1 Introduction

Polyurethane (PU) was discovered by Otto Bayer, a German chemist, and his coworkers in 1937 [1, 2]. PU materials are employed as alternatives to nylon and rubber because of their excellent mechanical properties, such as surface stability, flexibility, and adhesiveness [3,4,5,6,7,8,9,10]. PU materials are currently used in a wide range of applications in manufacturing and daily life, with a global market volume of over 25 million tons and a market value of approximately 65 billion USD [11, 12]. PUs are generally produced by reacting isocyanates with polyols, where the characteristics of the polyols, such as their hydroxyl functionalities, molecular weights (MWs), and terminal hydroxyl groups, have crucial effects on the chemical and physical properties, as well as the performances of the end products [13].

Polypropylene glycols (PPGs) derived from propylene oxide are the most widely used precursors in PU manufacturing because of their versatility and abundance. PPGs are commercially produced via anionic ring-opening polymerization of propylene oxide using basic catalysts (e.g., KOH) [14]. The ring-opening of propylene oxide can proceed via both α- and β-cleavage routes; however, in base-catalyzed anionic polymerization, β-cleavage of the epoxy ring is favored, resulting in PPGs with predominant secondary hydroxyl groups (up to 95%). Furthermore, basic catalysts can produce PPGs with significant monol contents or unsaturation values. These factors significantly affect the performance of PPG polyols in PU synthesis [15].

Considerable effort has been devoted to increasing the primary –OH content of the final PPG products and diminishing their monol contents. Especially, Lewis acid catalytic systems and double metal cyanide (DMC) catalysts have been explored [16,17,18,19]. Lewis-acid-mediated systems, such as boron compounds, can be used to afford PPG polyols with significant primary hydroxyl contents compared to those of the secondary congeners (50–70%). However, these catalytic systems are rarely used in PPG production because of their relatively high costs, difficulty in handling, and the possibility of side reactions. DMC compounds are well-known catalysts that are used industrially to produce high-quality PPG polyols with extremely low contents of unsaturated impurities. However, polyether polyols produced via DMC catalysis contain few primary –OH groups [20].

Another approach for increasing the primary –OH contents of PPG polyols is via post-synthesis modification of PPG by ethylene oxide end-capping using KOH or by propylene oxide end-capping wherein DMC and Lewis acidic boron compounds are combined in a dual catalytic system [21, 22]. DMC belongs to one of the oldest classes of inorganic complexes, also known as Prussian blue analogs (PBAs). These complexes have been employed in various fields such as catalysis [23,24,25,26,27,28,29,30,31], battery materials [32,33,34,35,36,37,38,39,40,41], biomedical applications [42,43,44], electrochemical sensors [45,46,47], and sorbents [48, 49]. Recently, we found that PBAs effectively catalyze the ring-opening of lactones using both monoinitiators and macroinitiators [50], which motivated us to investigate these complexes for the synthesis of lactone-capped PPG polyols.

In this study, a series of PBA catalysts is prepared in the absence of organic complexing agents using various central metal atoms (Co, Fe, Mn, and Zn). The activities of the as-prepared catalysts are evaluated for the ring-opening of ε-caprolactone (CL) using ethylene glycol (EG) as an initiator. Kinetic studies are conducted using a propylene glycol initiator to gain insight into the reactivity of the primary and secondary hydroxyl groups in the ring-opening of CL, paving the way for the synthesis of CL-capped PPG polyols. To highlight the advantages of the current approach, the reactivities of the resulting CL-capped polyols for PU synthesis are evaluated and compared with those of conventional PPG.

