Integrating Mussel Chemistry into a Bio-Based Polymer to Create

Jan 4, 2017 - (1, 3) With the advent of synthetic adhesives, renewable monomers and .... Degradation samples were solvent cast into these molds yieldi...
0 downloads 0 Views 2MB Size
Article pubs.acs.org/Macromolecules

Integrating Mussel Chemistry into a Bio-Based Polymer to Create Degradable Adhesives Courtney L. Jenkins,† Heather M. Siebert,† and Jonathan J. Wilker*,†,‡ †

Department of Chemistry, Purdue University, 560 Oval Drive, West Lafayette, Indiana 47907-2084, United States School of Materials Engineering, Purdue University, 701 West Stadium Avenue, West Lafayette, Indiana 47907-2045, United States



S Supporting Information *

ABSTRACT: Adhesives releasing carcinogenic formaldehyde are almost everywhere in our homes and offices. Most of these glues are permanent, preventing disassembly and recycling of the components. New materials are thus needed to bond and debond without releasing reactive pollutants. In order to develop the next generation of advanced adhesives we have turned to biology for inspiration. The bonding chemistry of mussel proteins was combined with preformed poly(lactic acid), a bio-based polymer, by utilizing side reactions of Sn(oct)2, to create catechol-containing copolymers. Structure−function studies revealed that bulk adhesion was comparable to that of several petroleum-based commercial glues. Bonds could then be degraded in a controlled fashion, separating substrates gradually using mild hydrolysis conditions. These results show that biomimetic design principles can bring about the next generation of adhesive materials. Such new copolymers may help replace permanent materials with renewable and degradable adhesives that do not create chronic exposure to toxins.



INTRODUCTION We are constantly surrounded by toxic adhesives. The plywood in our walls, the chairs we sit on, and the carpet beneath our feet are all off-gassing reactive chemicals.1 With ∼9 billion kilograms of glue manufactured annually in the United States, almost 4 billion kilograms contain formaldehyde.2 The permanent, petroleum-based nature of most adhesives also means that there are no practical ways to disassemble building materials, furniture, cars, or electronics for recycling. Renewable, nontoxic, and removable adhesives are thus in great demand to decrease our exposure to pollutants as well as waste in landfills. Early adhesives were made of natural materials such as starch and protein, permitting natural degradation.1,3 With the advent of synthetic adhesives, renewable monomers and degradability were lost in exchange for higher performance.4 The detrimental health and environmental effects of synthetic glues are becoming more of a concern, with alternatives being developed. Modifications to natural materials can improve adhesive properties.5−12 Some systems have made major gains in adhesion,6,7 whereas others find advances to be more modest.8−12 With biomedical adhesion being a popular target, successes include overcoming the challenges of degradation and biocompatibility. Difficulties inherent to binding biological substrates means that current strengths can still leave a desire for further improvements.5,10 Other systems can use nonrenewable components to obtain stronger bonding but require somewhat extreme conditions to bring about degradation13 or are degradable, but not quite to a complete extent.14,15 When considering how to design adhesives for our future, marine biology presents several compelling ideas. Mussels achieve strong bonding onto rocks with proteins that contain © XXXX American Chemical Society

3,4-dihydroxyphenylalanine (DOPA) for cross-linking and adhesion.16 Utilizing the catechol functionality of DOPA is inspiring the emergence of exciting new materials. This musselmimetic chemistry has sparked the development of copolymer,17−25 polypeptide,26−29 and coacervate30 systems for hydrogels, coatings, and sensors.16,31−33 Integrating biomimetic chemistry into otherwise nonadhesive polymers such as polystyrene has even generated systems able to outperform established commercial products including Super Glue.18 As the mussel-mimetic field has grown, several catecholcontaining polymers have been targeted toward biomedical applications. Keeping biocompatibility in mind, many such systems focus on poly(ethylene glycol) (PEG) with the addition of components including acrylics,34,35 polycaprolactone,18,35 and even some natural materials.36−38 Others utilize potential degradability in the main backbone including chitosan19,39 and keratin,40 although, at least so far, these systems appear to function more like sealants instead of high strength adhesives. Natural components such as caffeic acid, coumeric acid, and cinnamic acid can be combined for significant bonding, albeit without degradability.25 The general idea of using natural materials to achieve degradability appears to be successful27,38,39 and can even take advantage of enzymes for material breakdown.36,41 Nonrenewable materials can also be taken apart.42−44 Lower strengths can be quite useful for some biomedical needs, but additional areas will benefit from the development of stronger, renewable, and degradable adhesives. Received: October 10, 2016 Revised: December 19, 2016

