Bioinspired All-Polyester Diblock Copolymers Made from Poly(pentadecalactone) and Poly(3-hydroxycinnamate): Synthesis and Polymer Film Properties

Abstract

A bioinspired diblock copolymer was synthesized from pentadecalactone and 3-hydroxy cinnamic acid. Poly(pentadecalactone) (PPDL) with a molar mass of up to 43,000 g mol^-1 was obtained by ring-opening polymerization initiated propargyl alcohol. Poly(3-hydroxy cinnamate) (P3HCA) was obtained by polycondensation and end-functionalized with 3-azido propanol. The two functionalized homopolymers were connected via 1,3-dipolar Huisgen addition to yield the block copolymer PPDL-triazole-P3HCA. The structure the block copolymer was confirmed by proton NMR, FTIR spectroscopy and GPC. By analyzing the morphology of polymer films made from the homopolymers, from a 1:1 homopolymer blend, and from the PPDL-triazole-P3HCA block copolymer, clearly distinct micro- and nanostructures were revealed. Quantitative nanomechanical measurements revealed that the block copolymer PPDL-triazole-P3HCA had a DMT modulus of 22.3 ± 2.7 MPa, which was lower than that of the PPDL homopolymer (801 ± 42 MPa), yet significantly higher than that of the P3HCA homopolymer (1.77 ± 0.63 MPa). Thermal analytics showed that the melting point of PPDL-triazole-P3HCA was similar to PPDL (89-90 °C), while it had a glass transition was similar to P3HCA (123-124 °C). Thus, the semicrystalline, potentially degradable all-polyester block copolymer PPDL-triazole-P3HCA combines the thermal properties of either homopolymer, and has an intermediate elastic modulus.

1. Introduction

Cutin and suberin are hydrophobic, bioactive and water-impermeable biopolyesters used by nature for compartmentalization in plants. They are mostly found in the outer plant layers, e.g. their bark, leave cuticles, and roots. There, cutin and suberin protect plants against desiccation, microbial infections, and insects.^[1] They can be extremely stable materials, as evidenced by the suberin-based material cork, yet they are degradable by enzyme-catalysed hydrolysis and other degradation processes. The peculiar property combination of long-term stability, degradability, and antimicrobial activity makes cutin- and suberin-like polyesters an attractive object of study, and a potentially interesting group of biomaterials.

Cutin is the main structural component of plant leave cuticles. It is a branched polyester consisting mostly of fatty acids, alcohols and diacids with 16 to 24 carbon atoms. Many of these carry further hydroxyl or epoxy groups. Notably, cutin can contain up to 32 % ω-hydroxy fatty acids. Additionally, it can contain a low amount of glycerol (1-14 %) and phenols (0-1 %).^{[1, 2]} Suberin is found mostly in bark or the outer skin of roots, and also forms during plant wound healing to provide an effective barrier to the plant environment.^{[3, 4]} It consists of a polyester-based, aliphatic part and a non-hydrolysable polyaromatic part. Depolymerisation experiments have shown that the structure of the polyaliphatic domain is similar to cutin, consisting of aliphatic fatty acids with 16 to 32 carbon atoms. It also contains 14-26 % glycerol, and 1-10% phenolic compounds ^{[2, 5]} The structure of the polyaromatic domain is not yet fully understood. The available experimental evidence shows that there are two different kinds of aromatic domains. One consists of highly cross-linked, lignin-like phenolic compounds, and is attached to the primary and tertiary cell wall of suberinized cells (Figure 1a).^[6] The other is an aromatic polyester made from mainly hydroxycinnamic acids, glycerol, and hydroxyl-substituted fatty acids. This part is linked to the polyaliphatic domains inside the secondary cell wall (Figure 2a).^{[2, 5]} Together, the polyaliphatic and polyaromatic part in the secondary cell wall form lamellar nanostructures, as observed with transmission electron microscopy (TEM, Figure 1a). Exact details of this lamellar structure are still under debate,^[7] yet it is clear that the aliphatic polyester and aromatic components are spatially separated.^{[5, 7]} Several models for the macromolecular architecture of suberin have been suggested, for example the one for suberinized potato cell walls shown in Figure 1b.^[8]

Figure 1.

MALDI-TOF spectrum of P3HCA (M_n = 2700 g mol^-1): a) full spectrum; b) close-up ofthe region between m/z from 1800 to 1945 which showed chains with different end groups. DCTB was used as ionization matrix material, and the polymer was ionized with silver ions.

Figure 2.

¹H-NMR spectra of a) alkyne-functionalized PPDL (PPDL-alkyne), b) azide-functionalized P3HCA (P3HCA–azide) and c) the block copolymer PPDL-triazole-P3HCA (all in CDCl₃ at 250 MHz); d) section of the MALDI-TOF mass spectrum of P3HCA-azide with peak assignment (matrix: DCTB; ion: silver); e) GPC curves of the block copolymer PPDL-triazole-P3HCA (black line) and the corresponding homopolymers PPDL-alkyne (dashed line) and P3HCA-azide (dotted line).

The lamellar structure of suberin looks strikingly similar to the morphology of many block copolymers at a one to one block volume ratio. Inspired by this similarity, the aim of this work was to produce a synthetic all-polyester block copolymer consisting of an aliphatic ester block and an aromatic ester block (Scheme 1), and to investigate its phase behavior. This is synthetically far from trivial. While there is a rich literature on polyester-based block copolymers with one polyester block,^[9–13] attempts to obtain an all-polyester block copolymer are scarce.^{[14, 15]} This is on account of polymerization method-related problems: Polycondensation, the traditional method to produce polyesters (either from hydroxyalkanoates, or diacids with diols)^[16–20] does not yield defined blocks. More recently, living ring-opening polymerization (ROP) of cyclic lactones has been established,^{[20, 21]} which offers better control over the polymer structure, including the molecular mass, polydispersity, end-groups and polymer architecture. Even larger lactones can be polymerized via ROP; for example, poly(pentadecalactone) ^{[22, 23]} can be either obtained by enzyme catalysis,^{[22, 24, 25]} organocatalysis,^{[23, 26]} or metal catalysis.^[27–29] Due to the precise control over the end groups of ROP polymers, it should be possible to obtain block copolymers via sequential ROP, using the first polyester block as a macroinitiator for the second block.^[14] However, recent literature^{[30, 31]} as well as our own attempts show that this typically leads to transesterification between the blocks and thus ill-defined, not phase-separating polymer nanostructures. For this reason, one of the strategies to obtain diblock polyesters presented in this work was to separately end-functionalize each block with functional groups that could undergo an efficient coupling reaction, e.g. by 1,3-dipolar Huisgen addition (“click reaction”),^[32–34] as described for other block copolymers.^{[35, 36]}

Scheme 1.

