Fluorescent polymeric nanocomposite films generated by surface-mediated photoinitiation of polymerization
- First Online:
- Cite this article as:
- Avens, H.J., Chang, E.L., May, A.M. et al. J Nanopart Res (2011) 13: 331. doi:10.1007/s11051-010-0034-z
- 205 Views
Incorporation of nanoparticles (NPs) into polymer films represents a valuable strategy for achieving a variety of desirable physical, optical, mechanical, and electrical attributes. Here, we describe and characterize the creation of highly fluorescent polymer films by entrapment of fluorescent NPs into polymer matrices through surface-mediated eosin photoinitiation reactions. Performing surface-mediated polymerizations with NPs combines the benefits of a covalently anchored film with the unique material properties afforded by NPs. The effects of monomer type, crosslinker content, NP size, and NP surface chemistry were investigated to determine their impact on the relative amount of NPs entrapped in the surface-bound films. The density of entrapped NPs was increased up to 6-fold by decreasing the NP diameter. Increasing the crosslinking agent concentration enabled a greater than 2-fold increase in the amount of NPs entrapped. Additionally, the monomer chemistry played a significant role as poly(ethylene glycol) diacrylate (PEGDA)-based monomer formulations entrapped a 10-fold higher density of carboxy-functionalized NPs than did acrylamide/bisacrylamide formulations, though the latter formulations ultimately immobilized more fluorophores by generating thicker films. In the context of a polymerization-based microarray biodetection platform, these findings enabled tailoring of the monomer and NP selection to yield a 200-fold improvement in sensitivity from 31 (±1) to 0.16 (±0.01) biotinylated target molecules per square micron. Similarly, in polymerization-based cell staining applications, appropriate monomer and NP selection enabled facile visualization of microscale, sub-cellular features. Careful consideration of monomer and NP selection is critical to achieve the desired properties in applications that employ surface-mediated polymerization to entrap NPs.
KeywordsFluorescent polymersPolymeric nanocompositesPhotopolymerizationSurface-mediated polymerizationFluorescent immunocytochemistryProtein microarrays
Polymeric nanocomposites are of significant interest, heralded for achieving further enhancements in material properties as compared to microcomposite approaches (Allegra et al. 2008). Incorporation of nanomaterials in polymeric matrices has been employed in a variety of applications including generation of materials with enhanced optical, magnetic, electrical, thermal, and mechanical properties (Balazs et al. 2006; Dhibar et al. 2009; Durán et al. 2008; Luo et al. 2009; Verma et al. 2009). Additionally, nanocomposite coatings have been developed that achieve various desirable surface modifications including superhydrophobicity (Xu et al. 2009), corrosion resistance (Olad and Rashidzadeh 2008), reduced gloss (Balan et al. 2008), and enhanced mechanical strength (Zhu et al. 2007).
Recently, the use of fluorescent nanocomposite films has been reported for sensitive and photostable biodetection in microarrays and immunocytochemical staining (Avens and Bowman 2010; Avens et al., submitted to Journal of Histochemistry and Cytochemistry; Hansen et al. 2008a). This approach, coined “fluorescent polymerization-based amplification” (FPBA), utilizes biological probes selectively labeled with eosin photoinitiators. After the probe has bound its target on the microarray or within the cell, a solution of monomer, coinitiator, and fluorescent nanoparticles is applied and the test surface is exposed to light of wavelengths greater than 480 nm to initiate polymerization. This approach results in the formation of crosslinked polymer films specifically in regions of the microarray or cell where the probe has bound its target. As the polymer film forms, it entraps the fluorescent nanoparticles, rendering the film highly fluorescent. In this manner, the presence and localization of a biological target is evidenced by a highly fluorescent nanoscale polymer film. During the development of FPBA, other strategies for generating fluorescent films were considered, such as using monomers covalently attached to organic fluorophores; however, this approach was unsuitable because the organic fluorophores themselves non-specifically initiated polymerization and/or underwent extensive photobleaching (Avens and Bowman 2010). To overcome these challenges, it was necessary to use polystyrene nanoparticles with fluorophores embedded in the interior which enabled facile formation of fluorescent films with fewer problematic side reactions between the monomer and the fluorophores. FPBA has been demonstrated to yield approximately 100-fold brighter signals and a 100-fold improvement in detection sensitivity compared to using a molecular fluor-labeled probe (Avens and Bowman 2010; Avens HJ et al., submitted to Journal of Histochemistry and Cytochemistry). Additionally, in cell staining applications, FPBA compared favorably to the highly sensitive tyramide signal amplification (TSA) approach which relies on peroxidase enzymes. Moreover, while both TSA and FPBA generated similarly intense fluorescent signals, FPBA is advantageous in that it is not affected by endogenous cellular peroxidase enzymes which cause non-specific TSA staining (Avens HJ et al., submitted to Journal of Histochemistry and Cytochemistry).
