Solventless grafting of functional polymer coatings onto Parylene C

The integration of Parylene C into biomedical devices has been a subject of great interest due to its biocompatibility,¹ dielectric properties,² solvent resistance,³ and chemical stability.^2,4,5 These attractive properties have led to the wide use of Parylene C coatings on a variety of biomedical devices such as shunts,⁶ metal stents,⁷ and implantable neural prosthesis.^8,9 While Parylene C exhibits desirable properties, its hydrophobicity promotes the adsorption of cells and proteins, limiting its antifouling ability.^1,10–12 For this reason, multiple studies have investigated the surface modification of Parylene C to improve its hydrophilicity.^10,13

There are many ways to modify the surfaces of polymer substrates with polymer coatings. For example, spin coating has been used to apply poly(4-methyl-1-pentane) onto microporous polysulfone,¹⁴ poly(methyl methacrylate) has been dropcasted onto poly(dimethylsiloxane) micropillers,¹⁵ and poly(2-vinyl pyridine) has been polymerized in solution phase on a silicon wafer and then transferred to polyester membranes.¹⁶ These techniques create coatings that are physisorbed to their polymer substrate, which can lead to leeching or degradation.¹ For practical applications, it is necessary to covalently attach polymer coatings to Parylene C. For example, solution phase grafting was used to covalently attach poly(2-methacryloyloxethyl phosphorylcholine) to Parylene C to provide enhanced lubrication to an implantable biomedical device,¹⁰ and solution phase photografting was used to covalently attached poly(acrylamide) to Parylene C to provide improved hydrophilicity.¹³ 

Since solution-phase coating techniques have inherent limitations related to surface tension effects, solvent compatibility issues, and solvent leeching, the development of dry methods to covalently attach coatings is necessary. The use of initiated chemical vapor deposition (iCVD) is promising because it is an all-dry, vacuum process that could be used to deposit conformal polymer coatings with a wide range of functionalities. For example, iCVD has been used to conformally apply ultrathin zwitterionic coatings to polyamide membranes to prevent biofouling¹⁷ and has been used to apply poly(1H,1H,2H,2H-perfluorodecyl acrylate-co-ethylene glycol diacrylate) onto polydimethylsiloxane microfluidic channels to control droplets.¹⁸ In the iCVD process, a thermal initiator and a monomer are introduced into a vacuum chamber and the initiator is broken into free radicals by a heated filament array. These radicals, along with the monomer molecules, are adsorbed to the surface of the substrate and polymerized via a free-radical mechanism. The iCVD coatings are physisorbed to their substrates. A modified version of iCVD called photoinitiated chemical vapor deposition (piCVD) has been used to covalently attach polymers to other polymer substrates. For example, a previous study by Martin et al. showed that (dimethalamino)methyl styrene and (diethylamino)ethyl acrylate could be grafted onto nylon fabric allowing the nylon to retain its conformality while adding antimicrobial functionality.¹² In the piCVD method, a photoinitiator such as benzophenone is used instead of a thermal initiator.^11,12 Ultraviolet (UV) light causes benzophenone to create a diradical that abstracts a hydrogen atom from the surface of the polymer substrate.^13,19,20 The resulting radical on the surface then initiates free radical polymerization of the monomer.^11,12,20 Although piCVD can achieve grafting, the grafting densities can be low and the exposure to benzophenone and UV radiation throughout the polymerization process can result in undesirable side-reactions or the destruction of sensitive functionalities which ultimately limits the chemistries that can be deposited using piCVD.^11,12