2 Materials and methods

2.1 Materials

Potassium hexacyanocobaltate(III) (K3Co(CN)6, ≥ 97%) and hexamethylene diisocyanate (HDI, ≥ 98%) were purchased from Merck Korea (Seoul, Korea). Phosphoric acid (≥ 89%) was purchased from Tokyo Chemical Industry Co., Ltd. (Incheon, Korea). Zinc(II) chloride anhydrous (≥ 98%), ethylene glycol (EG, 99%), and ε-caprolactone (CL, 99%) were obtained from Fisher Scientific Korea (Incheon, Korea). Nitric acid (60–62%) was purchased from Junsei Chemical Co., Ltd. (Tokyo, Japan). Potassium cyanide (≥ 97.5%) was obtained from Duksan Pure Chemicals (Incheon, Korea). Manganese(II) chloride tetrahydrate (≥ 98%) and propylene glycol (PG, ≥ 99.5%) were purchased from DaeJung Chemical Co. (Gyeonggi-do, Korea). PPG (MW: 500, functionality: 2) was obtained from Kumho Petrochemical Co. (Ulsan, Korea). Propylene oxide was obtained from SKC Polyurethanes, Inc. (Ulsan, Korea). EG, PG, and PPG were distilled prior to use. Potassium manganese cyanide (K3Mn(CN)6) was prepared according to literature methods, recrystallized in warm 10% potassium cyanide solution, and dried before use [51]. Unless otherwise mentioned, all other reagents were used without further purification.

2.2 Catalyst Preparation

The PBA catalysts were prepared by mixing an aqueous solution of a metal salt with a metal cyanide salt. A stoichiometric ratio of the metal salt to metal cyanide salt was used to ensure the formation of PBAs with the desired composition. As an example, a solution of ZnCl2 (0.31 g, 2.28 mmol) in distilled water (2.5 mL) was added to a solution of K3Mn(CN)6 (0.5 g, 1.52 mmol) in distilled water (2.5 mL) under vigorous stirring at 25 °C. The resulting precipitate was washed with 10 mL distilled water and dried to produce Zn3[Mn(CN)6]2. Zn3[Co(CN)6]2 and Mn3[Mn(CN)6]2 were prepared using the same method.

2.3 Catalytic Reaction

All reactions were performed using standard Schlenk techniques. For the ring-opening reaction of CL, prescribed amounts of the PBA catalyst and hydroxyl-containing initiator were added to a 10 mL round-bottom flask, which was then immersed in an oil bath and heated to 90 °C under nitrogen flow for 30 min to remove traces of water. The temperature was raised to 160 °C, and prescribed amounts of CL were added under continuous stirring. Occasionally, small amounts of the reaction mixture were collected for further analysis.

Using a 50 mL round-bottom flask, the reactions of polyols with HDI was investigated. Typically, a prescribed amount of polyol was added to the reactor and purged with nitrogen for 30 min at 90 °C to remove traces of water. The reactor was then cooled to 60 °C, and the prescribed amount of HDI was added under vigorous stirring and the resulting mixture was allowed to react for 3 h, and samples of the reaction mixture were collected at different intervals to determine the conversion of –NCO groups by Fourier-transform infrared (FTIR) analysis.

2.4 Characterization

FTIR data were acquired using a Shimadzu IR Prestige 21 spectrometer (Shimadzu Co., Tokyo, Japan). The crystal structures of the catalysts were determined using a RINT2000 wide-angle diffractometer (Rigaku Co., Tokyo, Japan) equipped with a 185 mm radius goniometer and Cu-Kα radiation. The morphologies and particle sizes of the catalysts were investigated by scanning electron microscopy (SEM) using an S-3000 H microscope (Hitachi Ltd., Tokyo, Japan). Nuclear magnetic resonance (NMR) analysis was conducted using an INOVA 400 NMR spectrometer (Varian Inc., California, USA). Chemical shifts are presented as parts per million (ppm) relative to the peaks of the residual monomer used as the internal standard. Gel permeation chromatography (GPC) was performed using a Waters 150 instrument (Agilent Technologies, Santa Clara, California, USA). Tetrahydrofuran was used as the diluent at a flow rate of 1 mL min− 1, at 40 °C, and polystyrene was used as the standard for calibration.