Though a variety of surface-mediated initiation strategies are suitable for generation of NP-polymer composite films, the method presented here employs surface-immobilized eosin as a type II photoinitiator for polymerization. Upon photoexcitation with visible light, eosin undergoes energy, charge, and electron transfer with a coinitiator, commonly a tertiary amine such as N-methyldiethanolamine (MDEA), to yield a relatively stable eosin radical and an MDEA radical capable of initiating polymerization (Avens and Bowman 2009). Despite the fact that the initiating MDEA radicals are not tethered to the surface, the reaction is still considered surface mediated because the initiating radicals are generated only at the surface. Covalent attachment of the films to the surface is believed to occur through a variety of chain transfer reactions and other processes including termination of the polymer chains with the eosin radical which is reported to be inactive for initiation, but capable of reacting by termination (Kizilel et al. 2004). Eosin has been used for surface-mediated polymerization for an assortment of purposes including modification of nanoparticle surfaces to enhance dispersion (Satoh et al. 2005), islet cell encapsulation (Cruise et al. 1998), creation of protective surface coatings for arterial walls (An and Hubbell 2000), and generation of macroscopically visible and fluorescent films for sensitive signal amplification of biodetection (Hansen et al. 2008a; Avens and Bowman 2010). The fact that eosin uses visible light which is less damaging to biological systems than UV light, and the observation that eosin is less sensitive to oxygen inhibition than other photoinitiators (Avens and Bowman 2009) makes eosin an excellent choice for numerous surface modification applications. Importantly, photoinitiation, as opposed to other initiation strategies, readily enables precise spatial and temporal control of the polymerization process, factors which are often critical for achieving the desired material properties. (Bowman and Kloxin 2008). Though photoinitiation is the focus of this work, the NP incorporation results presented here are expected to be generalizable to other surface-mediated polymerization processes including other polymerization-based detection strategies (He et al. 2008; Lou et al. 2005).
For FPBA and other applications that employ surface-mediated polymerization with NPs, it is necessary to optimize the NP and monomer attributes to achieve the required polymer properties. To more aptly employ this type of nanocomposite film formation for biodetection and other applications, a detailed study is presented here investigating the effects of NP size, NP surface functionality, monomer type, and crosslinking agent content on the fluorescence and thickness of the polymer films. For FPBA, it is generally desired to incorporate as high a density of NPs as possible to maximize the subsequent fluorescent signal emanating from the surface. It is hypothesized that chemical and mechanical interactions occurring at the polymer–monomer–NP interface during the polymerization strongly influence the successful entrapment of NPs into the growing polymer matrix. To identify trends that are relevant across a broad range of initiation rates, surface-mediated polymerizations were carried out using a microarray format in which each row of spots comprises a different eosin surface density, such that a wide range of eosin photoinitiator densities were evaluated simultaneously. Reactions were performed with both an acrylamide-based monomer formulation and a poly(ethylene glycol) (PEG) acrylate-based formulation where the acrylamide polymerization is known to occur more rapidly and the PEG-based formulation is more hydrophilic. Trends that are common between these two monomer formulations are expected to be pertinent to other monomer types as well, while differences are correlated to the differences in the chemistry and reaction behavior of the formulations.
The significance of the trends presented here was investigated in the context of two specific applications. First, it was shown that through selection of the appropriate monomer type, crosslinker content and NP size one is able to significantly increase the absolute amount of fluorescent NPs immobilized on the surface, yielding a 200-fold improvement in detection sensitivity in a polymerization-based microarray biodetection platform compared to using sub-optimal reaction conditions. Second, by selecting parameters to optimize the NP density entrapped in the film, rather than just the total NP amount immobilized to the surface, it was possible to create highly fluorescent, yet thin and spatially restricted films suitable for visualization of microscale sub-cellular features in polymerization-based fluorescent immunostaining of cells. The principles outlined here are expected to be useful for tailoring monomer and NP attributes to achieve desired properties in additional applications in which NPs are being entrapped in surface-mediated polymerizations, including in the context of different polymer matrices and different types of NPs.