In order to modify the surface of Parylene C with robust functional coatings that have high surface coverage, we developed a procedure that involves a combination of piCVD and conventional iCVD. The coatings are composed of a grafted cross-linked anchoring bottom layer and a functional top layer. The cross-linked anchoring layer is made using benzophenone as the photoinitiator in order to covalently attach the coatings to the surface of Parylene C, whereas the functional top coating is made using a less reactive thermal initiator, di-tert-butyl peroxide (TBPO), to preserve surface functionality. We demonstrate the versatility of our method by depositing coatings with top layers consisting of a biocompatible polymer poly(vinyl pyrrolidone) (PVP) and a photoresponsive polymer poly(ortho-nitrobenzyl methacrylate) (PoNBMA). PVP was chosen because it is hydrophilic, biocompatible,^11,21,22 and the amide functionality provides a nitrogen environment to easily differentiate it from other components during x-ray photoelectron spectroscopy (XPS) analysis. PVP is also of interest for biomedical applications because of its antimicrobial and antifouling properties,^11,21,23 as well as its potential as a drug delivery agent.²⁴ PoNBMA was chosen for its potential as a photoresponsive coating.²⁵ We show that both top layers retain their functionalities with PVP remaining hydrophilic and PoNBMA retaining its photoresponsiveness by switching from a hydrophobic to a hydrophilic state after UV light exposure. Although we demonstrate our procedure on Parylene C surfaces, our method can easily be extended to any surface containing labile hydrogen atoms that can be abstracted by benzophenone radicals. Our process can also be extended to graft other functional polymers that can deposited via iCVD, including thermoresponsive poly(N-isopropyl acrylamide),²⁶ hydrophilic poly(2-hydroxyethyl methacrylate),²⁷ and hydrophobic poly(1H,1H,2H,2H-perfluorodecyl acrylate).²⁸ 

Benzophenone (BP) (99%, Sigma-Aldrich), TBPO (98%, Sigma-Aldrich), 1-vinyl-2-pyrrolidone (VP) (97%, Sigma-Aldrich), ethylene glycol diacrylate (EGDA) (90%, Sigma-Aldrich), ortho-nitrobenzyl methacrylate (oNBMA) (95%, Polysciences, Inc.), and methanol (Macron, absolute) were used without further purification. The coatings were deposited in a vacuum chamber (GVD Corporation, 250 mm diameter, 48 mm height) equipped with a nichrome filament array (80% Ni, 20% Cr, Omega Engineering). Reactor pressure was maintained using a throttle valve (MKS 153D) and measured using a capacitance manometer (MKS Baratron 622A01TDE). BP was heated to 70 °C, EGDA was heated to 35 °C, and VP was heated to 60 °C. The substrate temperature was controlled using a 4 × 4 × 0.47 cm thermoelectric cooler (TEC) (Custom Thermoelectric). The stage temperature was kept constant at 42 °C throughout the deposition using a recirculating chiller to prevent deposition onto the rest of the stage. Ultraviolet light irradiation was accomplished using a 250 MW Hg-lamp (UV-technik). The lamp was placed 6 cm above the samples to provide a light intensity of 70 mW cm⁻² at 254 nm through the quartz glass as measured by an optical power meter (PM100A, Thorlabs). The coatings were deposited onto reference silicon wafers and silicon wafers with a layer of 5 μm thick Parylene C. The Parylene C was deposited using the Gorham process in a Parylene deposition chamber (PDS 2010, Specialty Coating Systems, Inc.). The dimer, di-chloro-di-para-xylylene (Specialty Coating Systems, Inc.), was vaporized at 135 °C and 35 mTorr and pyrolyzed into precursors at 690 °C. The precursors then passed into the deposition chamber at room temperature where polymerization occurred on the silicon wafer substrate to form a film of poly(monochloro-p-xylylene). The total deposition time was 2 h to achieve a thickness of 5 μm. Fourier transform infrared spectroscopy (Thermo-Nicolet iS10) was conducted on Parylene C samples before and after exposure to 20 min of UV light within the reactor under vacuum. There are no observed oxidation peaks after exposure, which verifies that the Parylene C remains unchanged.