3 Results and Discussion

3.1 Catalyst Characterization

Figure 1a and Supplementary Fig. S1 show the FTIR spectra of the prepared PBA catalysts. The characteristic signals corresponding to the C ≡ N stretching vibrations of Zn3[Co(CN)6]2, Zn3[Mn(CN)6]2, and Mn3[Mn(CN)6]2 were observed at approximately 2184, 2220, and 2146 cm− 1, respectively. Additionally, signals attributed to the metal − CN bending vibrations were located at approximately 454, 466, and 503 cm− 1, respectively. The variations in the intensities and positions of these characteristic peaks provide insight into how the different central metal atoms (i.e., Zn, Co, and Mn) interact with cyanide in each PBA. The broad peaks at approximately 3550–3000 and 1680–1590 cm− 1 are attributed to the –OH stretching vibrations and H–O–H deformation of the physisorbed water involved in a hydrogen-bonded network [52]. In contrast, the sharp peaks at approximately 3700–3550 cm− 1 are assigned to the –OH stretching vibrations of chemisorbed water molecules involved in metal‒OH coordination [30]. The results of the FTIR analysis are presented in Supplementary Table S1.

Fig. 1
figure 1

Characterization of the representative PBA catalysts: (a) FTIR spectra, (b) XRD patterns, (c) TGA curves, (d) SEM images

Full size image

The crystal structures and crystal phases of the as-prepared PBAs were determined using X-ray diffraction (XRD) analysis. As shown in Fig. 1b and Supplementary Fig. S2, the Zn3[Co(CN)6]2, Co3[Mn(CN)6]2, Fe3[Mn(CN)6]2, and Mn3[Mn(CN)6]2 compounds exhibited a pure face-centered cubic structure (space group Fm-3m) with characteristic reflection peaks corresponding to the (111), (200), (220), (400), (420), (422), (440), (600), and (620) planes. Meanwhile, Zn3[Mn(CN)6]2 exhibited a monoclinic structure in the Pn-3m space group, with reflections corresponding to specified planes at 2θ = 21.3° (011), 26.0° (111), 37.2° (112), 43.4° (022), 48.8° (013), and 58.5° (123). The differences in the coordination environments of the central metal atoms had a noticeable impact on the crystalline structure and size of the catalyst clusters, as evidenced by the variations in the intensity and position of these characteristic reflections.

The thermal stability of the PBA catalysts was determined using thermogravimetric analysis (TGA) (Fig. 1c and Supplementary Fig. S3). All catalysts exhibited two-stage decomposition curves corresponding to the decomposition of the absorbed water and –C ≡ N– bonds. Zn3[Co(CN)6]2 exhibited the highest thermal stability, completely decomposing at approximately 550 °C, which is much higher than the decomposition temperature of Zn3[Mn(CN)6]2 (approximately 370 °C) and Mn3[Mn(CN)6]2 (approximately 200 °C). The amount of absorbed water in the PBA catalysts decreased in the order: Zn3[Co(CN)6]2 (11.6%) > Mn3[Mn(CN)6]2 (6.8%) > Zn3[Mn(CN)6]2 (2.2%), which is in good agreement with the FTIR data.

The morphologies and particle sizes of the catalysts were determined using SEM. As shown in Fig. 1d and Supplementary Fig. S4, Zn3[Co(CN)6]2 formed highly crystalline cubic particles (approximately 513.2 nm). The particle size of PBA decreased significantly when the central Co atom was replaced by Mn, as observed for Co3[Mn(CN)6]2 (56.7 nm), Fe3[Mn(CN)6]2 (63.5 nm), Zn3[Mn(CN)6]2 (53.3 nm), and Mn3[Mn(CN)6]2 (21.4 nm). Additionally, the edges of Zn3[Mn(CN)6]2 and Mn3[Mn(CN)6]2 underwent a significant shape transformation compared to those of Zn3[Co(CN)6]2, with slight debris on the plane surface, likely due to the dissociation of [Mn(CN)6]−3 and etching of the resulting PB crystal. The characterization data for the PBA catalysts are presented in Table 1.

Table 1 Structural properties of the representative PBA catalysts
Full size table

3.2 Catalyst Screening

The activity of the PBA catalysts was first investigated in the ring-opening of CL using EG as the hydroxyl initiator (Fig. 2). The reactions were conducted under a nitrogen atmosphere at 160 °C for 3 h, and the crude reaction mixture was analyzed at different intervals from 0.5 to 3 h using 1H NMR. The conversion of CL was determined based on the integration of the signals related to the methylene protons of CL at approximately 1.75, 1.84, 2.63, and 4.21 ppm (Supplementary Figs. S5–S15). The control reaction, which was performed in the absence of a catalyst, indicated negligible conversion of CL, and no adducts were observed. Among the catalysts, only Mn3[Mn(CN)6]2 and Zn3[Mn(CN)6]2 showed remarkable activity for the ring-opening of CL, whereas the other PBA catalysts exhibited relatively low activity (maximum CL conversion of 36% after 3 h). The Mn3[Mn(CN)6]2 catalyst exhibited the highest activity, achieving > 90% CL conversion after 30 min, whereas Zn3[Mn(CN)6]2 required 3 h to achieve a similar CL conversion.