Homopolymerization of a) pentadecalactone (PDL) via ring-opening polymerization), and b) trans-3-hydroxycinnamic acid (3HCA) via thermal condensation reaction. The first strategy was to obtain block copolymers from PDL and 3HCA c) using poly(pentadecalactone) (PPDL) as a macroinitiator for 3HCA polymerization. However, this yielded only PPDL and poly(3-hydroxycinnamate) (P3HCA) homopolymers. The second strategy was to end-functionalize the homopolymers for a poly-polymerization “click” reaction. d) For this, end-functionalized PPDL was obtained by initiation of PDL with propargyl alcohol (PA) to yield PPDL-alkyne. e) P3HCA was functionalized with 3-azidopropanol obtain P3HCA-azide, e). f) The functionalized homopolymers were connected via azide-alkyne 1,3-dipolar cycloaddition using a copper catalyst.

In the field of aromatic polyesters, the available structures are rather scarce, and the molecular mass of the obtained polymers are typically low. The polyaromatics most relevant in our context are alkyl and arylphenols, catechols or hydroxycinnamic acids. However, their oxidative polycondensation suffers from numerous side reactions.^{[30, 37–43]} In particular, polycondensations of hydroxycinnamic acids or its derivatives gave products with low molecular masses and broad polydispersity.^[44–46]

In the following, we describe our strategies to obtain all-polyester block copolymers made from pentadecalactone as the aliphatic block, and 3-hydroxy cinnamic acid as the aromatic block: first via a sequential approach, then by separate polymerization of each component, followed by end-functionalization and covalent connection via 1,3-dipolar Huisgen addition (Scheme 1).

2. Experimental

2.1. Materials

All chemicals were obtained as reagent grade from Sigma-Aldrich (St. Louis, MO, USA), Carl Roth (Karlsruhe, Germany) or Alfa Aeser (Haverhill, MA, USA) and used as received. High performance liquid chromatography (HPLC) grade solvents were purchased dry from Carl Roth and used as received. Trans-3-hydroxycinnamic acid (3HCA) was purified by column chromatography and pentadecalactone (PDL) was sublimed and dried under high vacuum for three days. Propargyl alcohol (PA) and benzyl alcohol (BA) were dried over calcium hydride and distilled under reduced pressure. Triazabicyclodecene (TBD) was recrystallized from diethyl ether and dried under high vacuum for three days. Toluene, chlorofom, PDL, PA, BA and TBD were stored and handled under nitrogen in a glovebox (MBRAUN, Garching, Germany). All aromatic species were handled and stored under exclusion of light.

2.2. Instrumentation

Gel permeation chromatography (GPC, in chloroform, calibrated with poly(methyl methacrylate) and polystyrene standards) was performed on PSS SDV 100, 1000 and 10000 Å columns (PSS, Mainz, Germany) using a 1260 Infinity RI detector (Agilent Technologies, Santa Clara, CA, USA). NMR spectra were recorded on a Bruker (Billerica, MA, USA) 250 MHz spectrometer using CDCl3 as solvent and tetramethylsilane (TMS) as internal reference. Fourier-transform infrared spectroscopy (FTIR) spectra were taken using a Cary 630 FTIR Spectrometer (Agilent Technologies, Santa Clara, CA, USA). The glass transition temperature and melting temperature were measured by differential scanning calorimetry (DSC). The spectra were recorded on a Phönix DSC 204 F1 device (Netzsch, Selb, Germany) in a temperature range of -20 to 250 °C and with a speed of 5 K/min. MALDI-TOF mass spectrometry measurements were conducted using an autoflexTOF/TOF (Bruker, Billerica, MA, USA) with trans-2-[3-(4-tert-butylphenyl)-2-methyl-2-propenylidene]malononitrile (DCTB) as the matrix material.

For the formation of polymeric layers on silicon wafers, a SPIN150 spin coater (SPS-Europe, Putten, Netherlands) was used with the following process parameters: 3000 rpm, 1000 rpm sec^-1 and 30 sec spinning time. The thickness of polymer layers was measured with a SE400adv ellipsometer (Sentech Instruments GmbH, Berlin) and the static, advancing, and receding contact angles were measured by using OCA 20 contact angle measurement set-up (Data Physics GmbH, Filderstadt, Germany). For each sample, the average value from three different positions was taken for each of those measurements. Atomic force microscopy (AFM) was used to analyze the surface topology. A Dimension Icon AFM from Bruker (Billerica, MA, USA) was used. RFESP-75 (width: 40 μm, length: 235 μm, spring constant: 4.18 N m^-1, resonance frequency: 76 kHz) as well as ScanAsyst-Air (width: 25 μm, length: 115 μm, spring constant: 0.4 N m^-1, resonance frequency: 70 kHz) cantilevers were used. In order to perform Quantitative Nanomechanical (PeakForce-QNM) measurements, the deflection sensitivity, the spring constant and the tip radius needed to be calibrated prior to use. A sapphire standard sample from Bruker (PFQNM-SMPKIT-12M, SAPPHIRE-12M) was used to calibrate the deflection sensitivity, which resulted in 94.96 nm V^-1 for the ScanAsyst-Air cantilever and 125.92 nm V^-1 for the RFESP-75. The spring constant for the ScanAsyst-Air cantilever was calibrated using thermal tune and resulted in a value of 0.463 N m^-1. The spring constant of the RFESP-75 was calculated using Sader method^[47] and resulted in a value of 4.18 N m^-1. In order to determine the tip radius, an absolute method was used. A titanium carbide standard sample from Bruker (PFQNM-SMPKIT-12M, RS-12M) was used to perform a Tip Check. Then, the titanium carbide sample was replaced by one of the samples. The value of the PeakForce (PF) setpoint was then manually adjusted to obtain roughly a 2 nm surface deformation, the curve of Force (nN) vs. z position (nm) was captured and the indentation depth was measured using the Nanoscope Analysis 1.5 software. The value of the indentation depth was inserted into the Tip Check image of the titanium carbide standard sample to calculate the tip radius. This resulted in a RFESP-75 tip radius for the PPDL sample of 6.05 nm and ScanAsyst-Air tip end radius of 9.13 nm for the PPDL-triazole-P3HCA sample and 8.12 nm for the P3HCA sample.

