Highlighting the Importance of Surface Grafting in Combination with a Layer-by-Layer Approach for Fabricating Advanced 3D Poly(l-lactide) Microsphere Scaffolds.

A combined surface treatment (i.e., surface grafting and a layer-by-layer (LbL) approach) is presented to create advanced biomaterials, i.e., 3D poly(l-lactide) (PLLA) microsphere scaffolds, at room temperature. The grafted surface plays a crucial role in assembling polyelectrolyte multilayers (PEMs) onto the surface of the microspheres, thus improving the physicochemical properties of the 3D microsphere scaffolds. The grafted surface of the PLLA microspheres demonstrates much better PEM adsorption, improved surface coverage at low pH, and smoother surfaces at high pH compared with those of nongrafted surfaces of PLLA microspheres during the assembly of PEMs. They induce more swelling than nongrafted surfaces after the assembly of the PEMs and exhibit blue emission after functionalization of the microsphere surface with a fluorescent dye molecule. The 3D scaffolds functionalized with and without nanosheets not only exhibit good mechanical performance similar to the compressive modulus of cancellous bone but also exhibit the porosity required for cancellous bone regeneration. The magnetic nanoparticle-functionalized 3D scaffolds result in an electrical conductivity in the high range of semiconducting materials (i.e., 1-250 S cm-1). Thus, these 3D microsphere scaffolds fabricated by surface grafting and the LbL approach are promising candidates for bone tissue engineering.

Introduction

Polymer particles have attracted much attention in many fields (e.g., the chemical industry, biotechnology, pharmacology, and the medical field). Their physicochemical properties, such as their electrical, optical, and chemical properties, can be adjusted via alterations in their composition, size, and structure.[1−4] The surface chemistry and morphology of the particles play a critical role in their assembly, and the morphology influences their physical interactions when they are utilized as new types of material building blocks for self-assembly.[5,6] Alterations in surface morphology can be achieved via surface modification to produce the desired properties, thereby controlling the subsequent surface interactions and responses needed for any particular application.[7] Surface grafting via UV light, which is commonly called photografting, is one of the surface modification techniques that can be used to modify polymer surfaces.[8−10] It is a mild reaction that selectively influences the sample surface,[9] and it is nondestructive.[11] The modified structure of particle surfaces offers a new opportunity for tailoring biomaterials to improve the cellular functions and interactions for tissue engineering for applications such as bone tissue regeneration.[5,6,12]

The three-dimensional (3D) scaffolds derived from polymeric materials for human cancellous or trabecular bone require suitable mechanical properties with a compressive modulus in the range of 10–2000 MPa,[13] and they require the appropriate microstructure to perform their bone functions in vivo.[14] Additives are also sometimes required for improved bone efficacy during bone tissue regeneration. However, the functionalization of 3D scaffolds may reduce the mechanical performance required for bone tissue regeneration.[15,16] The 3D bone scaffolds can be made via a variety of techniques, such as solvent casting and salt leaching,[17,18] chemical foaming,[19] foam gel,[20] freeze-drying,[21] 3D printing,[22,23] and 3D scaffolds based on sintered microparticles.[14,24] Microsphere-based 3D scaffolds with a porous interconnected network were initially fabricated using a sintering technique for bone tissue engineering. This approach requires a temperature in the range of 70–120 °C or even higher. The use of a high temperature during fabrication of the material can lead to undesired side effects because several polymers are thermally sensitive and the treatment causes detrimental effects on the surface morphology. The compressive modulus of functional macromolecule-incorporated and sintered microspheres has successfully been improved. However, high molecular weight polymers are required to create strong 3D scaffolds to enable the regeneration of cancellous bone, and only a single functional species can be added.[14] The fabrication of 3D microsphere scaffolds via a combined simple technique, the simultaneous enhancement of the mechanical performance, and the inclusion of functional additives to the 3D scaffolds remain challenging.

The layer-by-layer (LbL) technique is a simple, green, and highly versatile process that is based on the alternating deposition of oppositely charged polyelectrolytes.[25−27] The technique has been applied in the field of biomedical applications.[28−30] This assembly is fundamentally driven by not only electrostatic interactions[25,26] but also covalent,[31,32] hydrophobic,[33,34] and hydrogen-bonding interactions.[33,35] Microspheres have been used as a template for assembling multilayer thin films via a LbL approach, and the concept itself is an extension to produce 3D structures ranging from nano to micrometer scales.[36−38]

Graphene oxide (GO) nanosheets have recently been exploited to enhance the mechanical performance of tissue engineering scaffolds[39,40] due to the high Young’s modulus of GO.[41] GO nanosheets have also been shown to accelerate the differentiation mechanism of bone marrow-derived mesenchymal stem cells and myoblast C2C12 cells.[42,43] The same increase in cell differentiation has been demonstrated for Fe3O4 nanoparticles.[44,45] However, less attention has been given to a new basis that uses grafted surfaces as a base layer onto which polyelectrolyte multilayers (PEMs) can be physically adsorbed, and this combined surface treatment can later be used to produce advanced 3D microsphere scaffolds at room temperature. This constitutes an excellent surface treatment combination to provide materials for 3D scaffolds for, e.g., cancellous bone regeneration.

