- Email: [email protected]
Immobilisation of different surface-modified silica nanoparticles on polymer surfaces via melt processing
Colloids and Surfaces A xxx (xxxx) xxx–xxx
Contents lists available at ScienceDirect
Colloids and Surfaces A journal homepage: www.elsevier.com/locate/colsurfa
Immobilisation of different surface-modified silica nanoparticles on polymer surfaces via melt processing ⁎
Jürgen Nagela, , Felix Kroschwaldb, Cornelia Bellmanna, Simona Schwarza, Andreas Jankea, Gert Heinricha,c a b c
Leibniz-Institut für Polymerforschung Dresden e.V., 01069 Dresden, Hohe Straße 6, Germany Gerflor Mipolam GmbH, Mülheimer Str. 27, 53840 Troisdorf, Germany Technische Universität Dresden, Institute of Materials Science, 01069 Dresden, Hohe Straße 6, Germany
A R T I C L E I N F O
A B S T R A C T
Keywords: Nanoparticles Nanotechnology Surfaces Injection moulding
The immobilisation of silica nanoparticles on the surface of a flowing polymer melt is studied during injection moulding. We use silica nanoparticles of identical size of about 200 nm, but with different surface functionalities: plain unmodified particles with silanole groups as well as amino- and hydrocarbon modified particles. The particles were brought in contact with a polycarbonate melt at a temperature of 300 °C. In our investigation, we show that particles with polar surfaces are more embedded than those with non-polar surfaces. Immobilised polar particles exhibited also a higher adhesion, tested with an adhesive tape, whereas the nonpolar particles can be easily removed by peeling off the adhesive tape. This paper reveals that surface properties have a larger influence on embedding of nanoparticles than thermal conditions due to cooling of the melt.
1. Introduction Nanoparticles and nanostructured materials are the very promising materials for future demands. They are already today applied in many fields of applications. Many applications, e.g. as sensor [1], catalyst [2], require the immobilisation of nanoparticles on solid surfaces. This can be realised very efficiently by adsorption using the layer-by-layer technique [3–6] or by dispersion drying [7]. However, since the nanoparticles are only bound by weak chemical or physical interactions, the mechanical stability of those layers is limited. A higher stability could be approached by embedding in a polymer melt. Such an embedding of gold nanoparticles was realised in the group of Composto by moulding in a hot press [8]. Here, the particles were embedded, partly completely, in the polymer melt. The authors explained this by the high surface energy of the particles and capillary forces that resulted in extensive wetting by the melt. The time of contact with the melt at high temperature was arbitrary in these experiments, so that conditions for the adjustment of a thermodynamic equilibrium prevailed. However, fabrication of polymer parts by hot pressing is not very efficient. Our previous work dealt with embedding of small gold nanoparticles using injection moulding [9,10]. Embedding occurred also here. However, some parts of the surface area remained uncovered by the melt and were still accessible for small molecules from solution.
⁎
This was used for catalytic reactions. Wetting of the particle surface by melt occurred despite the high cooling rate once the hot melt hit the cool surface. However, this could be a result of the small particle diameter of about 20 nm. In this paper, we used nanoparticles with equal size and different surface functionalisation. Pure silica particles are polar and exhibit silanole groups on their surface. Amino-modified particles are also polar and have, in addition, groups that can reactively bind onto polycarbonate [11,12]. Hydrocarbon-modified particles are typically non-polar and exhibit a low surface energy. The particle diameters were about 200 nm and, thus, larger than those of the discussed gold nanoparticles. Since the nanoparticles affect the temperature diffusion from melt to mould, we used particles with uniform size to realise equal cooling conditions. The aim of this paper is, to find out, whether there is an influence of the surface modification on wetting and embedding of particles in the polymer melt under the special conditions of injection moulding, or if the rapid cooling dominates the structure formation. The question is to be answered, whether the embedding of nanoparticles into the surface of a thermoplastic part can be controlled by surface functionalisation. 2. Material and methods Uniform silica nanoparticles with a diameter of about 200 nm were
Corresponding author. E-mail address: [email protected] (J. Nagel).
http://dx.doi.org/10.1016/j.colsurfa.2017.05.016 Received 27 February 2017; Received in revised form 5 May 2017; Accepted 8 May 2017 0927-7757/ © 2017 Elsevier B.V. All rights reserved.