2.3. Synthesis

2.3.1. Synthesis of poly(pentadecalactone) using benzyl alcohol as an initiator

Poly(pentadecalactone) (PPDL) was synthesized via ring-opening polymerization under inert gas. The amount of each reagent used, and the molecular mass of the resulting polymers, as determined by GPC, can be found in Table 1. In the glovebox, PDL (1.50 g, 6.24 mmol), BA (81.1 mg, 0.750 mmol) and an equimolar amount of TBD (100 mg, 0.750 mmol) were put together in a Schlenk tube. The polymerization was either performed in bulk or in solution after dissolving the reaction mixture in dry toluene (2 mL). The reaction mixture was removed from the glovebox and stirred for three days at 100 °C. For all reactions, ¹H NMR of the crude product was measured to determine the total monomer conversion. After the reaction had cooled down, the crude product was precipitated twice with an excess of ethanol. The resulting polymer was dried in vacuum at room temperature for two days.

Table 1. Ring-opening polymerization of PDL using TBD as a catalyst and different reaction conditions.^a .

Initiator	Solvent	n(cat.)b / mmol	n(PDL) / mmol	Mcalculated c/ g mol-1	Mn d / g mol-1	PDId
Benzyl alcohol	toluene	0.75	6.24	2,000	2,200	1.88
	toluene	0.10	4.16	10,000	10,500	1.73
	-	0.20	8.32	10,000	38,000	1.71
	toluene	0.05	6.24	30,000	32,000	1.80
	-	0.02	2.08	30,000	43,000	1.86
Propargyl alcohol	toluene	0.15	1.25	2,000	2,300	1.35
	toluene	0.05	2.08	10,000	6,100	2.45
	toluene	0.10	8.30	20,000	21,000	1.72
	toluene	0.07	8.30	30,000	30,000	1.62
	toluene	0.05	8.30	40,000	38,000	1.65

^aReactions were performed for three days at 110°C with full monomer conversion.

^bEquimolar amount was used for the initiator and the catalyst.

^cCalculated molecular mass according to the ratio of the used reactants.

^dThe Number-average molecular mass and polydispersity index were determined by GPC in chloroform using poly(methylmethacrylate) standards.

¹H NMR (250 MHz, CDCl₃, δ): 7.35-7.41 (m, 5H, C_aromat H), 5.12 (s, 2H, CH ₂-C_aromat), 4.06 (t, J = 6.6 Hz, 2H, CH₂-O-C=O), 3.65 (t, J = 6.6 Hz, 2H, CH₂-OH), 2.36 (t, J = 7.58 Hz, 2H, CH₂-C=O), 2.29 (t, J = 7.58 Hz, 2H, CH₂-C=O), 1.58-1.68 (m, 4H, CH ₂), 1.22-1.38 (m, 20H, CH ₂); ¹³C NMR (63 MHz, CDCl₃, δ): 173.99, 173.09, 136.80, 128.84, 127.79, 66.40, 64.39, 63.09, 34.40, 33.95, 29.63, 29.60, 29.53, 29.48, 29.28, 29.27, 29.17, 28.65, 25.93, 25.01.

2.3.2. Synthesis of poly(3-hydroxycinnamic acid)

Poly(3-hydroxycinnamic acid) (P3HCA) was synthesized via a two-step reaction. The amount of each reagent used, and the molecular mass of the resulting polymers determined by GPC can be found in Table 2. The monomer 3HCA (10.0 g, 60.9 mmol) was mixed in a Schlenk flask with anhydride acetic acid (20.0 mL, 21.6 g, 212 mmol) as a condensation reagent and sodium acetate (50.0 mg, 0.609 mmol) as a catalyst. The reaction mixture was put in a preheated oil bath and stirred for two hours under nitrogen at 150 °C. In order to remove the excess anhydride acetic acid, high vacuum was applied and the acid was collected in a cooling trap. The polymerization was then carried out at 210 °C for different reaction times in vacuum. For all reactions, ¹H NMR of the crude product was measured to determine the total monomer conversion. After cooling down, the crude product was dissolved in chloroform and precipitated twice into an excess of ethanol. The precipitated polymer was isolated and dried in vacuum for two days.

Table 2. Properties of formed films from P3HCA, PPDL, a blend of the two homopolymers and the block copolymer.

	Mn / g mol-1	static CA / °	advancing CA / °	receding CA / °	thicknessa /nm
PPDL	32,000	93 ± 2	96 ± 2	61 ± 3	97 ± 2
P3HCA	2,700	92 ± 2	98 ± 2	56 ± 4	73 ± 2
Blend		95 ± 2	98 ± 2	57 ± 3	n.d.
PPDL-triazole-P3HCA	11,700	96 ± 2	99 ± 2	58 ± 3	101 ± 2

^aThe thickness was determined using ellipsometry.

¹H NMR (250 MHz, CDCl₃, δ): 7.62-7.96 (m, 1H, C=CH), 7.10-7.55 (m, 4H, C_aromat H), 6.39-6.75 (m, 1H, C=CH-C=O), 2.34 (s, 3H, CH ₃); ¹³C NMR (63 MHz, CDCl₃, δ): 169.22, 164.86, 151.20, 145.96, 135.56, 130.04, 125.89, 124.02, 121.11, 117.97, 21.08.