Our main objective is here to stress the importance of a grafted surface as a base layer for the assembly of PEMs on polylactide (PLA) microsphere surfaces and to introduce a straightforward method that involves surface grafting and a LbL approach for fabricating advanced 3D scaffolding materials. Aiming for bone tissue regeneration, mechanical properties of the scaffold should mimic the E-moduli of cancellous bone. As model systems, nanosized particles can also be incorporated into the 3D scaffolds to induce conductivity or to increase the mechanical strength of the materials. The incorporation of functional additives, i.e., nanomaterials, will create a so-called sandwich scaffold, thereby altering the physicochemical properties of the 3D material. In addition, any other additives can potentially be sandwiched, driven by entropic gain, between polyelectrolyte layers in order to provide high-value materials for certain applications. We hypothesize that surface grafting onto microsphere surfaces improves the physicochemical properties of the 3D microsphere scaffolds by ensuring strong interactions between the particle and the formed PEMs. To achieve our aim, different types of weak polyelectrolytes and neutral polymers, i.e., poly(acrylic acid) (PAA), poly(allylamine hydrochloride) (PAH), and poly(acrylamide) (PAAm), were utilized. Functional nanoadditives were used to increase the scaffold conductivity or strength by alternative assembly on the microsphere surfaces between the PEMs. This strategy should be applicable to different polymers and could create highly functional materials with tailored physicochemical properties for a wide range of applications.

Experimental Section

Materials

Poly(l-lactide) (PLLA) (Mn = 50 000 g mol–1; Đ = 1.32) was synthesized via ring-opening polymerization of l-lactide (LLA, Boehringer Ingelheim, Germany) at 110 °C for 72 h in bulk using ethylene glycol (Sigma-Aldrich, Germany) as the initiator and stannous octoate (Sn(Oct)2) (Sigma-Aldrich, Sweden) as the catalyst. The formed polymer was precipitated three consecutive times in cold hexane (LabScan, Ireland)/methanol (general purpose grade, Fisher Scientific, Germany) and used to fabricate the microspheres. Benzophenone (99%, Sigma-Aldrich), ethanol (96%, VWR, Sweden), sodium hydroxide (99%, Merck), hydrogen chloride (HCl) (2 M, Fisher Scientific), chloroform (>99%, Fisher Scientific), GO (4–10% edge oxidized, Sigma-Aldrich), 4′,6-diamidino-2-phenylindole (DAPI) (Mw = 457.5; Life Technologies, UK), hexadecylamine (90%, technical grade, Aldrich, Germany), FeCO5 (>99.9%, trace metal basis, Aldrich, Germany), oleic acid (90%, technical grade, Aldrich, Germany), n-octanol (≥99%, ACS reagent, Sigma-Aldrich), and sodium periodate (≥99.8%, ACS reagent, Sigma-Aldrich) were used as received. Acrylic acid (AA) (90%, Alfa Aesar) was purified by vacuum distillation at a temperature of 40–50 °C prior to use. The three different weak polyelectrolytes, PAA (99%, Sigma-Aldrich), PAH (99%, Alfa Aesar), and PAAm (99%, Polysciences), were used as received.

Microsphere Fabrication

The PLLA microspheres were fabricated using a spray dryer (Mini Spray Dryer BUCHI B-290, Flawill, Switzerland). Approximately 1 g of PLLA was dissolved in 100 mL of chloroform. The following system parameters were used: an inlet temperature (Tinlet) of 55–65 °C, an aspirator at 100% (35 m3 min–1), a pump feed rate of 20%, and a gas flow of 50 mm (750 NL h–1). The PLLA microspheres were recovered from a collection vessel located underneath the spray dryer cyclone. They were then surface-grafted with 20% (v/v) AA under UV exposure, as previously described.[11] The PAA-grafted PLLA microspheres were subsequently used as the main substrates to build-up PEMs.

Hydrophilic Fe3O4 Nanoparticle Synthesis

Magnetic Fe3O4 nanoparticles were synthesized using a method similar to that described by Si.[46] Briefly, 3 mL of oleic acid and 0.22 g of hexadecylamine were dissolved in 8.0 mL of n-octanol via stirring and ultrasonication. Afterward, 2 mL of Fe(CO)5 was added to the solution, and the blend was mixed for 5 min via stirring. The mixture was then transferred into a 50 mL Teflon container and heated at 200 °C for 12 h in an autoclave. The magnetic nanoparticles were collected using a magnet after the autoclave reached room temperature. The Fe3O4 nanoparticles were rinsed with ethanol and distilled water three consecutive times, and they were dried under vacuum overnight. To introduce hydrophilic groups onto the surface of the nanoparticles, 100 mg of the synthesized magnetic nanoparticles, which were dispersed in 15 mL of cyclohexane and 15 mL of 27 mg/mL sodium periodate aqueous solution, was added into a 20 mL mixture of ethyl acetate and acetonitrile with a volume ratio of 1:1 and stirred vigorously for 2 h. After the reaction, the nanoparticles were again collected using a magnet and rinsed with ethanol and distilled water several times. Then, the magnetic nanoparticles were redispersed in water.

LbL Formation

The PEMs were formed similar to those that are alternately assembled on a film substrate, as described elsewhere.[47] This technique involves hydrogen-bonding interactions between an anionic (PAA) and a neutral polymer (PAAm) polyelectrolyte at pH 2.5 (abbreviated PAA/PAAm) or electrostatic interactions between an anionic (PAA) and a cationic polymeric (PAH) polyelectrolyte at pH 2.5 and 5 (abbreviated as PAA/PAH). Both systems were alternately deposited onto PAA-grafted PLLA microspheres. Briefly, immersion into a 0.3 M polyelectrolyte solution (based on the repeating unit of molecular weight) in 18.2 MΩ cm Millipore water was performed for approximately 15 min under mechanical stirring, followed by centrifugation at 4000 rpm for 3 min. Water was added to physically remove the unabsorbed polyelectrolyte in the solution between each deposition step. The water volume was two times that of the polyelectrolyte solutions.