Fig. 2
figure 2

Screening of various PBA catalysts for the ring-opening of CL. Reaction conditions: cat. loading = 0.3 mol%, EG = 0.9 mmol, CL = 9 mmol, Tp = 160 °C, t = 3 h

Full size image

3.3 Kinetics of the ring-opening of CL Using Mn3[Mn(CN)6]2

Kinetic studies were conducted using the Mn3[Mn(CN)6]2 catalyst to evaluate the reactivity of the primary and secondary hydroxyl functional groups of EG and PG in the ring-opening of CL. Note that the ring opening of CL did not proceed in the absence of an initiator, as evidenced by the 1H NMR spectra collected after 3 h of the reaction (Supplementary Fig. S16). The conversion of CL and the initiators over time was monitored using in situ 1H NMR spectra collected at different intervals (Supplementary Figs. S17 and S18). The apparent rate constants (kapp) of the reactions were inferred from the plots of ln([CL0]/[CL]) versus time (Fig. 3). The linear relationship between ln([CL0]/[CL]) and time indicates that the ring opening of CL using the EG initiator is first-order with respect to the CL concentration, whereas the dependence on CL conversion in the reaction using the PG initiator was initially non-linear but transitioned to a linear regime at later stages. The deviation from this dependence after the early stages is attributed to the presence of both primary and secondary hydroxyl groups. The primary and secondary –OH groups were completely consumed after approximately 10 and 25 min, respectively, as evidenced by the disappearance of the 1H NMR signals at 1.15, 3.39, and 3.91 ppm, corresponding to the methyl, methylene, and methine protons adjacent to the hydroxyl groups, respectively (Fig. 4). After complete consumption of the secondary OH groups, the ring opening of CL proceeded solely via the primary OH groups and followed a first-order reaction. The reaction conditions were optimized by varying the molar ratio of CL to PG. As shown in Fig. 5 and Supplementary Fig. S19, complete consumption of the secondary –OH group was achieved when the CL/PG ratio was equal to or greater than 3:1.

Fig. 3
figure 3

Time-conversion plots of the ring-opening reaction of CL using EG or PG initiator and Mn3[Mn(CN)6]2 catalyst. Reaction conditions: cat. loading = 0.3 mol%, initiator = 5 mmol, CL = 50 mmol

Full size image
Fig. 4
figure 4

Expanded in situ 1H NMR (400 MHz, CD3Cl3) spectra of the crude reaction mixture in the ring-opening reaction of CL using PG and the Mn3[Mn(CN)6]2 catalyst, acquired at the early stage of the reaction

Full size image
Fig. 5
figure 5

Expanded 1H NMR (400 MHz, CD3Cl3) spectra of the crude reaction mixture in the ring-opening reaction of CL using various molar ratios of CL to PG