2.3.3. Synthesis of block copolymers using poly(pentadecalactone) as macroinitiator

The monomer 3HCA (300 mg, 1.83 mmol) and the PPDL macroinitiator (M_n = 38 kg mol^-1, 440 mg, 1.83 mmol PDL units, 0.012 mmol PPDL chains) were mixed in a Schlenk flask with anhydride acetic acid (0.500 mL, 540 mg, 5.29 mmol) as a condensation reagent and sodium acetate (1.50 mg, 0.018 mmol) as a catalyst. The reaction mixture was put in a preheated oil bath and stirred for two hours under nitrogen at 150 °C. In order to remove the excess acetic acid, high vacuum was applied and the acid was collected in a cooling trap. The polymerization was then carried out at 210 °C for six hours in vacuum. After cooling down, the crude product was dissolved in chloroform and precipitated twice into an excess of ethanol. The produced polymer was isolated and dried in vacuum for 2 days.

2.3.4. Synthesis of poly(pentadecalactone) using propargyl alcohol as an initiator

The amount of each reagent used and the molecular mass of the resulting polymers determined by GPC can be found in Table 1. In the glovebox, PDL (300 mg, 1.25 mmol), PA (8.41 mg, 0.150 mmol) and an equimolar amount of TBD (20.0 mg, 0.150 mmol) were weighed into in a Schlenk tube and dissolved with dry toluene (0.4 mL). The reaction mixture was removed from the glovebox and stirred for three days at 100 °C. For all reactions, ¹H NMR of the crude product was measured to determine the total monomer conversion. After the reaction cooled down, the crude product was precipitated twice into an excess of ethanol. The resulting polymer was dried in vacuum at room temperature for two days.

¹H NMR (250 MHz, CDCl₃, δ): 4.68 (d, J = 2.5 Hz, 2H, CH ₂-C≡C), 4.06 (t, J = 6.6 Hz, 2H, CH₂-O-C=O), 3.65 (t, J = 6.6 Hz, 2H, CH₂-OH), 2.47 (t, J = 2.5 Hz, 1H, H-C≡C), 2.36 (t, J = 7.58 Hz, 2H, CH₂-C=O), 2.29 (t, J = 7.58 Hz, 2H, CH₂-C=O), 1.58-1.68 (m, 4H, CH ₂), 1.22-1.38 (m, 20H, CH ₂); ¹³C NMR (63 MHz, CDCl₃, δ): 174.00, 172.97, 74.66, 64.39, 63.07, 51.72, 34.41, 33.99, 29.63, 29.60, 29.53, 29.48, 29.28, 29.27, 29.17, 28.65, 25.93, 25.01.

2.3.5. Functionalization of poly(3-hydroxycinnamic acid)

3-Azidopropanol was synthesized according to literature procedures.^[48] P3HCA (M_n = 2.70 kg mol^-1, 3.00 g, 1.11 mmol P3HCA chains), 3-azidopropanol (562 mg, 5.55 mmol) and N,N-dimethylaminopyridine (DMAP, 163 mg, 1.33 mmol) were dissolved in dry dichloromethane (DCM, 16 mL). The solution was stirred for ten minutes. A cooling bath was put under the flask for another ten minutes, after which 1-ethyl-3-(3-dimethylaminopropyl)carbodiimid (EDC, 345 mg, 2.22 mmol) was added. The solution was stirred for two days at room temperature. Afterwards, water was added and the reaction solution was washed two times with 1M hydrochloric acid (HCl, 15 mL). The organic phase was washed twice with saturated NaHCO3 solution (15 mL) followed by water (15 mL). The organic phases were dried over Na2SO4, filtered and the solvent was evaporated under reduced pressure. After precipitation into an excess of ethanol, the functionalized polymer was isolated and dried in vacuo for 2 days.

¹H NMR (250 MHz, CDCl₃, δ): 7.62-7.95 (m, 1H, C=CH), 7.09-7.55 (m, 4H, C_aromat H), 6.39-6.72 (m, 1H, C=CH-C=O), 4.31 (t, J = 6.1 Hz, 2H, CH ₂-O), 3.45 (t, J = 6.7 Hz, 2H, CH ₂-N₃), 2.34 (s, 3H, CH ₃), 1.99 (q, J = 6.4 Hz, 2H, CH ₂-CH₂); ¹³C NMR (63 MHz, CDCl₃, δ): 169.26, 166.52, 164.89, 151.15, 145.96, 135.53, 129.99, 125.88, 123.96, 121.07, 117.92, 60.38, 48.16, 28.16, 21.03.

2.3.6. Synthesis of the block copolymer PPDL-triazole-P3HCA

In the glovebox, PPDL-alkyne (M_n = 6.10 kg mol^-1, 51.0 mg, 0.008 mmol PPDL chains), P3HCA-azide (M_n = 2.70 kg mol^-1, 392 mg, 0.145 mmol P3HCA chains), Cu^IOAc (1.02 mg, 0.008 mmol) and an equimolar amount of TBD (1.16 mg, 0.008 mmol) were put together in a Schlenk flask and everything was dissolved in dry chloroform (4 mL). The reaction mixture was removed from the glovebox and stirred for two days at 40 °C. Afterwards, the crude product was first precipitated into an excess of ethanol, and twice into an excess of N,N-dimethylformamide (DMF). The resulting polymer was dried in vacuum at room temperature for two days.