The rinsing step was carried under vigorous stirring and was subsequently followed by centrifugation. Thus, a deposition/centrifugation/rinse/centrifugation cycle was performed to obtain the PEMs and to remove free polyelectrolytes from the solution. The pH of the rinsing water was adjusted to be the same as the polyelectrolyte solution, i.e., pH 2.5 and 5. The pH of the PAH and PAA dipping solutions was experimentally adjusted to 2.5 and 5, respectively, whereas a pH of only 2.5 was used to create the PAA and PAAm multilayers. To the system with the outermost layer of PAAm or PAH at pH 2.5 was deposited 0.002 wt % of a functional additive as a model nanomaterial (i.e., GO or Fe3O4) onto the microsphere surfaces. The nanomaterial, i.e., GO or Fe3O4, acted as a third component after sequential adsorption of polyanions and polycations/neutral polymers onto grafted surfaces, meaning that the sequential adsorption of these nanomaterials was carried out an equal number of times as the number of formed bilayers following the desired layer of polyanions and polycations/neutral polymers. Nongrafted PLLA microspheres were also prepared using the same procedure to build-up PEMs as reference (Table ).

Table 1

Surface-Grafted Microspheres Used during the Assembly of PEMs and Their Nongrafted Microspheres

microspheres	type of multilayer build-up	pH	name
nongrafted PLLA	PAA/PAAm	2.5	PAMn
	PAA/PAH	2.5	PAH2.5n
	PAA/PAH	5	PAH5n
PAA-grafted PLLA	PAA/PAAm	2.5	PAMg
	PAA/PAH	2.5	PAH2.5g
	PAA/PAH	5	PAH5g
	PAA/PAAm/GO	2.5	PAMgGO
	PAA/PAH/GO	2.5	PAHgGO
	PAA/PAAm/Fe3O4	2.5	PAMgFeO
	PAA/PAH/Fe3O4	2.5	PAHgFeO

3D Scaffold Fabrication

After rinsing with water, the PEM-coated PLLA microspheres whose outermost layers consisted of carboxylates were mixed with amide layers that contained microspheres in one batch. Later, they were rinsed with water and centrifuged. This procedure was similarly performed for the electrostatic-driven interaction with the PAA and PAH outermost layers. They were packed into a cylindrical cutoff plastic syringe with a diameter of 8–10 mm and a height of 4–8 mm (Figure ). The 3D microsphere scaffolds were subsequently dried for 1 week prior to characterization.

Figure 2

Schematic illustration of the surface grafting and build up of PEMs onto the grafted surface of PLLA microspheres.

Characterization

Size Exclusion Chromatography (SEC)

The molar mass of the PLLA microspheres was determined using a Waters 717 plus autosampler and a Waters model 510 apparatus equipped with two PL gel 10 μm mixed B columns (300 × 7.5 mm; Polymer Laboratories, UK). The instrument was calibrated using narrow molar mass distribution polystyrene standards with molar masses in the range 580–400 000 g mol–1.

Fourier Transform Infrared Spectroscopy (FTIR)

The formation of PEMs that were attached to the PLLA-based microspheres was characterized using FTIR. Spectra were recorded in the range 4000–600 cm–1 on a Spectrum 2000 PerkinElmer spectrometer using an attenuated total reflectance (ATR) accessory (Golden Gate).

X-ray Diffraction (XRD)

The crystalline structure of the GO-functionalized materials and their respective references was evaluated using XRD on a PANalytical XPert Pro instrument with Cu Kα radiation at a wavelength (λ) of 1.54 Å and a voltage and electric current of 45 kV and 45 mA, respectively. The diffractograms were recorded at 25 °C using a silicon monocrystal sample holder at a step size angle of 0.017°. The intensity of the diffractogram was measured as a function of 2θ ranging between 5 and 60°.

Scanning Electron Microscopy (SEM)

The surface morphology of the microsphere-based samples was determined using a Hitachi S-4800 scanning electron microscope at an acceleration of 1 kV. The samples were mounted on adhesive carbon black and were sputter-coated.

Compression Test

The mechanical testing was conducted on an Instron 5566 instrument (Instron, UK) with a load cell of 10 kN and a compressive rate of 10% thickness min–1. The 3D microsphere scaffolds had a height between 5 and 8 mm and a diameter between 8 and 10 mm.

Particle Size Distribution (PSD)

A predetermined amount of microspheres (approximately 600 mg) was prepared. The PSDs of neat PLLA, PLLA-g-PAA microspheres, and PEM-coated microspheres were determined at a distance of 10 mm from the spray outlet via a laser diffraction spectrometer with a He–Ne laser at a wavelength of 632.8 nm as the laser source (model HELOS/KR-H3522, Sympatec GmbH, Clausthal-Zellerfeld, Germany), according to ISO 13320:2009. The microspheres were previously dispersed using a RODOS dispersing unit at a reduced pressure of 2 bar. The HELOS was operated with a focal distance of fR3 = 100 mm to span a distance between 0.5 and 175 μm. The measurement was performed for 1.5 s for each sample. The PSD was then evaluated using the Fraunhofer diffraction model.