A

DOI: 10.1021/acs.macromol.6b02213 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules

Polymers were characterized by 1H NMR spectroscopy in acetoned6 on a Varian Inova-300 MHz spectrometer. Removal of the methylene protecting groups could be observed by 1H NMR spectrometry, comparing before versus after spectra. However, overlapping peaks meant that such changes were difficult to distinguish quantitatively. Also indicative of deprotection were color changes noted when the polymer was dissolved in acetone. Addition of Fe(NO3)3 could change the acetone solution from light orange-brown to black, revealing a reaction of the iron with a deprotected catechol. Titrations of this sort between the deprotected polymer and Fe3+ were attempted using UV−vis spectroscopy. However, the insoluble nature of the cross-linked material resulting from the polymer reacting with iron made quantification of the degree of free catechol difficult. Molecular weights were determined by GPC using a Polymer Laboratories PL-GPC20 with tetrahydrofuran mobile phase and polystyrene standards to calibrate the instrument. Differential scanning calorimetry (DSC) was used to find the polymer glass transition temperatures (Tg’s) and determine random versus block incorporation of 3,4-dihydroxymandelic acid into the PLA backbone. Adhesion Studies. Adhesion testing was conducted on 6061-T6 aluminum substrates purchased from Farmer’s Copper, cut into 8.9 × 1.2 × 0.3 cm rectangles and cleaned using the ASTM D2651−01 standard method.49 The cleaning procedure involved degreasing the metal substrates initially using trichloroethylene followed by washes in base and acid. Samples were rinsed with methanol and boiling water to remove residual iron contained in the acid bath. Polymer samples were dissolved in acetone (0.3 g mL−1). An overlap area of 1.2 × 1.2 cm between substrates in lap shear configuration was covered with the polymer solutions (45 μL). For cross-linked samples, tetrabutylammonium periodate dissolved in acetone (15 μL) was then added at a 3:1 catechol:[N(C4H9)4](IO4) ratio. Aluminum substrates were overlapped and allowed to set for 30 min at room temperature before curing at 37 °C for 22 h. Samples were then cooled to room temperature for 30 min. Lap shear testing was conducted at 2 mm min−1 crosshead speed following a modified ASTM D1002 standard method using an Instron 5544 Materials Testing System.50 The maximum force at point of failure was measured and divided by the substrate overlap area to determine adhesion strengths. Error bars indicate 95% confidence intervals determined by averaging 10 samples in the case of all dry adhesion testing. Adhesion was also examined on sanded steel and poly(tetrafluoroethylene) (Teflon), purchased and prepared by a method described previously.18 Commercial adhesives were tested on each substrate for reference points including Elmer’s Glue, Gorilla Glue, Titebond Hide Glue, and Lineco Wheat Starch Glue. To keep conditions consistent, the same quantity and curing were implemented for each material. Hydrolytic Degradation. The months-long timeline required for these studies meant that polymers were being hydrolyzed at the same time that we were determining the composition and molecular weights giving rise to the best adhesion. Consequently, the degradation studies were carried out with catechol30%-PLA70% of Mw ≈ 23 000 g mol−1, which was providing the highest adhesion strength when these studies began. Cylindrical molds were made out of polydimethylsiloxane. Degradation samples were solvent cast into these molds yielding ∼250 mg cylindrical polymer monoliths of ∼7 mm height and ∼6 mm diameter. The copolymer alone samples were dark brown and became almost black when cross-linked with [N(C4H9)4](IO4). Each sample was submerged in phosphate buffered saline (PBS) with a pH of 7.4 and stored at 37 °C. Degradation was assessed by mass loss of the polymer over time. Samples were removed from PBS and dried under reduced pressure for 3 h before obtaining the dry mass. Measurements were taken every 24 h for the first 7 days and weekly thereafter. Data were repeated in triplicate. The average dry mass was plotted as a function of time. A second set of cylindrical samples was cast with catechol29%-PLA71% (Mw ≈ 19,000 g mol−1). Samples were treated as described above, being stored in PBS with a pH of 7.4 at 37 °C. Degradation of polymer chains was examined by monitoring changes in molecular weight of the