¹H NMR (250 MHz, CDCl₃, δ): 7.62-7.95 (m, 1H_P3HCA, C=CH), 7.09-7.55 (m, 4H_P3HCA, C_aromat H and 1H_triazole, H-C=C), 6.39-6.72 (m, 1H_P3HCA, C=CH-C=O), 5.17-5.24 (m, 2H_PPDL, CH ₂-C=C), 4.46-4.56 (m, 2H_P3HCA, CH ₂-N), 4.23-4.32 (m, 2H_P3HCA, CH₂-O), 4.06 (t, J = 6.6 Hz, 2H_PPDL, CH₂-O-C=O), 3.65 (t, J = 6.6 Hz, 2H_PPDL, CH₂-OH), 2.15-2.41 (m, 2H_PPDL, CH₂-C=O and 3H_P3HCA, CH ₃), 1.97-2.12 (m, 2H_P3HCA, CH ₂-CH₂), 1.58-1.68 (m, 4H_PPDL, CH ₂), 1.22-1.38 (m, 20H_PPDL, CH ₂); ¹³C NMR (63 MHz, CDCl₃, δ): 173.96, 172.93, 169.19, 166.52, 164.83, 151.13, 145.90, 142.52, 135.56, 130.03, 125.87, 124.68, 123.96, 121.09, 117.98, 64.36, 63.06, 61.01, 56.70, 48.62, 34.37, 34.09, 29.60, 29.57, 29.50, 29.44, 29.24, 29.14, 28.62, 28.17, 25.90, 24.99, 21.07.

2.4. Film formation from the polymeric material

Polymer films were formed via spin coating the respective polymer solution with a concentration of 20 mg mL^-1 on silicon wafers pre-functionalized with triethoxybenzophenone silane (3EBP-silane), which has been synthesized as described in the literature.^[49] The pre-functionalization procedure of the wafers were performed as described in the literature.^[50] The resulting polymer layers were put on a 120 °C hot plate for five minutes to delete the processing history of the sample. Afterwards, the samples were heated at 90 °C for 30 minutes and stored overnight at room temperature.

3. Results and Discussion

Study Design

With the aim of synthesizing an all-polyester block copolymer from an aliphatic and an aromatic component, our target structure was poly(pentadecalactone-block-3-hydroxycinnamate). Pentadecalactone was chosen as the aliphatic monomer because its number of carbon atoms is close to that of the aliphatic components of cutin and suberin, and because it can be readily polymerized to poly(pentadecalactone) (PPDL) via ROP.^{[23, 26]} Likewise, 3-hydroxycinnamic acid was chosen due to its structural similarity to the aromatic components of suberin, and because its polymers have a substantially better solubility than those of the 4-hydroxycinnamic acid isomer. We followed two synthetic strategies to obtain the target polymers: first, a sequential approach, where poly(pentadecalactone) is obtained via ROP, and 3-hydroxycinnamic acid is then synthesized in the presence of that first polymer; second, we synthesized the desired polymers separately with functional end groups, and then connected them to block copolymers via alkyne-azide cycloaddition. The sequential approach, using PPDL as a “macroinitiator” for a successive step-growth reaction seems mechanistically questionable, yet it was demonstrated by Li et al. that this approach gave blocky poly(ε-caprolactone-co-3,4-dihydroxycinnamic acid) copolymers,^[14] and we wanted to either verify or falsify this approach for the here presented system.

Homopolymer Synthesis

To obtain the first block of the target polymer, pentadecalactone (PDL) was polymerized via ring-opening polymerization using benzyl alcohol (BA) as initiator and triazabicyclodecene (TBD) as catalyst (Scheme 1a). The polymer molar mass obtained depended not only on the amount of initiator used, but also on the ratio of monomer and catalyst. The three components were weighed into a Schlenk tube under inert atmosphere, and the reaction mixture was stirred at 110 °C for three days. This gave full monomer conversion, as confirmed by proton NMR spectra of the crude product. The reaction was performed either in bulk or in toluene solution, as specified in Table 1. To remove the catalyst, the crude product was precipitated twice into an excess of ethanol. This gave the desired poly(pentadecalactone) (PPDL) homopolymer. Table 1 lists the amount of used reactants and the resulting number-average molecular masses Mn, which were determined by gel permeation chromatography (GPC). The theoretically calculated molecular masses matched the experimentally determined ones when the reaction was performed in solution. Polymerizations carried out in bulk gave higher molecular masses. This was also found by van der Meulen et al., and was explained by the potentially poor solubility of the catalyst in the reaction mixture, which would limit the number of active sites.^[51] Indeed, it was more difficult to keep the bulk polymerization mixtures homogeneous over the course of the reaction due to their higher viscosity. Besides PPDL homopolymers with relatively high molecular masses (30,000 – 40,000 g mol^-1), lower molecular mass PPDL was also synthesized for end-group analytics by proton NMR (Figure S1 in the Supporting Information). In these samples, the ¹H-NMR peaks from the benzyl end group were clearly visible at 7.36 ppm (aromatic ring) and 5.12 ppm (CH₂ group). The intensity of the latter signal matched the one of the CH₂ group next to the hydroxyl end group at 3.65 ppm, and thereby confirmed that the PPDL chains were indeed initiated by benzyl alcohol. All signals from the NMR measurements matched literature data.^[22] The successful synthesis could also be verified via Fourier transform infrared spectra (FTIR) and differential scanning calorimetry (DSC) measurements (Figure S2 and Figure S3, Supporting Information). The FTIR spectra showed the characteristic signals of an aliphatic polyester (e.g. aliphatic C-H vibrations at 2915 and 2848 cm^-1, ester groups at 1730 cm^-1). DSC measurements indicated that the crystalline polymer had a melting point around 89 °C.

The homopolymer of 3-hydroxycinnammic acid (3HCA), poly(3-hydroxycinnamate) (P3HCA), was synthesized as a reference material for the target block copolymers. 3HCA underwent polycondensation using acetic acid anhydride as the condensation reagent and sodium acetate as a catalyst (Scheme 1b). All reaction steps were performed with protection from light, since the polymer is light sensitive and could undergo cycloaddition reactions at its double bonds.^[52] In the first step, 3HCA was reacted with an excess of acetic anhydride at 150 °C for 2 hours. Unreacted anhydride was removed by vacuum at room temperature. In the second step, the polymerization mixture was heated to 210°C for different time spans in vacuum. Reaction details are given in Table S1 (Supporting Information). Proton NMR spectra of the crude product indicated full monomer conversion. The polymer was precipitated into ethanol and dried under high vacuum. GPC measurements indicated that the polymer obtained had a low molecular mass and the expected broad polydispersity (M_n ~ 3,000 g mol^-1, PDI 2-3, Table S1, Supporting Information) when the reaction times were five to six hours. Longer reaction times gave cross-linked, insoluble material.