Characteristics of NPs used for these studies
Nominal diameter (nm)
Actual diameter (nm)
24 ± 4
43 ± 6
100 ± 6
210 ± 10
190 ± 11
24 ± 3
36 ± 5
210 ± 10
24 ± 4
Compare photoluminescence of NPs in solution and in polymer
To verify that encapsulation in polymer films does not dramatically alter NP photoluminescence, an Ocean Optics USB4000-FL detector was used to measure emission spectra of free 20 nm yellow/green NPs in water (0.05 wt% NPs) as well as NPs that had been entrapped in a polyacrylamide film analogous to the ones formed on the surfaces. The excitation source was a 370 nm CrystaLaser. The samples were prepared in 1 cm wide UV/VIS cuvettes.
Preparation of eosin dilution chips
Streptavidin–eosin was printed onto epoxy-functionalized glass slides using a VersArray Chip WriterTM Pro (Bio-Rad) with approximately 45% humidity and a solid pin yielding spots of approximately 500 μm diameter. 9 × 5 arrays of spots were prepared containing five replicate spots of 8 decreasing eosin concentrations, and a 9th row with no eosin that served as a negative control suitable for evaluating non-specific polymerization. The highest print concentration was 0.27 mg/mL SA–eosin (12 μM eosin, 1.3 eosin per streptavidin) in a final concentration of 1× printing buffer. Lower print concentrations were prepared by serial 1:3 dilutions into 1× print buffer with unmodified streptavidin such that all spots were printed with a total protein concentration of 0.27 mg/mL. Unbound SA–eosin was removed by three 15 min rinses in water with rapid rocking. Since eosin is a fluorophore, which also serves as the initiator, surface density was characterized with an Agilent Technologies Microarray Scanner. A Cy3 Scanner Calibration slide was used to convert fluorescence readings into surface density of eosin. The slides were then polymerized as described subsequently.
Preparation of biotin dilution chips
Microarray printing as described for SA–eosin was likewise performed to create slides with a dilution series of biotin-α-goat IgG in a 9 × 5 array containing five replicate spots of 8 decreasing antibody concentrations, and a 9th row with no antibody that served as a negative control suitable for evaluating non-specific polymerization. The highest print concentration was 85 μg/mL biotin-α-goat IgG in 1× printing buffer and 0.5 mg/mL BSA. Lower print concentrations were prepared by serial 1:3 dilutions into 1× print buffer with 0.5 mg/mL BSA, and the 9th row contained only 0.5 mg/mL BSA. The slides were stored unprocessed until they were used for binding and detection reactions. The surface concentration of bound antibodies was estimated as described previously (Avens and Bowman 2010).
Binding reactions with the biotin dilution chips
The biotin-α-goat IgG dilution chips were rinsed with water to remove unbound antibody, then the slides were incubated with 10 mg/mL BSA in 1× PBS for 1 h, followed by three 2 min rinses in PBS with rapid rocking. The dilution chips were then reacted 1 h with 1 μg/mL SA–eosin, using Chip Clips from Whatman to form wells on the chips and gentle rocking was used, followed by three 2 min rinses in PBS with rapid rocking and one rinse in water. The slides were then polymerized as described subsequently.
Cell staining procedures
Cells in micro-chamber slides were rinsed with 1× PBS then fixed with 4% paraformaldehyde for 15 min. The slides were then stored in 1× PBS and refrigerated until staining. At the time of staining, slides were rinsed 3 min with 1× PBS, permeabilized with 0.1% Triton-X in PBS for 5 min, then rinsed three times with PBS. Next, the slides were blocked for 1 h with 2% horse serum and 0.1% BSA in PBS, and then rinsed three times with 0.1% BSA in PBS (PBSA). Anti-NPC IgG at a dilution of 1:1000 in PBSA was applied for 1 h, followed by three rinses with PBSA. Next, biotin-anti-mouse IgG was applied for 1 h at a dilution of 1:400 in PBSA, followed by three rinses with PBSA. The slides were then contacted with 10 μg/mL SA–eosin for 30 min, followed by three rinses with PBS and one rinse with water. Polymerization reactions were then carried out as described in the following section. After polymerization, the nuclei of the cells were stained 2 min with 1 μg/mL DAPI in water.