To create the cross-linked anchoring layer, benzophenone was first flown in at a reactor base pressure of 30 mTorr and a substrate temperature of 40 °C. After 30 min, the flow of benzophenone was ceased and UV light irradiation was started. After a UV pretreatment time of 2 min, the cross-linker EGDA was flown in at 0.2 sccm, the TEC temperature was lowered to 32 °C, and the reactor pressure was set to 60 mTorr. After 8 min, the functional top layer was created by turning off the UV light and flowing the monomer (VP or oNBMA) along with the thermal initiator TPBO flow at 0.25 and 0.5 sccm, respectively. The reactor pressure was set to 120 mTorr and the filament array was heated to 240 °C. EGDA flow was gradually decreased over 2 min to prevent pressure deviations. The reaction then proceeded for an additional 18 min with the desired monomer (VP or oNBMA) flow and TPBO flow. To fabricate the grafted PVP coatings with no cross-linked anchoring layer (gPVP-noXL), the BP deposition and UV pretreatment time were conducted as described above. After the UV pretreatment time, VP was flown in at a rate of 0.25 sccm at a reactor pressure of 120 mTorr with the UV light on for 30 min. A quartz crystal microbalance (QCM) with a 6 MHz gold-plated crystal was used to estimate the adsorption of BP onto the Parylene C. BP was flown into the reactor and the measurements were performed at the reactor base pressure (30 mTorr) and the temperature of the QCM was kept at 40 °C. BP flow was ceased after 60 min.

To test the chemical composition and functionality of the cross-linked anchoring layer, the cross-linked anchoring layer was deposited onto four samples per deposition over three depositions for a total of 12 samples (E1-E12). These samples were then removed from the reactor and analyzed. To test the chemical composition and functionality of the grafted-PVP samples, the entire process (the cross-linked anchoring layer plus the functional top layer) was deposited onto four samples per deposition over three depositions for a total of 12 samples (V1-V12). Samples E1-E6 and V1-V6 were used for XPS analysis and samples E7-E12 and V7-V12 were used for contact angle goniometry. Three samples from each set (E1-3, V1-3, E7-9, and V7-9) were used for methanol sonication. Three samples from each set (E4-6, V4-6, E10-12, and V10-12) were used for the 1× phosphate buffered saline (PBS) soak test. For the oNBMA samples, two samples were prepared per deposition over two depositions (O1-O4). Samples O1 and O2 were used for XPS analysis and samples O3 and O4 were used for contact angle goniometry. All samples were washed with methanol and deionized (DI) water for 30 s each and dried with compressed air. These washed coatings were allowed to dry for at least 1 h under vacuum before analysis was performed.

The chemical composition of the samples was analyzed using x-ray photoelectron spectroscopy (Kratos Axis Ultra DLD) with a monochromatic Al Kα x-ray source. Survey spectra were taken from 0 to 900 eV with a step size of 1 eV a total of five times. Each sample (E1-6 and V1-6) was scanned in two different spots. Static, advancing, and receding contact angles were measured in triplicate using a goniometer (ramé-hart Model 290-FI) by depositing a 7 μl droplet of DI water onto different locations on each sample (E7-12 and V7-12). Advancing and receding contact angles were measured using the tilting-base method where the recorded angles where those right before the droplet rolled off or at 90° tilt. Captive bubble was performed using a DI water bath in a quartz glass container (ramé-hart) by dispensing a 7 μl air bubble using an automated dispenser. Each sample was measured in three different positions. For durability testing, samples E1-3, V1-3, E7-9, and V7-9 under went five 1 h rounds of sonication in methanol and samples E4-6, V4-6, E10-12, and V10-12 were incubated in 1× PBS solution at 37 °C for 30 days. After each test, the samples were removed, rinsed with DI water, and dried under vacuum for at least 1 h before analysis. Averages and standard deviations are reported from each sample set.

In order to obtain a high coverage of functional polymer on Parylene C using vapor phase deposition, we developed a stepwise process to create a layered polymer coating consisting of a photoinitiated cross-linked anchoring bottom layer and a functional top layer with the desired surface chemistry (Fig. 1). The cross-linked anchoring layer of grafted-PEGDA (gPEGDA) was formed by first adsorbing the photoinitiator onto the Parylene C surface. Benzophenone was chosen because it is a type-II photoinitiator that has previously been used in the vapor deposition process to successfully graft functional polymers onto substrates containing labile hydrogen atoms.^11,12 To obtain high surface coverage of our coatings, we had to ensure that enough BP saturated the surface. Martin et al. previously showed that a longer BP deposition time leads to a nonlinear increase in the thickness of the grafted polymer layer at a given substrate temperature.^12,13 We used a QCM to estimate that 30 min is required for benzophenone to saturate the surface at our TEC temperature (40 °C). After 30 min of adsorption, the BP flow was ceased and UV light was used to create benzophenone diradicals that abstract hydrogen from the Parylene C surface. The benzophenone and UV light were introduced sequentially to prevent the formation of excess BP radicals, which can lead to the termination of the radicals formed on the Parylene C surface.²⁰ 