Proton NMR spectra (Figure S4, Supporting Information) also confirmed that polymer had been obtained by a signal shift to higher ppm values and by peak broadening. This was further supported by FTIR measurements, which showed new signals at 1732 cm^-1 (ester groups) and 1767 cm^-1 (acetyl groups, Figure S5, Supporting Information). DSC measurements revealed the amorphous character of this polymer, with a glass transition temperature around 124 °C.

Since the nature of the end-groups of each homopolymer is important for successful block copolymer synthesis, a P3HCA sample with a M_n of 2700 g mol^-1 was analyzed by MALDI-TOF mass spectrometry (Figure 1). The most intense peaks were obtained for chains with about 11 repeat units, the longer chains had up to 36 repeat units. A closer look was taken at the region between m/z 1800 and 1950 to identify the different end groups (see Supporting Information, Table S2, for the potential end group structures). The signals at m/z 1812 and 1919 corresponded to P3HCA chains with one acetyl group and one acid group, which is the result that would be theoretically predicted from the polymerization mechanism. The other peak assignments are shown in Figure 1b. In the proton NMR spectrum, the different end groups could also be found: the singlet at 2.34 ppm corresponds to the protons of the acetyl end groups, and the shoulder at 2.33 ppm could correspond to the protons of the anhydride end group. Due to the overlap of those signals, a quantification of the amount each species by NMR was not possible. However, all the different P3HCA species had either acid or an anhydride end groups, and thus could be used for further functionalization reactions.

Sequential Approach

In the sequential approach to obtain the target block copolymer, PPDL with hydroxyl end groups and a molecular mass of 38,000 kg mol^-1 was used to copolymerize with 3HCA (Scheme 1c). For this, PPDL and 3HCA were heated using the same conditions as described for the P3HCA homopolymer. When analyzing the reaction product by proton NMR (Figure S7, Supporting Information), all signals characteristic for the PPDL and P3HCA repeat units could be seen, with no further signals. This indicates that 3HCA indeed polymerized to P3HCA, and that the PPDL block did not undergo transesterification. The FTIR spectrum also showed the characteristic signals for each homopolymer (Figure S8, Supporting Information). Yet neither NMR nor FTIR were suitable to differentiate between a potentially formed block copolymer and a mixture of two homopolymers. A proper distinction between the two cases was possible by GPC measurements (Figure S9, Supporting Information), which showed two well-separated signals. The one at higher molecular mass had the same elution times and shape as the PPDL homopolymer. When the two GPC peaks were collected separately and analyzed by proton NMR, it was confirmed that the high molecular mass peak was PPDL, while the low molecular mass peak was a P3HCA homopolymer. Thus, this approach gave homopolymers instead of the desired block copolymer. Interestingly though, even these relatively harsh conditions did not lead to transesterification between the two homopolymers, as characteristic NMR peaks for this process were missing. This result stands in contrast to the data reported by Li et al. for the poly(ε-caprolactone-co-3,4-dihydroxycinnamic acid) system.^[14] In that publication, it was reported that thermal polycondensation of 3,4-dihydroxycinnamic acid in the presence of a poly(ε-caprolactone) homopolymer (molecular mass of 51,000 g mol^-1, polydispersity index 1.32) gave complex multiblock copolymers with a molecular mass of 30,000 to 58,000 g mol^-1 and a polydispersity index of 1.78-1.96.^[14] GPC elugrams were not given in that paper. Since the polycondensation conditions were similar to the ones applied here, it seems as if the poly(ε-caprolactone) block had a higher propensity to undergo transesterification than the PPDL block used in this work

“Click” Approach

Since the sequential approach did not give the desired result, homopolymers with connectable end-groups were synthesized. To connect these functional polymers, the well-known copper-catalyzed 1,3-dipolar addition reaction (Huisgen addition) between an azide and an alkyne group was chosen. PPDL with an alkyne head group was synthesized via ring-opening polymerization by initiating PDL with propargyl alcohol (PA) (Scheme 1 d). The reaction was performed and worked up using the same conditions as for the BA-initiated homopolymer (110 °C for three days, precipitation into ethanol). Reaction details are given in Table 1. The experimentally determined molecular masses of the resulting PPDL-alkyne polymers matched the theoretically calculated ones (Table 1). As with the homopolymers bearing the benzyl end group, relatively high molecular masses could be obtained without broadening of the polydispersity. Proton NMR of PPDL-alkyne showed the characteristic signals of the homopolymer (Figure 2a), in addition to the end group signals at around 4.69 ppm (CH₂ group next to the alkyne) and 2.47 ppm (alkyne proton). The integral of this end group matched the one from the CH₂ group next to the opposite hydroxyl end group. Thus, all polymer chains have been alkyne-functionalized. FTIR spectra of PPDL-alkyne also showed the expected polymer signals (2915 cm^-1 and 2848 cm^-1 from to aliphatic C-H vibrations; C=O bands at 1730 cm^-1 from the ester group). Additionally, the band at 3307 cm^-1 was characteristic for the alkyne end group (Figure S10, Supporting Information). This is another confirmation of the successful alkyne functionalization.