The microarray or cell slide surfaces were contacted with the desired monomer solution. For 5 min prior to and throughout the entire light exposure, the slides were placed in a plastic bag with argon flow to reduce oxygen in the atmosphere. The light source was an Acticure (Exfo) high pressure mercury lamp with an in-house internal bandpass filter (350–650 nm) and an external 490 nm longpass filter (Edmund Optics) positioned at the end of a light guide and a collimating lens. The light intensity was measured using an International Light radiometer. After polymerization, unreacted monomer was removed with three-5 min water rinses, followed by air drying. To obtain sufficient replicates in the microarray experiments, each condition investigated was polymerized on at least 2–3 arrays in 2–3 separate polymerization sessions.
Film thickness measurements
Polymer film thicknesses were measured using a Dektak 6 M surface profilometer with 12.5 μm diameter tip and a stylus force of 1 mg.
Fluorescent imaging of films
Cells stained by fluorescent polymerization-based signal amplification were imaged by confocal scanning laser microscopy (Zeiss LSM 510 instrument) with a 40× oil objective. Yellow/green NP fluorescence on the microarray surfaces was measured using a Leica MZ FLIII stereomicroscope (Leica Microsystems, Wetzlar, Germany) with the blue filter set. Exposure times in the range of 5 to 60 s were used, depending on the intensity of the fluorescent signal. A Cy3 Scanner Calibration slide was imaged at each exposure time to identify appropriate scaling factors for the different exposure times, such that all data could be plotted using the same arbitrary fluorescence scale. Crimson and dark red NP fluorescence on the microarray surfaces were measured using the red channel of an Agilent Technologies Microarray Scanner. Significant positive signals are those that have a signal to noise (S/N) greater than 3, where [S/N = (signal − background signal)/(standard deviation of the background signal)]. The film fluorescence values have been normalized to account for the differences in relative quantum yield between NPs of different sizes or different surface chemistries, utilizing the relative quantum yields provided in Table 1. Additionally, for NPs of the same color, but different sizes or surface functionalities, the film fluorescence values were normalized to account for the slight differences in the NP fluorophore content per mass, as determined by absorbance measurements (Table 1). In this way, the reported fluorescence intensities are more indicative of the mass of NPs that has been immobilized in the films, rather than simply variations in the manufacturing of these NPs.
In cases where a “gain value” is reported, linear regression with a least squares parameter estimation was performed to best fit a line of the form y = mx (intercept set to 0) to the data. The slope “m” is termed “gain” and the 95% confidence interval on the slope is indicated.
Mesh size determinations
The observation that PEGA with no added crosslinker readily forms a crosslinked gel highlights the fact that a significant amount of chain transfer occurs in the PEGA and PEGDA systems, which creates ambiguity concerning the chemical composition of the chains between crosslinks. Thus, the average mesh sizes for the PEG-based gels were calculated by two distinct methods to provide reasonable bounds on the probable average mesh sizes: (1) n = 2Mc/Mr, where Mr is the molecular weight of the repeating monomer unit assuming no chain transfer; and (2) assuming extensive chain transfer, n = 3Mc/Mr, where Mr is the molecular weight of the PEG repeat unit. The value of three in the second method arises from the fact that the PEG repeat unit (COO) contains three bonds. In the case of polyacrylamide, n = 2Mc/Mr, where Mr is 71 g/mol.
Results and discussion
Compare the photoluminescence of NPs in solution and in the polymer film
Effect of NP size
In order to ascertain the impact of NP diameter on the ability of fluorescent NPs to become encapsulated in films generated from surface-mediated photoinitiation, polymerizations were performed with monomer solutions containing fluorescent NPs of a variety of distinct diameters ranging from 20 to 200 nm. Additionally, a microarray format was employed such that a wide range of eosin photoinitiator surface densities were evaluated simultaneously, enabling investigation of the effect of NP size at various initiator surface densities. Using various eosin initiator surface densities ensures that the observed trends for NP size are generalizable to varying initiation rates. Finally, two different monomer formulations were employed, an acrylamide-based formulation and a PEGDA-based formulation.
The enhanced incorporation of smaller NPs likely arises from the interfacial nature of the surface-initiated polymerization and the evolving crosslinked network that is forming. Though the polymerization is initiated from a uniform surface, heterogeneities and differential extension of the polymer film into the bulk monomer will undoubtedly arise, particularly in these crosslinked materials that are notorious for microgel formation and heterogeneity (Hutchison and Anseth 2001; Kloosterboer 1988). The smaller the NP is, the more likely it is to become entrapped in this evolving structure.