After the UV pretreatment time, the cross-linker EGDA was flown into the reactor in the vapor phase in the presence of the UV light to initiate polymerization off the radicals on the surface of the Parylene C (Fig. 1). Using a cross-linker as the anchoring layer provides the ability to obtain high grafted coverage of functional polymers onto the Parylene C surface because the EGDA molecule has two vinyl groups, thereby increasing the ability to bind to the radicals on the surface of the Parylene C and also bind to other EGDA molecules that are already grafted to the surface. The monomer partial pressure was kept relatively low (60 mTorr) because previous work has shown that a high monomer concentration at the surface promotes homopolymerization of ungrafted polymer, diminishing the attachment to the surface.^11,12 In order to verify that the gPEGDA was grafted onto Parylene C, we conducted XPS and contact angle goniometry analysis on samples E1-E12. Our XPS analysis of the top 5 nm of the surface of the sample after a methanol and DI water wash matches that of a PEGDA reference film indicating successful attachment of gPEGDA onto Parylene C (Table I). Contact angles were also measured to verify the successful attachment of gPEGDA by the increase in the hydrophilicity of the sample. Uncoated Parylene C has a contact angle of approximately 90°,¹³ whereas the contact angle on the sample was 55° ± 2°, which is similar to the contact angle on the PEGDA reference film (53° ± 2°). This evidence, coupled with the XPS survey scan data, shows that our gPEGDA coating completely covers the Parylene C substrate.

To fabricate samples with the functional top layer (V1-V12), the gPEGDA layer was deposited as described above. After the deposition of the gPEGDA layer, the initiator and monomer were introduced into the chamber immediately, the reactor pressure was set, and the filament was turned on to deposit the grafted-poly(vinyl pyrrolidone) (gPVP) top layer using conventional iCVD (Fig. 1). The EGDA flow was gradually turned off while maintaining the reactor pressure, which allows for continuity from the piCVD process to the conventional iCVD process. TBPO radicals were used to attach the gPVP layer to the gPEGDA layer instead of BP radicals because TBPO radicals are not reactive enough to abstract hydrogen and therefore there are no side reactions during the iCVD process and the functionality of the monomer is kept intact. TBPO radicals initiate polymerization from the unreacted vinyl groups of the gPEGDA layer, leading to the formation of the gPVP layer. After the EGDA flow was ceased, the deposition continued with only VP flow to ensure maximum coverage. Since this top layer was deposited using conventional iCVD, coatings grew significantly thicker than the gPEGDA layer but only a small portion of the thickness was grafted, while the rest was washed away. XPS analysis shows that the elemental composition of the gPVP layer after a methanol and DI water wash matches that of our iCVD PVP reference coatings (Table I), demonstrating the successful attachment of the gPVP layer. In order to demonstrate the necessity of the cross-linked anchoring layer, we deposited gPVP via piCVD with no cross-linker layer (gPVP-no XL). In this case, Parylene C is still exposed because of the finite grafting sites available.^11–13,20 The XPS survey scan of gPVP-noXL shows both chlorine (6%) and nitrogen (2%) are present, indicating that PVP is present with low graft density emphasizing the importance of the cross-linked anchoring layer.