Porosity Measurement

The porosity of the scaffolds was measured using a μCT device (SkyScan 1172 Scanner, Belgium) at 40 keV with a 2.4 μm voxel. Samples were randomly selected and characterized via CTan (CT analyzer, ver. 1.5.0, SkyScan) and CTvol (CT Vol Realistic, ver. 1.9.4, SkyScan) software. Three-dimensional models were later reconstructed using NRecon (ver.1.6.9, SkyScan) and CTan (v.1.14.4, SkyScan) software.

Electrical Conductivity Measurement

The 3D scaffolds were doped with 1 mol L–1 HCl for 24 h. Nondoped 3D scaffolds were also prepared as a reference. After Cl– doping, the scaffolds were dried in a vacuum oven for 48 h. Thus, the conductivity was not influenced by the water content. Microsphere pellets with diameters of 0.6–0.9 cm and thicknesses of 0.3–0.4 mm were made via compression molding using a pressure of 150 kN m–2 at 100 °C for 10 min. The pellets were placed between two stainless steel plates, and an electric current of 1 mA was then passed through the plates. The voltage of the nondoped and doped 3D microsphere scaffolds was measured using symmetrical cylindrical cells. The recorded voltage was later used to calculate the electrical conductivity according to Pouillet’s law.[48]

Tunneling Electron Microscopy (TEM)

The structure of the nanomaterials was determined using a Hitachi HT7700 transmission electron microscope at a voltage of 100 kV. The samples were deposited on 400 mesh copper grids that were coated with carbon (Ted Pella, Inc., USA) from ultrasonicated water suspensions of the nanomaterials.

Zeta Potential Measurement

The zeta potential of the PEMs that were attached to the grafted microspheres was determined using a Nanozetasizer with a He–Ne laser source at a wavelength of 633 nm (Zetasizer ZEN 3600, Malvern Instruments, UK). The microspheres were injected into the zeta potential cell and allowed to equilibrate for 2 min. Duplicate samples were measured. During the measurement, the Slomuchowski method was selected to obtain the zeta potential.

Fluorescence Microscopy

Fluorescence measurements were performed using a Nikon Ti–S spectrometer. Prior to the surface fluorescence tests, the PEM-coated microspheres were kept in the oven after drying at a temperature of 55–65 °C for 1 week to create strong interactions and good stability for the particulate samples at a high pH of 7.4. Approximately 10–20 mg of the particulate samples was placed in glass vials and immersed in 0.5 mL of 3 mM DAPI (pH 7.4) at room temperature for 30 min. The vials containing DAPI solution and the microsphere-based samples were protected from light. The PEM-coated microspheres were pipetted onto a microscopic glass slide and covered with a coverslip.

Swelling Test

The swelling ratio (SR) of the particulate scaffolds was determined by immersing dry 3D microsphere scaffolds in water at pH 2.5 and 5. The weight at different time points of the swollen-state samples (ms,t) was recorded after removing excess water with tissue paper. Then, the SR was calculated from the following equationand the equilibrium swelling, Q∞, is given bywhere md and m∞ are defined as the weight of the samples in the dry state and at equilibrium, respectively.

X-ray Microtomography

A volume of 1 mm (diameter) × 1 mm (length) for the 3D microsphere scaffolds was acquired using microtomography (Xradia MicroXCT-200). The distances from the X-ray source and the detector were 40 and 7 mm, respectively. The following scanning conditions were then set up: a power of 6 W, an X-ray source voltage of 60 kV, 1105 projections, and an exposition time of 25 s projection–1. The magnification was 20× with a pixel resolution of 1.1519 μm.

Results and Discussion

Here, we show the importance of an AA-grafted PLA surface as a microsphere base layer for alternately assembled PEMs for the fabrication of strong and functional 3D microsphere scaffolds that can be made to be conductive and/or stronger.

After surface grafting and PEM assembly (polyanionic/polycationic layers or polyanionic/neutral polymer layers) onto the surfaces of the PLLA microspheres, the microspheres were molded at room temperature in a form (Figure , left, and Figure ), resulting in 3D scaffolds held together by intermolecular forces through electrostatic and/or hydrogen-bonding interactions. For the electrostatic interactions, strong attractive forces were created between oppositely charged surfaces. Interpenetration of polyelectrolytes could also occur to form polyelectrolyte complexes during this interaction, resulting in strong adhesive forces. The 3D scaffolds exhibited strong resistance to high-shear mixing (Video S1). Without adsorption of the PEMs, the PLLA microspheres could not be shaped into a permanent cylindrical 3D shape and could be very easily ruptured after 3D template removal. In addition, PLLA microspheres without any polyelectrolytes or grafted chains sintered at 55–65 °C were also easily ruptured under high ultrasonication or high-shear mixing.

Figure 1

Comparison of the fabricated 3D microsphere scaffolds to 10 cent Euro coins: (left) PAH2.5g, (middle) PAHgGO, and (right) PAHgFeO. All 3D microsphere scaffolds were assembled at pH 2.5.

The 3D scaffolds were also functionalized with model additives i.e., GO, Fe3O4, and DAPI, to illustrate the versatility of the technique. Functionalization of GO nanosheets into a 3D microsphere scaffold that was assembled via the LbL technique resulted in a darker gray color (Figure , middle) compared with that of the Fe3O4-functionalized 3D scaffold (Figure , right).

