 (m, 4H), 2.59 (s, 2H, OH), 2.82–3.15 (m, 4H), 3.41–3.61 (m, 2H), 3.62–3.90 (m, 4H), 3.82–3.99 (m, 2H), 4.28–4.41 (m, 2H), 5.01 (d, 1H), 5.15 (d, 1H); 13C NMR (CDCl3, 75 MHz): δ 33.0 (32.9), 36.6, 47.6 (47.5), 49.5 (49.4), 51.0, 51.9 (51.8), 71.7 (71.6), 74.2 (73.9), 90.2 (90.0); LC-MS, m/z: 307.10 (M+1)+.

3c was prepared using a method similar to that used in the preparation of 3b through the reaction of dialdehyde and ethylene glycol in a 90% yield. 3c: 1H NMR (CDCl3, 300 MHz): δ 1.35 (m, 2H), 1.81–2.05 (m, 2H), 2.44 (s, 2H, OH), 3.39–3.46 (m, 2H), 3.60–4.01 (m, 12H), 4.73 (d, 1.8H, OCH2O), 5.02 (d, 0.2H, OCH2O); 13C NMR (CDCl3, 75 MHz): δ 32.2, 46.3, 49.8, 65.3 (65.2), 74.0, 106.6 (106.4); LC-MS, m/z: 275.17 (M+1)+.

Synthesis of linear polyurethanes 4a, 4b, 4c

Diol 3a (0.65 g, 2.0 mmol), IPDI (0.20 g, 2.0 mmol) and dibutyltin dilaurate (0.12 ml, 0.20 mmol) were placed in a 25-ml two-neck flask with a reflux condenser and dissolved with anhydrous THF (3.8 ml) under N2 atmosphere. The reaction mixture was heated for 18 h at 50 °C and cooled. Methanol (30 ml) was added to the mixture, and the mixture was stored in a refrigerator overnight for polymer precipitation. The polymer was filtered and dried in a vacuum oven (50 °C) to yield polyurethane 4a (0.60 g, 70%). 1H NMR (CDCl3, 300 MHz): δ 0.65–1.00 (m, 6H), 1.02 (s, 3H), 1.25 (s, 2H), 1.42–1.80 (br, 6H), 2.15–2.40 (s, 4H), 2.81–3.05 (s, 2H, CH2-N), 3.06–3.38 (s, 8H, CH2-S), 3.61–3.88 (br, 1H, CH-N), 4.00–4.27 (br, 4H, CH2-O), 4.78 (s, 2H, S-CH2-S), 4.65–5.21 (br, 2H, NH). GPC (THF): Mn=19.4 kDa, Mw=34.9 kDa.

The polyurethanes 4b and 4c were synthesized with 75 and 85% yields, respectively, through a similar reaction as 4a.

Polyurethane 4b: 1H NMR (CDCl3, 300 MHz): δ 0.68–0.96 (m, 6H), 1.02 (s, 3H), 1.21 (s, 2H), 1.35–1.58 (br, 2H), 1.58–1.78 (m, 4H), 2.16–2.45 (br, 4H), 2.80–3.05 (m, 4H, CH2-S), 3.60–3.82 (m, 3H, CH2-N, CH-N), 3.99–4.22 (br, 4H, CH2-OCO), 4.22–4.41 (s, 4H, CH2-O), 4.60–5.21 (br, 2H, NH), 5.06 (s, 1H, S-CH2-O), 5.20 (s, 1H, S-CH2-O). GPC (THF): Mn=19.5 kDa, Mw=31.2 kDa.

Polyurethane 4c: 1H NMR (CDCl3, 300 MHz): δ 0.93 (s, 6H), 1.02 (s, 3H), 1.20 (s, 2H), 1.61–1.78 (br, 2H), 1.78–1.99 (m, 4H), 2.02–2.44 (m, 4H), 2.92 (s, 4H, CH2-N), 3.41–4.35 (m, 12H, CH2-O, CH2-OCO), 4.72 (s, 1H, O-CH2-O), 4.78 (s, 1H, O-CH2-O), 4.85–5.60 (br, 2H, NH). GPC (THF): Mn=16.4 kDa, Mw=29.5 kDa.

Synthesis of UV-cured poly(methacrylate) 4d

Diacrylates (1.0 g, 5 or 6 or 5/6 composite) were dissolved in 1,2,3-trichloropropane (2.0 ml) and mixed with irgacure 819 (5.0 mg) as a photo-initiator. The solution was filtered through a 0.45-μm polytetrafluoroethylene syringe filter and was left standing for 2 h to remove air bubbles. The resultant liquid was spin-coated on a glass plate and then illuminated with a mercury lamp for 10 min under N2 atmosphere. The prepared film was baked for 5 min in a vacuum oven set at 50 °C. The curing reaction was monitored by IR absorption spectroscopy and Raman spectroscopy. The acrylic double bond rapidly disappeared during UV-irradiation, and the reaction was completed in 10 min with 85% conversion (see supplementary data).