Contact angle hysteresis was used to obtain a true measure of hydrophilicity because it has been previously shown that grafted polymer systems tend to rearrange, depending on their dry or wet conditions²⁹ and the functionality of the polymer.^30,31 The static contact angles of 41° ± 5° of our gPVP coatings do not match PVP reference static contact angles of 11° ± 1° (Table II). However, at 90° tilt, the receding contact angle of our gPVP coatings approached that of the PVP reference static contact angle, demonstrating that some rearrangement is occurring.³¹ PVP is extremely hydrophilic; therefore, it tends to orient itself away from the dry environment (air) and when a water droplet is placed onto the surface, some molecules orient into the water droplet. The captive bubble method was used to elucidate the wet environment hydrophilicity of the coatings by dispensing an air droplet onto the coating in a DI water bath. This data is also useful because implantable Parylene C-based biomedical devices are constantly exposed to an aqueous environment.^{2–4,8,10,32,33} The contact angle is immediately lowered after immersion from that of the static sessile drop contact angle, from 41° ± 5° to 27° ± 2°. The coatings were immersed for 60 min to allow time for rearrangement of the gPVP molecules, ultimately reaching an angle of 19° ± 3° (Table II). In comparison, uncoated Parylene C shows no decrease in contact angle after immersion because there is no other favorable orientation. The gPEGDA coatings decrease by only 10° ± 1° after 60 min of immersion showing that some rearrangement is occurring. Figure 2 depicts the increase in hydrophilicity in the coatings after the various stages in the process.

In order to show the durability of our grafting process, the coatings were tested by sonicating them for 5 h in methanol to show that they can withstand physical forces. The durability of the coatings is evident since after 5 h of sonication in methanol, little to no functionality is lost in the gPVP coatings (Table II) as evidenced by the negligible change in contact angle. The elemental composition from the XPS spectra also confirms that our coatings are still on the surface of the Parylene C substrate evidenced by the lack of a chlorine signal showing that our coatings still cover the Parylene C (Table III). Since Parylene C is used for biomedical implants,^2,3,5,6,32 the coatings were also soaked in 1× PBS solution at 37 °C for 30 days to show that the functionality can be retained, long-term in in vitro environments. After 30 days in the 1× PBS soak at 37 °C, the coatings showed little to no degradation in functionality, as evidenced by the lack of change in contact angle (Table II) and elemental composition (Table III).

To show the generality of our procedure, the photoresponsive polymer PoNBMA was also deposited as the functional top layer. This photoresponsive polymer can be converted from hydrophobic to hydrophilic by exposure to UV light which cleaves some of the nitrobenzyl groups and thereby converts them to methacrylic acid groups.²⁵ To create the grafted-PoNBMA (gPoNMBA) layer, we first deposited the cross-linked anchoring layer onto Parylene C using the photoinitiator and then grafted PoNBMA using the thermal initiator as described above. The grafted samples were washed with methanol and DI water to remove ungrafted homopolymer. The static sessile drop angle was measured to be 77° ± 3°, which changed to 56° ± 3° after 1 h of UV light exposure (Fig. 3), confirming that the photoresponsiveness is retained. In comparison, the reference of ungrafted PoNMBA coatings on Parylene C had a static sessile drop angle of 102° ± 4°, which decreased to 72° ± 3° after 1 h of UV light irradiation. Since the gPoNBMA coatings can rearrange, captive bubble measurements were taken to see the extent of rearrangement of the functional groups in an aqueous environment before and after UV exposure. We found that the captive bubble contact angles before UV light exposure decreased from 70° ± 3° to 20° ± 3° after exposure (Fig. 3), showing the gPoNBMA coatings after exposure showed significant rearrangement when placed into an aqueous environment.

We have demonstrated a novel vapor deposition process where we used photoinitiated CVD, followed by conventional iCVD to graft functional polymers onto Parylene C. We used XPS and contact angle goniometry to show that our grafted coatings retained the desired functionality and fully covered the Parylene C substrate. Durability testing showed that our grafted coatings retained their functionality and elemental composition, showing promising long-term efficacy as an implantable material. Since our process uses standard iCVD parameters and precursors, this work could be used to deposit any cross-linker and functional polymer combination that has been deposited using iCVD. Therefore, the generality of this process gives access to a wide library of iCVD precursors that could be used. Additionally, this process can be extended to any substrate with a labile hydrogen atom. Also, given the conformal nature of the iCVD process, our process can be used on substrates with a wide array of geometries.

This work was supported by the National Science Foundation Award No. 1332394. M.D.L. is supported by a National Science Foundation Graduate Research Fellowship under Grant No. DGE-1418060.