Thus, 3D microsphere scaffolds were successfully fabricated and functionalized with different functional additives, i.e., GO, Fe3O4, and DAPI (Figure S5).

Functionalities of the PEMs on the Surface of PLLA Microspheres

The presence of PAA-grafted PLLA surfaces as a base layer significantly affected the physical adsorption of weak polyelectrolytes, resulting in denser absorption bands for the functionalities than with the nongrafted surfaces. The surface chemistry of the grafted microspheres after the assembly of the PEMs was evaluated using FTIR (Figure ) and compared with that of the nongrafted microspheres (Figure S3). During FTIR measurement, the higher intensity of the functional groups located on the grafted surfaces (Figure A,B) than on nongrafted surfaces (Figure S3A,B) confirmed the 2D coverage of PEMs and highlighted the importance of surface grafting to achieve dense multilayer films. These functional groups were primarily observed in two wavenumber (υ) regions, i.e., υ ∼ 1680–1500 cm–1 (the area where the carboxylic acid group and the ionized carboxylate groups of PAA and the amide I and II groups of PAAm exists) and υ ∼ 3800–2300 cm–1 (the absorption bands of the amine and methylene).

Figure 3

ATR-FTIR spectra of neat PLLA microspheres (black) and PEMs alternately assembled onto the surface of PLLA-g-PAA microspheres with PAMg (red), PAMgGO (magenta), PAMgFeO (orange), PAH2.5g (green), PAHgGO (violet), PAHgFeO (gray), and PAH5g (blue) in the region of (A) 1680–1500 cm–1 and (B) 3800–2300 cm–1. (C) XRD patterns of the 3D microsphere scaffolds with and without GO functionalization.

As expected, the neat PLLA microspheres exhibited an absorption band representing ester groups at 1747 cm–1 (Figure A). The FTIR spectra confirmed the growth from the first to the sixth PEM bilayer on the PAA-grafted surfaces (Figure A,B) and nongrafted analogues (Figure S3A,B) by the increase in the intensities of the corresponding functional groups. The zeta potential also confirmed the stepwise growth of the PEMs onto the grafted surfaces via charge reversal (Figure S2). The surface grafting had a strong affect on the peak area of the PEMs in the 1680–1500 and 3800–2300 cm–1 regions relative to the nongrafted microspheres (Table ). The physical adsorption of the PEMs on the surfaces of both the PAA-grafted and nongrafted PLLA microspheres is a complex process, and it could influence the conformational entropy of the acrylic acid groups of PAA and the amine-methylene groups of PAH during PEM assembly.

Table 2

Peak Areas of Nongrafted and Grafted PLLA Microspheres after Assembly of the Sixth PEM Bilayer

	peak area (A cm–1)a
	nongrafted		grafted
materials	υ ∼ 1680–1500 cm–1	υ ∼ 3800–2300 cm–1	υ ∼ 1680–1500 cm–1	υ ∼ 3800–2300 cm–1
PAMn	18	8
PAMg			20	15
PAH2.5n	20	15
PAH2.5g			20	19
PAH5n	15	5
PAH5g			15	13
PAMgGO			23	12
PAHgGO			19	12
PAMgFeO			17	13
PAHgFeO			19	18

Calibrated by subtracting the peak area of PEM-coated microspheres from the peak area of neat PLLA, which was approximately 0.3 at 1680–1500 cm–1 and 0.2 at 3800–2300 cm–1. The results of the peak area were subsequently rounded.

For the GO-functionalized particles (Figure S1A), the peak area in the 1680–1550 cm–1 region differed between PAH2.5g and PAMg, whereas Fe3O4 nanomaterials (Figure S1B) impacted the peak area in two different regions, i.e., 1680–1500 and 3800–2300 cm–1, for those two materials (Table ). The absorption peak for the Fe–O stretching bond[46] in the range of 600–580 cm–1 was difficult to detect using FTIR spectroscopy (Figure S3C), suggesting that the IR penetration depth and the concentration may have influenced the detection of the functionalities of the material.

XRD measurements of the two GO-functionalized materials and their respective references showed a low intensity of 12.3° for the detection of the GO stacking layer at 2θ in the PAHgGO material, which corresponded to an interlayer spacing of 0.72 nm for the GO nanosheets that was calculated from Bragg’s law (arrow, Figure C). This result confirmed the existence of stacked GO layers[49] between the PEM films. The XRD pattern of GO was quite low, which could be related to the very small amount of GO incorporated in the PEMs.

The XRD pattern of GO was observed in PAHgGO and not in PAMgGO. This might be related to the primary or secondary physical adsorption of nanomaterials interpenetrated through the PEM network. Primary physical adsorption (deep adsorption) might have occurred during GO functionalization into PAA/PAAm multilayers, whereas GO nanomaterials might have attached to the outermost layers of polyelectrolytes through secondary physical adsorption for PAA/PAH multilayers; thus, XRD could detect the existence of GO nanomaterials located at the outermost surface of the PEM rather than the ones adsorbed deeper in the material. Three sharp peaks were located at 2θ values of 16.9, 18.9, and 22.3°, which are characteristic of PLLA[50] (Figure C). Hence, grafting of the AA monomer on the surface of the PLLA microspheres provided a denser PEM assembly; thus, the grafted surfaces constitute a crucial factor for alternately assembling (PAA/PAAm)6 and (PAA/PAH)6 multilayers. Similar to the surface chemistry, the grafted surfaces also influenced the swelling behavior of the PEMs. It is worth noting that (PAA/PAH)6 multilayers demonstrated a larger swelling effect than (PAA/PAAm)6 multilayers at pH 2.5 (Figure S6).