The aromatic P3HCA homopolymer was to be functionalized with an azide group. As MALDI-TOF MS results (Figure 1) had indicated that all P3HCA homopolymer chains had an acid or anhydride end group, this homopolymer was to be reacted with 3- azidopropanol in the presence of the coupling agents 1-ethyl-3-(3-dimethylaminopropyl)carbodiimid (EDC) and N,N-dimethylaminopyridine (DMAP, Scheme 1e). For this, 3-azidopropanol was synthesized after literature procedures.^[48]

P3HCA, azidopropanol and DMAP were dissolved in dry dichloromethane and the solution was stirred for ten minutes. After cooling to 0 °C, EDC was added and the solution was stirred for two days at room temperature. After work up, the proton NMR spectra of P3HCA-azide (Figure 2b) featured the signals of the newly formed azide end groups (triplet at 4.31 ppm corresponds to the CH₂ group next to oxygen; triplet at 3.45 ppm to the CH₂ group next to nitrogen; quintet at 1.99 ppm corresponds to the CH₂ group in-between). The integral of this end group was 0.17 and could be used to determine the average length of the functionalized polymer, which was 12 repeat units. This roughly matches the MALDI-TOF mass spectrum results (Table S3, Supporting Information). In the section of the MALDI TOF mass spectrum of P3HCA-azide shown in Figure 2d, new signals from the azide species can be identified. Polymer chains with one hydroxyl end group and one and azide end group gave new signals at m/z 1814 and 1853. Acetyl-P3HCA with an additional azide end group were found at at m/z 1857 and 1894. Phenyl-P3HCA-azide had a signal at m/z 1943. Signals corresponding to the P3HCA without azide groups could still be detected, which clearly shows that not all of the polymer chains could be functionalized. This was also verified by proton NMR, where the singlet at 2.34 ppm of the acetyl end group still shows a small shoulder, indicating that there were unreacted anhydride groups. The amount of these species could not be precisely quantified due to the peak overlap and the relatively small peak intensity.

Due to the presence of a small fraction of chains without proper end groups for coupling, the reaction between P3HCA-azide and PPDL-alkyne was performed with an excess of P3HCA-azide. The advantage of this approach was that unreacted P3HCA species can be removed from the crude product by washing or precipitation with N,N-dimethylformamide (DMF), in which all P3HCA species are soluble, while the block copolymer and PPDL are not. To connect the two homopolymers, PPDL-alkyne (M_n = 6,100 g mol^-1), P3HCA-azide (2,700 g mol^-1), copper(I) acetate (Cu^IOAc) and an equimolar amount of TBD were reacted dry chloroform at 40 °C for 2 days (Scheme 1f). Precipitation of the crude polymer into an excess of ethanol removed the catalyst as well as the base, and precipitation into an excess of DMF removed excess P3HCA species. Figure 2c shows the NMR spectrum of the block copolymer obtained. All the characteristic signals of protons from the homopolymer repeat units could be assigned. The signal at 3.45 ppm from the CH₂ group next to the nitrogen atom from P3HCA-azide as well as the signals at 4.68 ppm and 2.47 ppm from protons of the alkyne group of PPDL-alkyne vanished. A new signal at 5.22 ppm appeared, which could be assigned to the CH₂ group of PPDL next to the newly formed triazole group. The signal at 4.51 ppm belongs to the CH₂ group next to the nitrogen of the triazole bond. The signal from the CH group of the triazole bond could be found in the aromatic region and overlaps with the signals of the P3HCA repeat units. The integrals from the signals of and next to the new formed triazole group match the integrals from the other chain end, e.g. the one at 3.65 ppm belonging to CH₂-OH of the PPDL chain. This is solid evidence that the desired block copolymer PPDL-triazole-P3HCA was indeed obtained. Furthermore, no signals indicating transesterification could be found. From the peak intensities of the signals from each block, a calculation of the ratio of each species was performed, which revealed that the block copolymer contained 28% P3HCA. Those aromatic chains had approximately 7 repeating units, which was calculated from the ratio of integrals of the signals of one proton from the double bond at around 6.5 ppm, and the signals of the CH₂ group from the former azide end group. The signals from the acetyl end group of the P3HCA chains overlapped with the signals of the C(=O)CH₂ group of the PPDL chain, so that further correlations with the number of acetyl end groups was not possible.

The shift of the GPC curve of the block copolymer PPDL-triazole-P3HCA relative to both homopolymers also confirms that a block copolymer was obtained (Figure 2e). The experimentally determined M_n of the block copolymer was 11,700 g mol^-1, with a polydispersity index of 1.79. No shoulders or further peaks were visible, indicating that no unreacted homopolymer was present. Additionally, the oligomer signals at an elution volume of 27 to 32 mL previously found in the homopolymer elugrams were not present in the block copolymer elugram, indicating that these low molecular mass entities had combined to higher molecular mass polymers. Figure S10 in the Supporting Information shows the FTIR spectra of the block copolymer PPDL-triazole-P3HCA in comparison to the corresponding homopolymers P3HCA-azide and PPDL-alkyne. Bands from characteristic groups of each homopolymer can be found in the copolymer, indicating that both species must be present. The characteristic signal of the azide group at 2096 cm^-1 as well as the signal from the alkyne group at 3307 cm^-1 vanished. The new formed triazole signals (1636 cm^-1 and 1463 cm^-1, belonging to C=C and C-H vibrations) overlap with signals from the homopolymers and therefore cannot be clearly identified.

DSC measurements of PPDL-triazole-P3HCA showed a semi-crystalline behavior. A melting temperature of around 90 °C as well as a recrystallization temperature of 78 °C was found (Figure S11, Supporting Information), which originates from the PPDL block. A low intensity glass transition temperature at around 123 °C could also be found, which was assigned to the P3HCA block. The low intensity of that transition can be explained by the block ratio of P3HCA and PPDL. Furthermore, the presence of only one melting point and one glass transition temperature similar to the ones of the homopolymers is in line with the expectations for a block copolymer (in contrast to a statistical copolymer).

Firm Formation Studies

In order to investigate the film formation and phase separation behavior of the synthesized block copolymer, films were formed by spin coating. The respective homopolymers and a 1:1 blend of both homopolymers were investigated as reference materials. In each case, polymer solutions at a concentration of 20 mg mL^-1 were spin-cast onto pre-functionalized silicon wafers. The pre-functionalization with the hydrophobic silane 3EBP made the wafer surfaces more hydrophobic und thus a better match to the hydrophobicity of the polymers. The resulting polymer layers were heated on a hot plate at 120 °C hot plate for five minutes to delete their processing history. Afterwards, the samples were heated at 90 °C for 30 minutes and stored overnight at room temperature.