Renewable, degradable polymers such as poly(lactic acid) (PLA) have received significant attention in the search for environmentally benign alternatives to petroleum-based plastics.2,45 Because of low toxicity, incorporation of lactic acid into copolymers has grown steadily with interest in biomedical materials.42,45 Such copolymers have been synthesized via several methods including grafting, ring-opening polymerization, and polycondensation. Often, polycondensation reactions are plagued by low molecular weights.42,45 Combining the positive attributes of PLA and mussel adhesive chemistry may generate a new class of materials (Figure 1). Our experience, however, found that access to such

Figure 1. Adhesive copolymers designed by combining the renewability of poly(lactic acid), which can be sourced from corn, with the adhesive chemistry of marine mussels into an adhesive copolymer system.

copolymers was often thwarted by several synthetic troubles. These challenges were met by taking advantage of the polymerization mechanism exhibited by the tin(II) 2-ethylhexanoate, Sn(oct)2 catalyst. The scission and chain rearrangement mechanism of Sn(oct)2 was used to create adhesive copolymers from preformed PLA.46 Avoiding the use of formaldehyde for curing, as well as utilizing components previously shown to be nontoxic, limit the likelihood of health issues associated with this new adhesive system.45,47 Results presented here show that a promising new adhesive system can be derived from a renewable resource, display high strength bonding, and degrade in a controlled fashion.



EXPERIMENTAL SECTION

Synthesis and Characterization. To synthesize the copolymers of interest, poly(DL-lactic acid) (PLA) was obtained from Akina Incorporated. Monomeric 3,4-(methylenedioxy)mandelic acid was purchased from Santa Cruz Biotechnology. The tetrabutylammonium periodate, [N(C4H9)4](IO4), cross-linker was synthesized using an established protocol.48 UV−vis absorption spectroscopy, 1H NMR spectroscopy, and melting point determinations confirmed the product. Additional materials were obtained from Sigma-Aldrich and Fisher Scientific. Oligomerization of 3,4-(methylenedioxy)mandelic acid was conducted by heating the monomer at 150 °C under vacuum until it melted into a brown, viscous liquid. High temperatures were maintained for 3 h at reduced pressure to aid removal of water. This condensation reaction yielded oligomers shown by 1H NMR spectroscopy to have a degree of polymerization typically between 4 to 5 (Mn ≈ 500−700) determined by gel permeation chromatography (GPC). Oligo(3,4-methylenedioxymandelic acid) and PLA were melted together at 150 °C before adding the Sn(oct)2 catalyst (0.5− 2.5 wt %) and p-toluenesulfonic acid (TSA) cocatalyst (1:1 molar Sn(oct)2:TSA). This melt polycondensation took place for 2 h under reduced pressure via vacuum before adding additional TSA (10 wt %). The reaction proceeded for an additional 1 h under argon to complete the polymerization and remove the methylene protecting group to reveal the catechol-functionalized poly[(3,4-dihydroxymandelic acid)co-(lactic acid)] (“catechol−PLA”). The PLA homopolymers containing no catechol were made by performing a ring-opening polymerization with lactide catalyzed by Sn(oct)2 (2 wt %) with equimolar TSA cocatalyst. B

DOI: 10.1021/acs.macromol.6b02213 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules cylindrical monoliths as well as species in the surrounding buffer solution by GPC. Measurements were taken at daily for 5 days. Polymer degradation was also assessed by the loss of bond strength over time. Adhesion testing on aluminum substrates was set up and cured in a manner similar to that described above. Bonded samples were then submerged in pH 7.4 PBS at 37 °C. The water baths were changed if the pH dropped below 7.0. Bond areas were rinsed with deionized water before testing to remove any salt along the bond line. Trials were tested after 24 h, daily for 7 days, and then weekly for 10 weeks or until no material remained bound at a given time point. Each sample was measured at least five times and averaged.