Please cite this article as: Nagel, J., Colloids and Surfaces A (2017), http://dx.doi.org/10.1016/j.colsurfa.2017.05.016
Colloids and Surfaces A xxx (xxxx) xxx–xxx
J. Nagel et al.
obtained from Geltech (Orlando, USA). Octadecyltrimethoxysilane (OTMS), N-(3-Trimethoxysilylpropyl)-diethylenetriamine (DETA), sodium poly(ethylene sulphonate) (PES) and Poly(diallyldimethylammoniumchlorid) (PDADMAC) were ordered at Sigma-Aldrich, Germany. Polyethyleneimine (hyperbranched, Mw 750 kg mol−1, Sigma-Aldrich, Germany, PEI) was ordered as 50 wt.% aqueous solution. Polycarbonate (Makrolon LED 2245, Bayer Materials Science GmbH, now Covestro AG, Germany, PC) was used for moulding experiments. The plain silica nanoparticles were used as received. They exhibited a low porosity with a surface area of 4.5 m2 g−1, as measured by BET using nitrogen gas. This value equalled that given from the manufacturer. Surface modification with amino groups was carried out according to a protocol taken from literature in a 500 ml flask under nitrogen purging [13]. 0.15 g acetic acid, 250 ml water and 2.5 ml DETA were mixed. 0.6 g silica nanoparticles were added and stirred for 6 h at room temperature. After that time, they were centrifuged, and the supernatant solution was changed against water. This procedure was repeated three times. The yield was about 84%. Surface modification with OTMS was carried out in a 100 ml flask with nitrogen purging. 500 mg silica nanoparticles were dispersed in 10 ml toluene in an ultrasonic bath. Then a solution of 3 mmol OTMS in 10 ml toluene was added dropwise via a septum. The solution was stirred first 1 h at 80 °C and then 2 h under reflux. Finally, the mixture was centrifuged three times with exchange of the supernatant solution against toluene. The product was dried in vacuum. The yield of the hydrocarbon-modified nanoparticles was about 62%. A Zetasizer Nano ZS (Malvern Instruments GmbH, UK) was used for measurements of zeta potential using electrophoresis. Polyelectrolyte titration was carried out at pH 5.8 using a 702 SM Titrino (Metrohm GmbH & Co. KG, Germany) and with PDADMAC and PES as titrant, respectively. Glass microscope slides with the dimensions 1 inch x 3 inch were used as substrates. They were cleaned in NoChromix solution and rinsed with water prior to use. Layer of amino-modified nanoparticles were prepared using cleaned plain glass slides. The adsorption was carried out using a stirred dispersion with a nanoparticle concentration of 2 g l−1 at pH 6. After an adsorption time of 30 min, the glass slides were rinsed for 15 min with water. For the adsorption of plain silica particles, PEI was adsorbed in advance onto the glass slides. The solution was prepared by dissolving 10 g of the concentrated PEI solution and 29 g NaCl in 400 ml water. Then, the pH was adjusted to 8 by adding HCl, and the volume was adjusted to 500 ml with water. The glass slides were rinsed in this solution for 15 min under stirring and then rinsed in water for 15 min. The adsorption of plain silica nanoparticles was carried out from a colloidal solution with a concentration of 1.2 g l−1 at pH 6. Hydrocarbon-modified nanoparticles were dispersed in toluene at a concentration of 52 mg ml−1. Layers were prepared by dispersion drying of a film of 32 μl of the suspension over the substrate. Substrates with a dense layer of nanoparticles on the surface were translucent. The substrates covered with nanoparticle layers were mounted in the mould cavity of a plunger injection moulding machine (construction made in-house). PC was dried before use at 110 °C for 5 h in vacuum. PC melt was injected at a melt temperature of 300 °C and with a mould temperature of 80 °C using typical processing parameters of PC (injection rate of 7.3 mm3 s−1, injection pressure of 100 bar, cooling time of 26 s). The specimens produced were round plates with a diameter of 50 mm and a thickness of about 2 mm. The substrates were removed from the formed part. They were clear after the moulding, but the surface of the parts fabricated were translucent, if the substrate were translucent before moulding, pointing to a successful transfer. In some cases, an adhesive tape with an adhesive strength of 5.4 N per 18 mm was mounted on the part surface and removed to characterise the adhesive strength of the particle immobilisation. Scanning electron images were taken with an Ultra Plus Gemini (Carl Zeiss SMT AG, Germany). The acceleration voltage was varied
Fig. 1. Zeta potentials of plain and amino modified silica nanoparticles as function of pH. Measured in 10−3 M KCl.