Morphology of the PEMs on the Grafted Surfaces

The change in morphology between neat PLLA and PLLA-g-PAA microspheres demonstrated significant surface texture differences. Neat PLLA microspheres with an average diameter of 2 μm (Table ) had a rough surface (Figure A), whereas PLLA-g-PAA microspheres had a “crystal”-like structure (Figure B). The crystal structure might be a consequence of the benzophenone that terminates the free-radical polymerization under UV irradiation[51] or a solvent-induced crystallization process.[52] The difference in the morphology of these two base substrates subsequently influences the sequential adsorption of PEMs since the growth of PEMs consists of three main zones, where zone I is a function of the substrate surface and zone III demonstrates that excess charge affects the final structure of the PEMs.[53] PAMg also had a rough morphology (Figure C), whereas a smooth surface was observed for PAMn (i.e., the nongrafted surface) after PEM assembly (Figure S4A). A significant change in the morphology was observed for PAH2.5g because the polyelectrolyte charge induced the conformational transition directly[54] (Figure D). In other words, the polyanion/polycation multilayers provided a rough surface at low pH because of the low charge density and because the formation of segmental loops and tails is more favored than at high pH, resulting in thicker PEMs than those from sequential adsorption at high pH. Additionally, the grafted surfaces demonstrated large polyanion/polycation multilayer surface coverage at low pH (Figure D) compared to that of the nongrafted analogues (Figure S4B). In contrast to the surface morphology, no significant change in emission color was observed for a fluorescent dye molecule that was functionalized into the PEMs and physically adsorbed on the grafted and nongrafted surfaces (Figure S5).

Table 3

Microsphere Characteristics, Layer Thickness of the Microsphere-Based Samples, and Porosity Data for the 3D Scaffolds

	microsphere properties
material	mean diameter,ad (μm)	Smb	layer thicknessc (μm)	porosity of 3D scaffold in air (%)
PLLA	2	1.2		d
PLLA-g-PAA	3.5 ± 0.7	0.6	1.5 ± 0.7	d
PAH2.5g	11.5 ± 0.7	0.2	9.5 ± 0.7	48
PAMg	7	0.3	5	58
PAHgGO	4	0.6	2	50
PAMgGO	12	0.2	10	50

Calculated based on Sauter’s method. This diameter may be calculated from single or aggregated microspheres.

Sm = specific surface area per unit mass.

The layer thickness of the PEMs was calculated by subtracting the diameter of PEM-coated microspheres from that of neat PLLA microspheres. The diameter was measured using laser diffraction.

Not possible to produce 3D scaffolds.

Figure 4

Surface morphologies of the microspheres of (A) neat PLLA, (B) PLLA-g-PAA, (C) PAMg, (D) PAH2.5g, (E) PAH5g, (F) PAMgGO, (G) PAHgGO, (H) PAMgFeO, and (I) PAHgFeO.

A smoother surface morphology was observed for PAH5g compared to that for PAH2.5g after the adsorption of the PEMs (Figure E). This is believed to be caused by a conformational transition in the (PAA/PAH)6 multilayers when the pH of the solution was increased to 5. At higher pH, the enthalpy required to maintain the charged polymer in a stretched conformation is able to overcome the entropic penalty of the globule conformational transition at a certain critical charge density. The result is that the polyelectrolyte chains that were adsorbed on the grafted microsphere surfaces may have formed a larger number of trains, resulting in stretched conformations.[47] Thus, the surface of the nongrafted PLLA microspheres remained rough after increasing the pH of the solution, in contrast with the grafted PLLA microspheres (Figure S4C).

At low pH, functionalization with GO or Fe3O4 nanomaterials improved the surface coverage of the grafted microspheres and altered their morphology after the adsorption of the (PAA/PAAm)6 (Figure F,H) and (PAA/PAH)6 (Figure G,I) multilayers, whereas no significant alteration in the surface morphology of the (PAA/PAAm)6 and (PAA/PAH)6 multilayers occurred after GO functionalization at low pH (Figure F,G). In summary, the surface grafting influenced the morphology of the microsphere surfaces during the assembly of the PEMs. Additionally, the grafted surfaces also demonstrated an improved surface coverage of the PEMs compared with that of the nongrafted analogues.

Mechanical Properties of the 3D Microsphere Scaffolds and Their Comparison to the Cancellous Bone

The 3D microsphere scaffolds (Figure A) were mechanically compressed under a load cell of 10 kN to evaluate their mechanical performance in comparison to that of natural materials, such as cancellous or trabecular bone. PAH2.5g had an E-modulus of 141 ± 19 MPa, whereas PAMg had a lower E-modulus of 115 ± 5 MPa. These two 3D microsphere scaffolds demonstrated a higher Young’s modulus than PAH5g (Figure B). The PAMg and PAH2.5g 3D scaffolds were selected, and new scaffolds, previously functionalized with GO nanosheets, were created, namely, PAMgGO and PAHgGO. After GO functionalization, the E-modulus of PAMgGO improved significantly to 146 ± 6 MPa, whereas PAH2.5gGO decreased significantly to a very low E-modulus of 16 ± 2 MPa.