The methylene protecting group prevented catechol crosslinking, which occurs at high temperatures and generates charred, insoluble materials. Catalytic quantities of ptoluenesulfonic acid were insufficient to remove the methylene protecting group, indicated by a lack of color change upon the second addition of acid. Additional acid caused some degradation to occur, but the presence of Sn(oct)2, along with short reaction times, limited the effect. This methylene group removal with acid yielded the final copolymer poly[(3,4dihydroxymandelic acid)-co-(lactic acid)] (“catechol−PLA”). Copolymers were characterized using 1H NMR spectroscopy to provide the final compositions of catechol:PLA (Figure 2B), gel permeation chromatography (GPC) for molecular weights, and differential scanning calorimetry (DSC) to observe thermal transitions showing random (not block) copolymers (Tables S1 and S2). Low Tg values were observed for samples consisting of low molecular weights. The presence of oligomers such as oligo(LA) can suppress the observed Tg of polymers. A subset of the polymers made here contained minimal levels of persisting oligomer, in particular the copolymer of Mw = 39 500 g/mol, hence the Tg value seemingly out of the expected trend (Table S2). The final polymer compositions displayed relatively good agreement with the starting feed monomer ratios. Minor losses of catechol content appeared to be due to cross-linking even under an inert atmosphere, but could be limited by adjusting timing of the second p-toluenesulfonic acid addition (Tables S1 and S2). Omitting additional p-toluenesulfonic acid as well as altering the reaction time after addition (30 min, 1 h, and 2 h) determined that 1 h provided the shortest exposure to acid while also deprotecting all polymers consistently. This synthetic scheme proved to be quite flexible, permitting synthesis of varied molecular weights and ratios of catechol:lactic acid within the copolymers. Adhesion Studies. For determining the adhesive performance of catechol−PLA, copolymers were dissolved into acetone, a common solvent for PLA.51 Joints were formed by depositing the copolymer solution between aluminum substrates overlapped in a typical lap shear configuration.18 Copolymer adhesion was tested both alone and when cross-linked with tetrabutylammonium periodate, [N(C4H9)4](IO4). This oxidant has been used with other biomimetic systems18,52−54 and was employed here at a 3:1 catechol:(IO4)− ratio to approximate the cross-linking found in mussel adhesive plaques.55 Prior work demonstrated different and often improved bonding strengths when catechol-containing polymers were cross-linked.18,52−54 Both oxidative and metallic reagents have been used to bring about cross-linking via chelation, oxidation, and radical generation.27,56 Cross-linking of DOPA-containing proteins and catechol-containing polymers is generally believed to proceed via Michael addition chemistry.27,56 Cross-linkers begin this process by oxidizing the catechol group by one electron, to a semiquinone, or two electrons, to a quinone. Nucleophiles present in macromolecule such as phenolic −OH’s then react with the oxidized species to yield covalent cross-links. Predicting the ideal copolymer to yield the strongest bonding is not straightforward. Mussel adhesive plaques are made of six different proteins with a range of molecular weights (6000− 110000 g mol−1) and DOPA contents (2−30% of all amino acids).57 Prior work with catechol-containing copolymers has shown that modulating the molecular weight and catechol content both impact adhesion significantly.52,53Consequently, here we examined how altering the ratio of lactic acid to



RESULTS AND DISCUSSION Synthesis and Characterization. There are several potential ways to integrate catechol groups into PLA. Experiments in the laboratory, nonetheless, proved otherwise. Seemingly logical synthetic routes to catechol-containing PLA copolymers were explored. Ring-opening copolymerization with cyclic lactide and a DOPA mimic, combining lactic acid with DOPA via an acid chloride, and various combinations of monomers, dimers, or oligomers were all attempted under multiple reaction conditions. Different protecting groups as well as not using any protecting groups were also tried. Problems arose with each of these routes including unwanted cross-linking reactions, insufficient molecular weights, and low yields. A successful route to catechol-functionalized PLA was developed and is shown in Scheme 1. The 3,4-methylenedioxScheme 1. 3,4-Methylenedioxymandelic Acid Monomer Oligomerized and Subsequently Combined with Poly(lactic acid) in a Polycondensation Reactiona

a

Acid aids in the polymerization reaction while simultaneously deprotecting the pendant catechol to reveal poly[(3,4-dihydroxymandelic acid)-co-(lactic acid)] (“catechol−PLA”).