between 1 and 20 kV. The working distance was 3 mm 3 nm platinum were sputtered on the sample surfaces in advance using a Leica EM SCD 050 (Leica Mikrosysteme Vertriebs GmbH, Germany). Atomic force microscopy (AFM) images were taken with a Dimension 3100 Nanoscope (Veeco Instruments Inc., USA). The measurements were carried out in tapping mode with silicon SPM sensors (Budget Sensors, Bulgaria) having a spring constant of 4 N m−1 and a resonance frequency of 75 kHz. The radius of the tip was smaller than 10 nm. The dynamic contact angles were measured using an OCA35 XL (DataPhysics Instruments GmbH, Germany). An average was calculated from at least four measurements per sample. 3. Results and discussion 3.1. Characterisation of the particles Fig. 1 shows that the isoelectric point (IEP) of the plain silica nanoparticles was around a pH value of 2.5. The zeta potential decreased on increasing pH until a value of −58 mV was approached at pH 10. This is the typical behaviour for silica surface [14]. On the other side, the amino-modified nanoparticles showed an IEP of about 9.7, which is typical for basic functionalities. The plateau value of 50 mV is characteristic for complete protonated functional groups on the particle surface. This result is typical for amino-modified surfaces [15] and pointed to a stable binding of the amino silane onto the silica nanoparticle. The dispersions were titrated to zero potential with PES or PDADMAC, respectively, and the results are shown in Table 1. The consumption of ions was related to the total surface area, calculated from the particle diameter and the total amount of particles in the dispersion. The surface charge density of plain silica nanoparticles was −5.4∙10−2C m‐2. The alternation by treatment with the amino silane to 8.1∙10−2 pointed to binding of the amino silane onto the silica surface. The hydrocarbon-treated silica nanoparticles could not be dispersed in Table 1 Surface charge densities from polyelectrolyte titrations at pH value of 5.8.
2
Type of nanoparticle
Surface charge density/C m−2
Plain silica Amino-modified Hydrocarbon-modified
− 5.4∙10−2 + 8.1∙10−2 n.a.
Colloids and Surfaces A xxx (xxxx) xxx–xxx
J. Nagel et al.
Fig. 3. Heigth profiles of different types of silica nanoparticles on the surface of PC parts, measured by AFM. The arrow has a length equal to the particle diameter, 200 nm. The circle emphasises a peak with gaps at both sides.
Fig. 2. Schematic structure of the used silica particles.
water. Moreover, FTIR spectra of their products after pyrolysis exhibited hydrocarbon fragments (see Supplementary material). Since the OTMS particles could not be dispersed in water, no zeta potential measurement and titration could be carried out. Nevertheless, this behaviour revealed a hydrophobic surface of the particles, i.e., that binding of the hydrocarbon chains was successful. This is in contrast to the other particles, which could be easily dispersed in water and are, thus, hydrophilic. Consequently, the particles shown in Fig. 2 have formed.
established, although the receding angle should be increased for real applications [17]. This could probably be realised by a more controlled structuring of the particles on the substrate surface. The PC surfaces with immobilised silica nanoparticles were investigated by different microscopic approaches. SEM images (see figure SM-4, Supplemental material) as well as The AFM topography images (see figure SM-5, Supplemental material) showed a distribution similar to that on a glass surface with single particles and relatively dense islands of particles, but no complete coverage. In these investigations, the immobilisation of single particles was in focus. Fig. 3 shows the height profiles of the PC surface for the different types of particles. The curve of the immobilised plain silica nanoparticles partly exhibited a flat base line with some peaks. Most of the peaks were close to other, pointing to the formation of islands, as already suggested from the AFM image. The peaks had a width at the base of less than 200 nm. Moreover, the peak height was in a range between 10 and 40 nm. This suggests that the main parts of the particles were surrounded by the PC melt and, thus, were effectively embedded. However, a considerable part of the surface area of a particle was uncovered by PC and, thus, potentially useful for further interactions. The surface with immobilised amino-modified particles exhibited a straight line with only some peaks. The peaks had, again, a width of less than 200 nm. The peak height was about 30–40 nm above the base line, pointing to an effective embedding, as in the case of the plain silica particles. Moreover, there were negative peaks on both sides of each peak (see circle in Fig. 3), suggesting that there was a small gap between the particle and the PC, which increased the accessible surface area. The gap had an apparent depth of less than 20 nm, but for a detailed analysis, the curvature of the AFM tip should be taken into account. The profile of the surface with transferred hydrocarbon-modified particles showed peaks with a width of more than 200 nm. This is more than the particle diameter and was a result of the curvature of the AFM tip. The peaks were mostly single, but close to other peaks. However, the height of the peaks was about 200 nm, too, as shown by the arrow in Fig. 3. This suggested that the particles lay on top of the PC surface rather than being embedded. In some cases, even particles were immobilised on top of the first layer. The SEM image of a PC surface with immobilised silica particles in Fig. 4 exhibited random distribution of the particles, with some formation of clusters. The particle surface density was estimated to be similar to that on the substrate. This pointed to high transfer ratio, but it should be investigated more in detail, if the collective properties of the layer are coming more into focus. The plain particles in Fig. 4A are mostly wetted by the melt and permanently immobilised on the PC surface. This conclusion agrees to that drawn from the AFM profiles.