Figure 5

(A) Illustration of the 3D microsphere scaffold. (B) Summary of the compressive Young’s modulus (E-modulus) of the 3D microsphere scaffolds (nongrafted microspheres resulted in irreproducible data). (C) Ashby plot of E-modulus vs the density of natural materials and the 3D microsphere scaffolds. (D) X-ray microtomography of the 3D microsphere scaffold.

The mechanical properties of the scaffold increased (or decrease) for PAMg (or PAH2.5g) after alternatingly adsorbing GO to the particles, which could be related to an increase (or decrease) in the thickness of the PEM layer of PAMg (or PAH2.5g). Low pH generally results in GO nanosheet agglomeration.[55] At pH 2.5, PAHgGO demonstrated thicker layers than PAMgGO (Table ). The thickness of the charged polymers depends on the surface charge density and the charge density of the adsorbing polymer.[47] The PAA/PAH multilayers (polyanionic/polycationic combination) may have demonstrated a higher surface charge density than the PAA/PAAm multilayers (polyanionic/neutral polymer), resulting in more GO nanosheets binding to these charged polymers per unit area. The interpenetration of GO nanomaterials during LbL assembly, then, might have interrupted the chain conformations of the PAA/PAH multilayers, thus physically suppressing the thickness of the PEMs. When the thickness decreased, weaker mechanical performance was observed. The E-modulus of the nongrafted 3D scaffolds could not be evaluated due to the weak 3D scaffolds. More specifically, the 3D scaffold fabricated from (PAA/PAH)6 multilayers on nongrafted surfaces at pH 2.5 cracked after the 3D template was removed. The 3D scaffolds of PAMg, PAH2.5g, and PAMgGO demonstrated similar mechanical performances to those of sintered microspheres of diameters between 100 and 600 μm when they were sintered between 60 and 70 °C,[56] at 100 °C for 4 h,[14] and at 160 °C for 2 h.[57] However, these 3D microsphere scaffolds had a higher mechanical performance compared with that of solvent/nonsolvent sintered microsphere scaffolds.[58] As previously noted, an increase in the temperature and sintering time resulted in a detrimental effect on the surface morphology. Additionally, the porosity was decreased after the sintering process.[14,56] Furthermore, the E-moduli of all of the 3D microsphere scaffolds that were fabricated via surface grafting and LbL assembly were in the range of cancellous bone from a low to high E-modulus, i.e., 10–2000 MPa. However, the cancellous human bone described in the Ashby plot database[59] demonstrated a high range for the E-modulus (>100 MPa)[60] at medium density, and three of our 3D microsphere scaffolds closely matched natural human bone, namely, PAH2.5g, PAMg, and PAMgGO (Figure C). Similar to the XRD analysis, X-ray microtomography of the 3D scaffold, i.e., PAMgGO, did not detect a signal for GO inside the 3D scaffold. Thus, the distribution of GO could not be calculated, but it showed that the 3D biomaterial was homogenous (Figure D and Video S2). However, the 3D microsphere scaffolds fabricated via LbL assembly were successfully functionalized with GO and retained good mechanical performance from a low to high Young’s modulus.

Physical Properties of the Microspheres

Combined surface treatments (surface grafting and physical adsorption of PEMs) on the PLLA microspheres had a large impact on the diameter and thickness of the PEMs (Table ). The particle diameter was evaluated using Sauter’s method since this method is commonly used to evaluate the diameter of a particle when the surface is active. Such a substantial increase in layer thickness would, at first glance, indicate particle aggregation. However, a shape factor equivalent to 1, referring to a spherical geometry, was observed for all of the particulate samples during particle size measurement, indicating single particles rather than spherical aggregates. An aggregated spherical complex of particles would be substantially larger than the measured values in Table .

It is worth noting that traditional LbL assembly is generally used to create ultrathin PEMs (from several tenths to a hundred nanometers in thickness),[47,61] and several hundred-nanometer-thick layers were achieved after numerous bilayers were assembled, which depended on parameters such as pH.[62] For a specific application, such as 3D biomaterials, micrometer-thick PEMs might be favored over nanometer-thick PEMs since the mechanical stability can be enhanced. Thick PEMs, i.e., in the micrometer range, could also be obtained after the addition of numerous bilayers.[63,64] Here, micrometer-thick films were obtained on the particulate surfaces after grafting due to exponential growth of the PEMs and due to the conditions used during LbL assembly, i.e., the stirring rate and polymer concentration in the solution. The mechanism of an exponential growth regime is not well understood.[65] It is speculated that exponential growth of PEMs can be obtained through a complexation process of polyanions/polycations[66] or diffusion “in and out” of the polyanions/polycations, which subsequently forms polyelectrolyte complexes at the outermost layer of the physically adsorbed film.[65,67] At one-tenth of the polyelectrolyte concentration used, the 3D scaffold ruptured under the high-shear mixing that was used as a pretest for mechanical stability, which could be due to the presence of a linear growth regime at a concentration that is lower than the actual concentration we used. Thus, the concentration of polymer in the solution and the stirring process used in this work might have accelerated the exponential growth process to support the integrity and mechanical performance of the 3D scaffold. Surface grafting on PLLA microspheres yielded a polymer graft thickness of 800 nm to 2.2 μm in the dry state. The (PAA/PAH)6 multilayers adsorbed onto the PAA-grafted PLLA microspheres had a diameter of 11.5 ± 0.7 μm, which was 4–5 μm thicker than that of the (PAA/PAAm)6 multilayers. This could be due to the formation of a loop conformation on the PLLA surface (Figure C). Functionalization of the GO nanosheets resulted in an alteration of the properties of the microsphere material. The (PAA/PAAm)6 multilayers showed an increase in the diameter and thickness and a decrease in the surface area of PLLA microspheres compared to those of the (PAA/PAH)6 multilayers assembled on grafted PLLA microspheres at low pH. The higher E-moduli of two of the 3D microsphere scaffolds, PAH2.5g and PAMg (Figure B), may also be related to the microsphere properties because of their thicker layer (Table ), which improves the mechanical performance. Importantly, the porosity of PAMg and PAMgGO in air matched that of human cancellous bone (50–90%),[60] whereas the porosity of PAH2.5g (48%) was comparable to that of porous scaffolds that have been used to regenerate femoral defects in rabbits.[68]