Synthesis of diacrylate 5

A mixture containing compound 3a (1.53 g, 4.5 mmol), methacrylic anhydride (1.36 ml, 9.2 mmol), triethylamine (1.92 ml, 13.8 mmol) and a catalytic amount of 4-dimethylaminopyridine (27 mg) were added to dichloromethane (30 ml). The reaction mixture was stirred at room temperature for 12 h. The resulting mixture was diluted with dichloromethane (100 ml), washed with water (100 ml) and dried over anhydrous MgSO4. The solvent was removed by evaporation under a reduced pressure. The residue was subjected to column chromatography on silica gel using n-hexane and ethyl acetate as coeluents to yield diacrylate 5 (1.64 g, 77%). 1H NMR (CDCl3, 300 MHz): δ 1.99 (s, 6H), 2.21–2.42 (m, 6H), 3.17–3.29 (m, 8H, CH2-S), 4.22 (m, 4H, CH2-OCO), 4.78 (d, 2H, S-CH2-S), 5.58 (s, 2H, vinyl), 6.08 (s, 2H, vinyl); 13C NMR (CDCl3, 75 MHz): δ 18.7, 39.0, 39.1, 45.6, 47.6, 57.5, 64.9, 126.2, 136.2, 167.3; IR (KBr): νmax 3088, 2959, 2913, 1705, 1629, 1447, 1295.

Synthesis of diacrylates 6, 7

IPDI (1.0 g, 10.0 mmol), 2-hydroxyethyl methacrylate (2.6 g, 20.1 mmol) and dibutyltin dilaurate (0.60 ml, 1.0 mmol) were placed at a 100-ml two-neck flask with a reflux condenser and dissolved in anhydrous THF (30 ml) under N2 atmosphere. The reaction mixture was heated for 12 h at 50 °C. The mixture was diluted with ethyl acetate (100 ml), washed with water twice, dried over MgSO4 and filtered. The solvent was removed using rotary evaporation, and flash chromatography separation yielded pure diacrylate 6 with a 94% yield (4.5 g). Isomers: 1H NMR (CDCl3, 300 MHz): δ 1.88 (s, 3H), 1.00 (s, 6H), 1.14–1.31 (m, 2H), 1.55–1.81 (m, 4H), 1.98 (s, 6H), 2.82 (d, 2H, CH2-N), 3.77 (br, 1H, CH-N), 4.27 (s, 8H, CH2-O), 4.62 (m, 1H, NH-CH2), 4.83 (t, 1H, NH-CH), 5.58 (s, 2H, vinyl), 6.12 (s, 2H, vinyl); 13C NMR (CDCl3, 75 MHz): δ 18.2, 23.2, 27.5, 31.7, 35.0, 36.4, 41.5, 44.7, 46.2 (46.8), 53.5 (54.8), 62.3 (62.5), 62.9, 125.9, 136.0, 155.4, 156.6 (156.5), 167.1.

Compound 7 was similarly prepared with a 95% yield. Isomers: 1H NMR (CDCl3, 300 MHz): δ 1.89–0.95 (m, 12H), 1.00 (s, 3H), 1.34 (m, 4H), 1.41–1.81 (m, 6H), 2.82 (d, 2H, CH2-N), 3.74 (br, 1H, CH-N), 3.99 (m, 4H, CH2-O), 4.62 (br, 1H, NH-CH2), 4.91 (t, 1H, NH-CH); 13C NMR (CDCl3, 75 MHz): δ 13.5, 18.7, 22.8, 27.2, 30.7, 31.4, 34.5 (34.7), 36.0, 41.7 44.0, 46.1 (46.8), 54.3, 60.0, 62.0, 64.3 (64.1), 155.7, 156.9; IR (KBr): νmax 3330, 2948, 1688, 1544, 1244. LC-MS, m/z: 371.29 (M+1)+.

Urethane 7 (1.0 g) was dissolved in THF (8.0 ml) and stirred with 1.0 M-DCl in deuterium oxide (3.0 ml) at 60 °C for 2 h. The mixture was diluted with THF (15 ml) and dried over anhydrous MgSO4. The filtered solution was concentrated using a rotary evaporator to yield deuterium-substituted 7 (0.95 g, 94%). 1H NMR (CDCl3, 300 MHz): δ 1.84–1.00 (m, 12H), 1.02 (s, 3H), 1.34 (m, 4H), 1.41–1.81 (m, 6H), 2.90 (s, 2H, CH2-N), 3.79 (m, 1H, CH-N), 4.04 (m, 4H, CH2-O); IR (KBr): νmax 2969, 2484, 1699, 1441, 1162.