ymandelic acid monomer could be oligomerized, demonstrated by 1H NMR spectroscopy in Figure 2A obtaining a typical degree of polymerization of 4 to 5 (Mn ≈ 500−700). The oligomer was then melted into preformed poly(DL-lactic acid) via a polycondensation reaction. Racemic PLA was chosen due to the amorphous properties that may yield a better adhesive compared to the rigid and crystalline nature of stereoregular poly(L-lactic acid). Combining racemic PLA with oligomerized 3,4-methylenedioxymandelic acid and a Sn(oct)2 catalyst brought about both intermolecular and intramolecular transesterification.46 Chain rearrangements, scission, and reincorporation formed a new random copolymer.46 Although the reaction mechanism of Sn(oct)2 does not lend itself to complete control over polymer molecular weights, often creating relatively broad polydispersities, the final molecular weight can, indeed, be influenced.46 Our experience here shows that altering the starting molecular weight of preformed PLA as well as the ratio of catalyst to PLA or PLA plus oligomer results in final products of varied chain lengths. C

DOI: 10.1021/acs.macromol.6b02213 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules

Figure 2. 1H NMR spectra in acetone-d6 of (A) oligo(3,4-dimethoxymandelic acid) and (B) poly[(3,4-dihydroxymandelic acid)-co-(lactic acid)].

Figure 3. Structure−function studies to determine optimal adhesive bonding for catechol−PLA. (A) The effect of copolymer composition on lap shear adhesion strength of poly[(3,4-dihydroxymandelic acid)0−49%-co-(lactic acid)51−100%] was examined. The percent of catechol-containing monomer, 3,4-dihydroxymandelic acid, was varied with lactic acid comprising the remainder of the copolymer. Molecular weights of each polymer were in the range of ∼12 000 to ∼34 000 g mol−1. (B) Influence of polymer molecular weight upon adhesion. These copolymers were composed of ∼7% 3,4-dihydroxymandelic acid and ∼93% lactic acid.

maintained at ∼12 000 to ∼34 000 g mol−1. Data in Figure 3A reveal that the strongest adhesion was derived from a crosslinked copolymer composed of ∼7 mol percent of 3,4dihydroxymandelic acid and ∼93% lactic acid. This unexpected result contrasts with prior work on a polystyrene-based system in which ∼33% of the catechol-containing monomer yielded the highest performance.53 Needing so little catechol for the highest adhesion may be a result of enhanced interchain

catechol-containing monomer and also the copolymer molecular weight would influence performance. Catechol content could be controlled by altering the 3,4methylenedioxymandelic acid to PLA ratio in the starting polymerization feed. A range of 0−49 mol percent 3,4dihydroxymandelic acid was incorporated into the PLA host polymer, with the remainder monomers being lactic acid (Table S1). For these studies similar molecular weights were D

DOI: 10.1021/acs.macromol.6b02213 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules Table 1. Adhesion of Catechol−PLA Compared to Commercial Glues adhesive

adhesion [MPa] in aira

adhesion [MPa] under water for 24 hb

± ± ± ± ± ±

0 2.5 ± 0.8 0 0 0.10 ± 0.05 1.0 ± 0.3

Elmer’s Glue (poly(vinyl acetate)) Gorilla Glue (polyurethane) starch glue hide glue poly(lactic acid) biomimetic copolymer (catechol6%−PLA94% + IO4−)

3 2.8 2.4 0.8 0.21 2.6

1 0.7 0.4 0.1 0.06 0.4

Bonds were lap shear joints between two pieces of aluminum, cured for 24 h at 37 °C, and tested immediately. bOr they were then submerged into buffer for 24 h and measured.

a

Table 2. Adhesion Strength of Catechol−PLA Bonding Different Substrates in Air aluminum [MPa] Elmer’s Glue (poly(vinyl acetate)) Gorilla Glue (polyurethane) starch glue hide glue poly(lactic acid) biomimetic copolymer [catechol6%−PLA94% + (IO4)−]

3 2.8 2.4 0.8 0.21 2.6

± ± ± ± ± ±

1 0.7 0.4 0.1 0.06 0.4

sanded steel [MPa] 1.6 1.7 2.0 1.2 0.21 1.7

± ± ± ± ± ±

0.6 0.3 0.3 0.4 0.05 0.5

teflon [MPa] 0.23