During bone regeneration, an electric field can exogenously stimulate new bone formation at a minimum frequency between 10 and 30 Hz. The bone tissue will later respond to a minimum electrical signal in the range of 1–10 μV cm–1.[69] Thus, the conductivity of the scaffolds also plays a critical role in electrical signaling. The conductivity of the 3D scaffolds consisting of different nanomaterials was measured, and neither the nonfunctionalized materials nor the GO-functionalized scaffolds were conductive. GO itself is an electrically insulating material, but an electrical conductivity of approximately 2.58 × 10–6 S cm–1 was previously observed.[70] It could also be that the GO-functionalized scaffolds were not conductive because of the very low concentration of GO nanomaterials, which are required to pass the electric current, in the scaffolds. In contrast to the GO-functionalized scaffolds, the Fe3O4-functionalized scaffolds resulted in a high electrical conductivity, which could be due to the regular organization of Fe3O4 nanoparticles between layers after sequential adsorption. In addition, Fe3O4 nanomaterials can transfer electric current through their aligned electron domains in parallel, a phenomenon called electronic conduction. The hot press PAHgFeO and PAMgFeO films had electrical conductivity values of approximately 180 and 130 S cm–1, respectively, after 1 mA electric current was passed through 0.3–0.4 mm thick films, although only 0.002 wt % of Fe3O4 nanomaterials was functionalized on the surface of the 3D scaffolds. After doping, the electrical conductivity of PAH2.5g decreased approximately 60 times, i.e., 3 S cm–1, compared with that of the nondoped microparticle-based scaffolds. In contrast to PAHgFeO, PAMgFeO exhibited a much higher electrical conductivity of approximately 240 S cm–1 compared with that of nondoped PAMgFeO. Humidity has been known to affect electronic conductivity,[71] but humidity was not controlled during conductivity measurements in this work. A possible explanation for this phenomenon might be related to the swelling effect of PAA/PAH multilayers during HCl doping. In addition, PAA/PAH multilayers seemed to demonstrate more swelling than PAA/PAAm multilayers, causing an extension of the chain conformations. As a consequence, the orientation of iron nanoparticles was also altered, causing a large gap between these nanoparticles in the PAA/PAH multilayers. The pathways of unconnected nanoparticles seemed to be constant in the PAA/PAH multilayers, even after they were dried and hot pressed. The tunnelling probability of electrons will then decrease when the electric current is passed through this material in air. Thus, the electronic conductivity of PAHgFeO, where iron nanoparticles were located between PAA/PAH multilayers, decreased dramatically more than PAMgFeO, which consisted of iron nanoparticles sandwiched in PAA/PAAm multilayers. The functionalization of Fe3O4 nanoparticles interpenetrated into (PAA/PAAm)6 multilayers had a stronger impact on the magnetic field than that of the (PAA/PAH)6 multilayers when a magnetic source was located 4–5 mm on top of the surface of the 3D scaffolds (Video S3).

The success of the 3D microsphere scaffolds fabricated via surface grafting in combination with the LbL approach at room temperature demonstrates that these scaffolds can be functionalized with inorganic nanomaterial or biofunctional species between different layers while retaining porosities and mechanical performances that are similar to those of human cancellous bone.

Conclusions

Surface grafting was shown to be a crucial technique to create a base layer for the alternating assembly of weak polyelectrolytes onto the surface of PLLA microspheres. This simple method can be used to fabricate advanced 3D scaffolding materials that are strong, functional, and/or highly conductive. Conformational changes in the PEMs at different pH values resulted in variations in the layer thickness, thus affecting the mechanical performance of the 3D microsphere scaffolds. Grafted surfaces of PLLA microspheres exhibited a larger peak area for the absorption bands of PEMs compared with that of nongrafted surfaces. Alteration of the morphology produced an improved surface coverage for the PEMs assembled at low pH on the grafted surfaces compared with that of the nongrafted surfaces. A blue emission color indicated the success of the functionalization of a fluorescent dye onto two different substrate surfaces. The PEMs that were physically adsorbed on the grafted surfaces had more pronounced swelling behaviors than those of nongrafted surfaces. The 3D microsphere scaffolds were within the compressive range of human cancellous bone, in addition to having a similar porosity. The functionalization of magnetic nanoparticles at low concentration into the PEMs resulted in a high electrical conductivity. The 3D scaffolds of polyanionic and polycationic systems reduced their electrical conductivity after doping, in contrast to the polyanionic and neutral polymers. Thus, surface grafting of PLLA microspheres in combination with the LbL approach and functionalization with nanomaterials created 3D scaffolds that are suitable for bone tissue engineering.