Results and Discussion

The refractive indices of the core and cladding layers in the optical waveguides were precisely adjusted for light confinement and long transmission. Fluorination, which reduces the intrinsic absorption loss from hydrocarbons over a specific near infra-red range, has been reported to lead to a decrease in the refractive index.12, 13 Sulfur–carbon covalent bonds increase the refractive index. Recently, cyclic 1,3-dithianes and 1,3-dithiolane developed through acrylic derivation were reported to have a high refractive index.11 The diol derivative of 1,3-dithiolane reacts with a diisocyanate to yield a urethane polymer as an alternative waveguide material. A norbornene derivative 1 was synthesized via reduction with LiAlH4 and successive protection from cis-5-norbornene-endo-2,3-dicarboxylic anhydride.11 The double bond of compound 1 was oxidized by O3 to dialdehyde 2, where the configuration was fixed. The acid-catalyzed condensation of compound 2 with 1,2-ethanedithiol and successive benzoyl-deprotection yielded diol 3a. Similarly, the diols 3b and 3c were synthesized by the condensation of compound 2 using 2-mercaptoethanol and ethylene glycol, respectively. All the carbon configurations in 3a were retained to yield a single product, which was confirmed by the doublet peak at 4.6 p.p.m. (S-CH2-S) from 1H-NMR spectroscopy. However, unwanted isomerization was observed in both compounds 3b and 3c, which occurred at the carbon atoms adjacent to the formyl groups of compound 2 through keto-enol tautomerization. Diols 3b and 3c were isolated as mixtures containing ∼50 and 10% isomer, respectively, from 1H-NMR spectroscopy. The two doublets at 5.01 and 5.15 p.p.m. (S-CH2-O) and two doublets at 5.02 and 4.73 p.p.m. (O-CH2-O) indicated the formation of isomers of compounds 3b and 3c, respectively. The slow condensation of compounds 3b and 3c allowed the isomerization of compound 2 over 8 h. Isomerization was not observed during the synthesis of compound 3a owing to rapid condensation using 1,2-ethanedithiol. Linear polyurethanes 4(a–c) were synthesized when the diols 3(a–c) were reacted with IPDI in the presence of catalytic dibutyltin dilaurate. Figure 1 shows the synthetic procedure of the polyurethanes and diols. The reaction of IPDI and 2-hydroxyethyl methacrylate (HEMA) produced diacrylate 6. UV-irradiation of a mixture of IPDI and diacrylate 5 produced the acrylic polymer films. The polymerization occurred with ∼85% conversion based on the IR and Raman spectroscopy results. The band intensity of C=C bonds of 15% remained near 1600 cm−1, which originates from the unreacted acrylic bonds after the UV-polymerization. The content of a urethane linker in the film 4d was determined from the monomer ratio. The linear polyurethanes 4(a–c) were analyzed using gel permeation chromatography. Table 1 lists the molecular weights and polydispersity of the three polyurethanes. Differential scanning calorimetry and thermogravimetric analysis were performed to examine the thermal properties of the polymer samples, as shown in Figure 2. The glass transition of polyurethanes 4b and 4c occurred at 71 and 78 °C, respectively, and the glass transition of 1,3-dithiolane polyurethane 4a occurred at 133 °C. Thermal degradation of the polyurethanes was determined to have occurred at the temperature where a 5 wt% loss was recorded. The polyurethanes were stable up to ∼250 °C. Thermoplastic polyurethanes belong to a class of thermoplastic elastomers and exhibit high thermal expansion. The refractive index of most organic polymers decreases with increasing temperature owing mainly to thermal volume expansion. Table 2 lists the refractive index of all the prepared polymer films at important wavelengths for current optical communication. The birefringence (Δ=nTE−nTM) of the urethane polymers, which was derived from the mode dependence of the propagating light, was <0.0004. A low birefringence is often achieved with amorphous materials or a strain relief process, which is favorable for a low glass transition material.4, 14 The prepared polyurethanes are promising waveguide materials, even though the polyurethanes exhibited higher glass transitions than the conventional rubbery polyurethanes. 1,3-dithiolane-attached polyurethane 4a exhibited the highest refractive index among the three polyurethanes examined. The increase in refractive index was attributed to the presence of covalent sulfur. The thermal dependence of the refractive index was examined, and a linear correlation was obtained. The refractive index decreases gradually with increasing temperature. The linear graph was derived using the least squares method, which produced a slope (dn/dT), representing the TOC. The three linear polyurethanes in Figure 3 exhibited unequal thermal behavior with different slopes. The slopes were expected to be similar because all the polyurethanes are composed of similar urethane linkers, which control the thermal volume expansion. Polyurethane 4a exhibited a large TOC of −3.63 × 10−4. The coefficient of polymer 4c was unusually small (−2.79 × 10−4) and that of polymer 4b increased to some degree. The reason for the different thermo-optic effects is unclear but involved a contribution from 1,3-oxolane or 1,3-dithiolane.