3.2. Nanoparticles immobilised on PC Layers on glass slides were formed by adsorption or dispersion drying. This resulted in random layers with loosely packed nanoparticles (see Supplementary material). The diameter of the particles as found by analysis of the SEM images on a glass slide, see figure SM-3 in the Supplemental materials, was about 200 nm. These layers were mounted in an injection moulding machine and brought in contact with the PC melt during injection. The surfaces of the formed parts were investigated, see Table 2. The advancing contact angle of plain PC was 83°. Immobilisation of silica particles resulted in a slightly lower advancing angle. The receding angle was largely reduced to 31°. The reason for that could be the interaction between water and the surface of the immobilised silica particles, which resulted in a structured surface. However, PEI was also immobilised by covalent bonding to the PC surface, since it was used as adsorption layer during the formation of the silica particle layer on the substrate surface [12,16]. The adsorption of aminomodified silica particles was carried out without the use of PEI, because it adsorbed spontaneously on glass surface. Nevertheless, the contact angles of the PC surface with amino-modified nanoparticles were lower than those of plain PC. This pointed to an accessibility of the surface of the hydrophilic particles immobilised on the PC surface. On the other hand, the contact angles of the PC surface with immobilised hydrocarbon-modified silica particles had higher advancing and receding angles than that of plain PC. A trend to super hydrophobic behaviour is Table 2 Dynamic contact angles of the PC surfaces against water. Surface
Advancing angle/°
Receding angle/°
PC PC-silica (PEI) PC-silica amino-modified PC-silica hydrocarbon-modified
83 78 72 150
66 31 50 112
3
Colloids and Surfaces A xxx (xxxx) xxx–xxx
J. Nagel et al.
Fig. 4. SEM images of different types of silica nanoparticles on the surface of PC parts, after removing of adhesive tapes. A) Plain, B) hydrocarbon-modified, C) amino-modified nanoparticles, D) enlargement of an area with amino-modified particles.
4. Conclusions
The wetting is based on the high surface energy of the silica nanoparticles. Thus, they were mechanically fixed. They withstand even the adhesion of the adhesive tape. Only some particles, which most likely were not deeply embedded during the melt contact, were removed on treatment with the tape. On the other hand, the hydrocarbon-modified particles (see Fig. 4B) were removed almost completely on treatment with the adhesive tape. The reason for that can be explained by the non-polar character of the particle surfaces. In case of a higher surface energy of the melt (advancing angle against water < 90°) compared to the surface energy of the solid, only a partial wetting can be expected. Consequently, the particles were not really embedded but lay only on top of the PC surface. This agrees with results of the profile measurements. The amino-modified nanoparticles in Fig. 4C, in contrast, were almost completely embedded and almost completely wetted, like in the case of the plain silica particles. Unlike them, there is always a small gap between the particle and the melt. This is particularly visible in the enlargement on the right side, Fig. 4D. This supports the assumption from the appropriate profile in Fig. 3. Small molecules from the environment can enter this gap. Therefore, the assemblies of aminomodified nanoparticles on the PC surface exhibited obviously the highest free surface area of the assembly types. The particles could not be removed by treatment with the adhesive tape, and, thus, were strongly bound to the PC. Probably, the particles were fixed by chemical bonds between the amino groups and the carbonate groups of PC under formation of urethane linkages, in addition to a mechanical fusing [16]. Recently, the thermal processes in nanometre dimensions at the interface during a contact of a hot thermoplastic melt and a mould or a thin film on a mould was investigated by numerical simulations. It was found, that the melt was cooled down rapidly and freezes usually instantly due to the low temperature and high thermal conductivity of the mould. Using the processing conditions for PC as applied in this paper, the simulations revealed a time at a temperature above the glass transition temperature of PC of only about 1 μs [18]. From this point of view, an instant freezing of the melt on contact with the nanoparticle surface should be expected. However, the presented experiments suggest that the interface interactions played a dominant role for the wetting process even in nanometre dimensions, despite the rapid cooling.