Table 2 summarizes the surface properties of the films obtained. Films made from PPDL and the block copolymer had a thickness of around 100 nm, as determined by ellipsometry. P3HCA had a lower film thickness of 73 ± 2 nm, which could be explained by the lower molecular mass of the polymer. Contact angle measurements revealed that all the formed films were hydrophobic, with a static contact angle of around 94 °. The advancing contact angle was around 97 °, and the receding contact angle was around 60 °. Thus, the hydrophobicity of the components of PPDL-triazole-P3HCA is similar.

Atomic force microscopy was used to examine the surface morphology of the films (Figure 3). PPDL (M_n = 32,000 g mol^-1) had the expected crystalline behavior, with large spherulites of a typical diameter of > 250 μm, consisting of densely packed lamellae. The PPDL homopolymers with smaller molecular masses also formed spherulites. The surface roughness of PPDL was 4.5 - 8.5 nm, depending on the size of the sample area evaluated (Figure 3a). In contrast to this, the P3HCA homopolymer was amorphous, with a roughness of only 0.3 nm (Figure 3c). The 1:1 blend of PPDL and P3HCA showed the expected microphase separation, with spherical, unstructured P3HCA domains with a height of 10-150 nm and a width of 0.1-1.8 μm embedded into a structured PPDL matrix and an overall roughness of 32.2 nm. The PPDL-triazole-P3HCA block copolymer (with 21 % P3HCA) also formed spherulites. In contrast to the PPDL homopolymer, the size of the block copolymer spherulites was smaller, with diameters of between 28 to 56 μm, and lower film roughness of only 1.8 – 3.0 nm (Figure 3d). The lamellar crystals that form the spherulites also seemed finer than in pure PPDL homopolymer. Furthermore, small vacancies are observed between the lamellae. These have a depth of up to 7 nm and a width of around 200 nm. The differences of the morphology of PPDL-triazole-P3HCA compared to both homopolymers and the blend clearly demonstrates that a chemically different polymer with a distinct micro- and nanostructure has been synthesized.

Figure 3.

AFM height images and roughness data (R_q, root-mean-square roughness) of the films made from a) PPDL homopolymer (M_n = 32,000 g mol^-1), b) a 1:1 blend of the PPDL and P3HCA homopolymers, c) P3HCA homopolymer (M_n = 2,700 g mol^-1), and d) the block copolymer PPDL-triazole-P3HCA.

To further analyze the different polymer films, quantitative nanomechanical AFM studies (QNM-AMF) were performed. The QNM-AFM images can be found in the Supporting Information in Figures S13-S15. From these, the elastic modulus of the samples was calculated using the Derjaguin-Muller-Toporov (DMT) model. The crystalline PPDL had a DMT modulus of 801 ± 42 MPa, in contrast to the amorphous P3HCA with only 1.77 ± 0.63 MPa. The block copolymer had an intermediate DMT modulus of 22.3 ± 2.7 MPa. This again proves that the block copolymer had physically distinct properties compared to the two homopolymers.

4. Conclusion

In the above, we presented two strategies to synthesize all-polyester diblock copolymers from aliphatic pentadecalactone and aromatic 3-hydroxycinnamic acid. Poly(pentadecalactone) (PPDL) homopolymers were obtained by ring-opening polymerization. Functionalized PPDL with alkyne groups (PPDL-alkyne) was readily obtained when propargyl alcohol was used as an initiator. Poly(3-hydroxycinnamate) (P3HCA) could be obtained by polycondensation, and functionalization with 3-azidopropanol gave the corresponding polymer with azide end groups (P3HCA-azide). Since the sequential approach to synthesize the block copolymer (quite expectedly) failed, the two end-functionalized homopolymers were connected via 1,3-dipolar Huisgen reaction to form the block copolymer PPDL-triazole-P3HCA. The block-like structure of this polymer was confirmed by NMR and FTIR spectroscopy, GPC, and atomic force microscopy. Notably, polymer films formed from PPDL-triazole-P3HCA had a distinct morphology and a modulus that was in between the ones of the two homopolymers. The polymer also retained the thermal characteristics (one melting point, one glass transition temperature) of the parent homopolymers.

While styrenic or acrylic monomers can be easily polymerized sequentially to di-, tri- and multiblock copolymers consisting of all-carbon backbones, the synthetic possibilities for polymer architectures containing ester blocks are limited. The high reaction temperatures needed in polycondensation reactions, and even in sequential ring-opening polymerizations, can lead to transesterification between the blocks. Alternatively, the components can prove incompatible when it is attempted to synthesize a second block in the presence of the first block. This was observed in the here presented case. In consequence, the synthesis of block copolymer PPDL-triazole-P3HCA was quite involved and required separate synthesis of each block, followed by functionalization with reactive end groups and covalent connection of the two blocks in the final reaction step. In spite of complications (such as only partial end-functionalization of the P3HCA homopolymer), it was possible to separate the block copolymer obtained from homopolymer contaminations by careful choice of the work-up conditions. Thus, the structural properties of the resulting PPDL-triazole-P3HCA block copolymer could be studied in detail. The PPDL-triazole-P3HCA polymer obtained did not show a lamellar phase separation because its components did not have a 1:1 volume ratio. The volume ratio of the P3HCA block was limited because the polycondensation of 3HCA only gave low molecular masses. However, the morphological studies clearly showed that the nanostructure and the modulus of PPDL-triazole-P3HCA was altered by even a low P3HCA content. Thus, future work in this direction will be dedicated to finding a suitable aromatic component to obtain polymer blocks with higher molecular masses, so that all-polyester block copolymers with a more defined nanostructure can be obtained. Such degradable materials could be sustainable alternatives to the currently used all-carbon block copolymers, for example as phase compatibilizers or component of packaging materials.

Fig. 2.

(A) Electron micrograph and cartoon structure of a suberinized cork cell; (B) Predicted molecular structure of suberin lamellae.^{[5, 6]} Reproduced with permission.