Figure 1
figure 1

The synthesis of linear polyurethanes and UV-acrylic polymers.

Full size image
Table 1 Thermal and molecular weight analyses of the linear polyurethanes
Full size table
Figure 2
figure 2

The refractive index change as a result of temperature and thermo-optic coefficient (dn/dT) of linear polyurethanes (left) and UV-cured poly(methacrylate)s (right) at 1550 nm.

Full size image
Table 2 The refractive index and thermo-optic effect of the polyurethanes and acrylic polymers
Full size table
Figure 3
figure 3

The thermal behavior of linear polyurethanes measured by differential scanning calorimetry (left) and thermogravimetric analysis (right). A full color version of this figure is available at Polymer Journal online.

Full size image

The thermo-optic effect of the UV-cured acrylic polymers was examined. The acrylic monomer 6 was designed to have urethane bonds under the expectation of a large thermo-optic effect. A UV-polymer 4d (0/10) prepared from monomer 6 had a dn/dT value of −1.98 × 10−4, which is smaller than that of the linear polyurethanes. A diacrylate monomer undergoes cross-linking under UV-polymerization conditions, and the resulting network structure limits thermal expansion. The contribution of urethane to the TOC of polymer 4d (0/10) was clear because a homo-polymer 4d (10/0) from diacrylate monomer 5 yielded dn/dT=−1.58 × 10−4. The composite polymers from the monomer mixtures of compounds 5 and 6 provided an interesting result. Two composite polymers, 4d (2/8) and 4d (6/4), exhibited a higher thermo-optic effect than the homo-polymer 4d (0/10) despite the low monomer 6 content. No quantitative relationship between the monomer 6 content and the thermo-optic effect of the composite polymers was observed. Polymer 4d (4/6) exhibited the lowest thermo-optic effect among the three composite polymers examined. The thermal volume expansion of the polymers was calculated using the Lorentz-Lorenz theory, where the refractive index is expressed as the density and polarizability.15 The volume expansion coefficient, β, provides a clear comparison of the thermal effect in Table 2. Urethane bonds were observed to enhance the thermo-optic effect, which is not proportional to the quantity of the urethane bonds of cross-linked poly(methacrylate)s.

Intrinsic absorption by chemical bonds has been observed in the material loss of a waveguide that is reduced by fluorine substitution. The presence of N-H in the urethane bonds will generate optical loss, as suggested by the O-H group,16 however, the loss is predicted to be insignificant at specific wavelengths, such as 1100 and 1310 nm through spectral analysis. The analysis was performed with a liquid small molecule because of the low absorption coefficients of the hydrocarbons in the near infrared region. Figure 4 presents the near infrared absorption spectrum of urethane 7, which was prepared from the condensation of IPDI and n-butanol. Several absorptions peaks were assigned to the overtone or combination vibrations of C-C, C-H and N-H bonds. A strong band at 1478 nm was assigned to N-H absorption, which disappeared after an N-D exchange reaction under a deuterium-substitution condition of compound 7. As shown in Figure 4, no absorption was observed near 1100 and 1310 nm, indicating that the additional loss from the N-H combined overtones is very low in the two wavelengths. For example, the propagation loss of polymer 4d (2/8) film was 0.090 (1100 nm), 0.244 (1310 nm) and 0.288 dB cm−1 (1550 nm) through loss measurements using an immersion oil.17 The losses of the N-D exchanged film were improved to 0.078 (1310 nm) and 0.085 dB cm−1 (1550 nm).

Figure 4
figure 4

The near infrared absorption spectrum of urethane 7 and a deuterium-substituted urethane solution (5 wt%) in CDCl3 (left) and mid-IR absorption spectrum (right). A full color version of this figure is available at Polymer Journal online.

Full size image

Conclusions

Polyurethanes were evaluated as thermo-optic waveguide materials. The linear polyurethane in the study exhibited a large TOC of −3.63 × 10−4. UV-cured acrylic polymers were prepared using a urethane-containing acrylic monomer for improved thermo-optic effect. The high urethane content in the acrylic polymers promoted thermal volume expansion. The D-isotope exchange of the urethane N-H group not only retains H-bond interactions for the thermal expansion property of polyurethanes but also reduces near infrared- absorption as a waveguide material. The urethane-containing polymers exhibited low optical losses and low birefringence as promising optical waveguide materials.