Different types of functional groups, i.e. amino groups and hydrocarbon chains, were bound to silica nanoparticles. Layers of these particles were deposited onto the surface of a glass substrate and transferred to a PC melt by injection moulding. The embedding and the bond strength of the immobilised particles were dependent on the type of surface functionalisation of the nanoparticles. Particles with high surface energy, i.e. plain and amino-modified silica particles, were almost completely embedded according to their high surface energy. This resulted in a form-fit immobilisation as shown in Fig. 5A. The main parts of the particles were covered by PC, but a certain amount of surface area was still accessible for molecules from the environment. In addition to that, a chemical bonding on PC is suggested for amino-
Fig. 5. Scheme of different types of wetting and embedding of surface-modified silica nanoparticles with PC melt. A) Almost complete wetting of the hydrophilic silica particles (plain, amino-modified), B) Partly wetting of the hydrophobic silica particles (hydrocarbon-modified).
4
Colloids and Surfaces A xxx (xxxx) xxx–xxx
J. Nagel et al.
detection, ACS Appl. Mater. Interfaces. 5 (2013) 2800–2814. [2] L.M. Rossi, N.J.S. Costa, F.P. Silva, R. Wojcieszak, Magnetic nanomaterials in catalysis: advanced catalysts for magnetic separation and beyond, Green Chem. 16 (2014) 2906–2933. [3] Y. Lvov, F. Caruso, Biocolloids with ordered urease multilayer shells as enzymatic reactors, Anal. Chem. 73 (2001) 4212–4217. [4] Z.J. Deng, S.W. Morton, E. Ben-Akiva, E.C. Dreaden, K.E. Shopsowitz, P.T. Hammond, Layer-by-Layer nanoparticles for systemic codelivery of an anticancer drug and siRNA for potential triple-Negative Breast cancer treatment, ACS Nano 7 (2013) 9571–9584. [5] J.J. Richardson, M. Bjoernmalm, F. Caruso, Technology-driven layer-by-layer assembly of nanofilms, Science 348 (2015) 2491. [6] G. Decher, J. Hong, J. Schmitt, Buildup of ultrathin multilayer films by a selfAssembly process .3. consecutively alternating adsorption of anionic and cationic polyelectrolytes on charged surfaces, Thin Solid Films 210 (1992) 831–835. [7] T.R. Andersen, T.T. Larsen-Olsen, B. Andreasen, A.P.L. Bottiger, J.E. Carle, M. Helgesen, E. Bundgaard, K. Norrman, J.W. Andreasen, M. Jorgensen, F.C. Krebs, Aqueous processing of low-band-gap polymer solar cells using roll-to-roll methods, ACS Nano 5 (2011) 4188–4196. [8] R.D. Deshmukh, R.J. Composto, Direct observation of nanoparticle embedding into the surface of a polymer melt, Langmuir 23 (2007) 13169–13173. [9] J. Nagel, P. Chunsod, C. Zimmerer, F. Simon, A. Janke, G. Heinrich, Immobilization of gold nanoparticles on a polycarbonate surface layer during molding, Mater. Chem. Phys. 129 (2011) 599–604. [10] F. Kroschwald, J. Nagel, A. Janke, F. Simon, C. Zimmerer, G. Heinrich, B. Voit, Gold nanoparticle layers from multi-step adsorption immobilised on a polymer surface during injection molding, J. Appl. Polym. Sci. 133 (2016) 43608. [11] J. Nagel, M. Bräuer, B. Hupfer, K. Grundke, S. Schwarz, D. Lehmann, Investigations on the reactive surface modification of polycarbonate by surface-reactive injection molding, J. Appl. Polym. Sci. 93 (2004) 1186–1191. [12] C. Zimmerer, J. Nagel, G. Steiner, G. Heinrich, Nondestructive molecular characterization of polycarbonate-Polyvinylamine composites after thermally induced aminolysis, Macromol. Mater. Eng. 301 (2016) 648–652. [13] M. Qhobosheane, S. Santra, P. Zhang, W. Tan, Biochemically functionalized silica nanoparticles, Analyst 126 (2001) 1274–1278. [14] G. Petzold, R. Rojas, M. Mende, S. Schwarz, Application relevant characterization of aqueous silica nanodispersions more, J. Dispers. Sci. Technol. 30 (2009) 1216–1222. [15] M. Mihai, I. Stoica, S. Schwarz, pH-sensitive nanostructured architectures based on synthetic and/or natural weak polyelectrolytes more, Colloid Polym. Sci. 289 (2011) 1387–1396. [16] M. Jang, S. Park, N.Y. Lee, Polycarbonate bonding assisted by surface chemical modification without plasma treatment and its application for the construction of plastic-based cell arrays, Sens. Actuator A-Phys. 206 (2014) 57–66. [17] M. Ma, R.M. Hill, Superhydrophobic surfaces, Curr. Opin. Colloid Interface Sci. 11 (2006) 193–202. [18] J. Nagel, G. Heinrich, Temperature transitions on the surface of a thermoplastic melt during injection moulding and its use for chemical reactions, Int. J. Heat Mass Trans. 55 (2012) 6890–6896.