Full size image

3.4 Synthesis of CL-terminated PPG Polyols

The ring-opening of CL using PPG was similar to that using PG, except that the content of secondary –OH groups in PPG was higher than that in PG (up to 95%). The former required a larger amount of CL to completely convert the secondary alcohols. Based on the kinetics data presented above, a series of CL-terminated PPG polyols was synthesized by varying the molar ratio of CL to PPG using Mn3[Mn(CN)6]2 as the catalyst. A comparison of 1H NMR spectra of PPG and the resulting CL-terminated PPG polyols is shown in Supplementary Fig. S20. The conversion of the CL and –OH groups of PPG at various CL/PPG ratios is reflected in the expanded NMR spectra in Fig. 6. Accordingly, signals corresponding to the methyl and methylene protons adjacent to the terminal –OH groups of PPG overlapped with those of the repeating units at approximately 1.08–1.18 ppm and 3.15–3.54 ppm, respectively, whereas signals related to the methine protons were observed at 3.93 ppm. After the reaction with CL, the methine signals of the proton were gradually depleted, while new signals appeared at 1.22 and 5.02 ppm, assigned to the methyl and methine protons, respectively, adjacent to the reacted terminal alcohols. The intensity of these signals increased in proportion to the CL/PPG ratio and reached a maximum when the CL/PPG ratio was equal to or higher than 5:1. The GPC curves of the resultant CL-capped PPG polyols exhibited a narrow MW distribution (1.12–1.31), with MWs of 600–1800 g mol− 1, in good agreement with the calculated values (Supplementary Fig. S21). The results obtained in the synthesis of the CL-capped PPG polyols are summarized in Table 2.

Fig. 6
figure 6

Expanded 1H NMR (400 MHz, CD3Cl3) spectra of the crude reaction mixture in the ring-opening reaction of CL using various CL: PPG ratios

Full size image
Table 2 Results for the synthesis of CL-terminated PPG polyols
Full size table

The advantages of producing CL-terminated PPG polyols over conventional PPG were investigated using a model reaction in which HDI was used to synthesize polyurethane. The reactions were performed under vigorous stirring at 60 °C for 3 h in the absence of a catalyst. The reactions were monitored by in situ 1H NMR analysis to determine the NCO conversion over time (Supplementary Figs. S22–S26). For conventional PPG, the signal assigned to the methylene protons adjacent to the ‒NCO groups at 3.32 ppm remained almost unchanged, indicating relatively poor reactivity for PU synthesis (NCO conversion of 2.7% after 3 h). As shown in Fig. 7, the NCO conversion increased in proportion to the CL/PPG ratio (from 10.7% for PPG-CL1 (CL/PPG = 1:1) to 41% for PPG-CL3 (CL/PPG = 3:1)) and reached a maximum of 48% at CL/PPG = 5:1 (PPG-CL5). The reactions were conducted without any catalyst, where an excess of HDI to polyol (HDI/polyol = 2:1) was used. The in situ FTIR spectra of the crude reaction mixture of PPG-CL and HDI were further investigated to corroborate the NMR data (Supplementary Figs. S27–S31). The consumption of NCO was monitored based on the reduction of the signals corresponding to the ν(–NCO) vibrations at approximately 2260 cm− 1. These results were in good agreement with the NMR data, highlighting the superior reactivity of CL-capped polyols compared to conventional polyols.

Fig. 7
figure 7

Reactivity of CL-terminated PPG polyols with HDI. Reaction conditions: polyol = 5 g, HDI/polyol molar ratio = 2 : 1, Tp = 60 °C, t = 3 h

Full size image

4 Conclusions

A series of Prussian blue analogs catalysts was investigated for the synthesis of α,ω-primary hydroxyl-terminated polyols via the ring-opening reaction of ε-caprolactone. The Mn3[Mn(CN)6]2 catalyst exhibited the highest activity and was used for further kinetic studies. The reactivity of the primary and secondary hydroxyls toward the ring opening of ε-caprolactone was determined using ethylene glycol and propylene glycol as representative initiators. The reaction using the ethylene glycol initiator proceeded via a simple coordinative mechanism with a first-order dependence on the ε-caprolactone concentration, whereas the reaction using the propylene glycol initiator was characterized by a two-stage kinetics plot. In the early stages, both the primary and secondary hydroxyl groups initiated the reaction. After complete consumption of the secondary hydroxyl groups, the reaction proceeded via a mechanism similar to that with the EG initiator. A ε-caprolactone/propylene glycol ratio of 3/1 was found to be suitable for the complete conversion of the secondary hydroxyls. Based on these findings, α,ω-primary hydroxyl-terminated polyols were successfully synthesized using polypropylene glycol as a macroinitiator. These polyols exhibited superior performance in polyurethane synthesis compared to the original polypropylene glycol, even in the absence of a catalyst. This work offers a promising approach for the synthesis of α,ω-primary hydroxyl-terminated precursors for a wide range of applications in the polyurethane industry.