Effect of NP surface chemistry
Effect of crosslinker concentration
Gain values measured for fluorescent films formed from varying concentrations of PEGA and PEGDA
Overall gain (NP fluorescence/(eosin/μm2))
Fluorescence gain (NP Fluorescence/nm)
Thickness gain (nm/(eosin/μm2))
Estimated molecular weight between crosslinks (Mc) and mesh size in polyacrylamide gels prepared with varying amounts of bisacrylamide crosslinker
Bisacrylamide content (wt%)
Mesh size (nm)
5.5 (±0.1) × 103
2.9 (±0.1) × 103
Estimated molecular weight between crosslinks (Mc) and mesh size in PEG gels prepared with varying amounts of PEGDA crosslinker
Mesh size range (nm)a
1.2 (±0.1) × 104
7.9 (±0.7) × 103
2.2 (±0.1) × 103
2.4 × 102–1.5 × 103 b
Effect of monomer type
Impact on sensitivity in a biodetection application
Fluorescent polymerization-based amplification (FPBA) has been demonstrated previously to be a sensitive array-based biodetection platform for signal amplification that enables a 100-fold improvement in sensitivity as compared to traditional fluorescence detection methods (Avens and Bowman 2010; Hansen et al. 2008b). FPBA relies on surface-mediated initiation of polymerization to entrap fluorescent NPs within PEGDA or polyacrylamide polymer films. Specifically, the assay employs DNA or protein probes that are coupled to eosin photoinitiators, thereby immobilizing eosin photoinitiators on the surface wherever the probe has bound its target. In the last step of the assay, the surface is contacted with monomer, MDEA coinitiator, and fluorescent NPs and exposed to visible light, generating a highly fluorescent film specifically in regions of the surface where target is present. To achieve the highest sensitivity for fluorescence detection, it is necessary to maximize the mass of fluorescent NPs immobilized per surface area in response to the eosin-initiated polymerization.
The FPBA array-based biodetection platform was used to assess the combined effect of selecting monomer and NP attributes that are more or less favorable for immobilization of fluorescent NPs to the surface. As a model system for biodetection, microarray slides were used that contain a range of biotin-labeled antibody surface densities. These slides were contacted with eosin-labeled streptavidin which specifically binds to biotin, thereby immobilizing eosin photoinitiators to the surface. The biotin–streptavidin system is a commonly used label-probe combination employed in many detection assays. Two different monomer formulations were compared: (1) an acrylamide monomer formulation containing 2 wt% crosslinker and 20 nm yellow/green NPs; and, (2) a formulation containing 1 wt% PEGDA, 21 wt% PEGA and 100 nm yellow/green NPs. Based on the aforementioned NP-incorporation studies, the first formulation is expected to generate significantly stronger amplified fluorescence signals and allow more sensitive detection due to the smaller NP diameter, the higher crosslinker content, and the ability of acrylamide to grow thicker films.
Impact on the ability to obtain localized staining of a small cellular feature
These two biodetection examples demonstrate the importance of understanding the NP incorporation behavior and their impact on polymerization as each different application places distinct requirements on the polymerization and detection process. Here, the optimal NP–monomer combination is dramatically different depending on whether one requires maximum sensitivity or whether one desires maximum spatial resolution.
The results presented here highlight the fact that careful consideration of monomer and NP selection is essential to achieving the desired properties in applications that employ surface-mediated polymerizations to entrap NPs. In choosing a monomer formulation, higher crosslinker content tends to be more efficient at entrapping NPs, while consideration of the interactions between the polymer and the NP surface is also important. Choice of suitable NP attributes requires consideration of both NP size and surface chemistry, with smaller sizes tending to favor entrapment in polymer matrices. Although this study employed relatively low NP concentrations (0.05 wt%) such that light attenuation of the NPs would not excessively hinder photoinitiation, it is expected that the trends outlined here will be similar and perhaps even more pronounced with increased NP loading. It is additionally anticipated that these findings will be helpful in designing approaches for surface-mediated generation of polymeric nanocomposites composed of other polymers, polymerization types, and NP types.
This material is based upon work supported by National Institutes of Health R21 CA 127884, a National Science Foundation Graduate Research Fellowship to HJA, and NSF Research Experience for Undergraduates funding to KRV. Also, this work has been supported by the State of Colorado and the University of Colorado Technology Transfer Office.