modified particles. Because of the controlled increase of effective surface area compared to a flat surface, those surfaces with immobilised particles have potential for use in sensing applications. If the layers on the substrate were ordered, e.g. in photonic crystals, then it might be possible to create also permanently immobilised ordered structures on PC. To address the different fields of application, particle properties like size and surface functionalities should be properly adjusted. Future work will deal with those problems. Silica nanoparticles with hydrocarbon-modified surfaces were only slightly embedded by the PC melt. The situation is sketched in Fig. 5B. The particles could be removed by an adhesive tape and, thus, were not permanently immobilised. The remaining PC surfaces exhibited imprints according to the particle size and shape. Those cavities may be used as reaction sites or, according to a structuring, as negative photonic crystals. An approach was presented in this paper for the fabrication of nanostructured polymer surface that could probably be realised for different polymer-nanoparticle combinations and for arbitrary size of parts and surface areas. The embedding of the nanoparticles in the polymer matrix can be controlled by the surface properties. Wetting by the melt is a dominant factor, despite the fast freezing of the melt on contact with the cold objects. Acknowledgements This work was supported by the German Research Foundation (HE 4466/17-1). The authors thank Mrs. Martina Priebs and Mrs. Anja Caspari for measuring of the contact angles and the zeta potentials, respectively. Covestro AG is thanked for supplying the Polycarbonate. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.colsurfa.2017.05.016. References [1] Z. Yue, F. Lisdat, W.J. Parak, S.G. Hickey, L. Tu, N. Sabir, D. Dorfs, N.C. Bigall, Quantum-Dot-Based photoelectrochemical sensors for chemical and biological
5
Contents lists available at ScienceDirect
Colloids and Surfaces A journal homepage: www.elsevier.com/locate/colsurfa
Immobilisation of different surface-modified silica nanoparticles on polymer surfaces via melt processing ⁎
Jürgen Nagela, , Felix Kroschwaldb, Cornelia Bellmanna, Simona Schwarza, Andreas Jankea, Gert Heinricha,c a b c
Leibniz-Institut für Polymerforschung Dresden e.V., 01069 Dresden, Hohe Straße 6, Germany Gerflor Mipolam GmbH, Mülheimer Str. 27, 53840 Troisdorf, Germany Technische Universität Dresden, Institute of Materials Science, 01069 Dresden, Hohe Straße 6, Germany
A R T I C L E I N F O
A B S T R A C T
Keywords: Nanoparticles Nanotechnology Surfaces Injection moulding
The immobilisation of silica nanoparticles on the surface of a flowing polymer melt is studied during injection moulding. We use silica nanoparticles of identical size of about 200 nm, but with different surface functionalities: plain unmodified particles with silanole groups as well as amino- and hydrocarbon modified particles. The particles were brought in contact with a polycarbonate melt at a temperature of 300 °C. In our investigation, we show that particles with polar surfaces are more embedded than those with non-polar surfaces. Immobilised polar particles exhibited also a higher adhesion, tested with an adhesive tape, whereas the nonpolar particles can be easily removed by peeling off the adhesive tape. This paper reveals that surface properties have a larger influence on embedding of nanoparticles than thermal conditions due to cooling of the melt.
1. Introduction Nanoparticles and nanostructured materials are the very promising materials for future demands. They are already today applied in many fields of applications. Many applications, e.g. as sensor [1], catalyst [2], require the immobilisation of nanoparticles on solid surfaces. This can be realised very efficiently by adsorption using the layer-by-layer technique [3–6] or by dispersion drying [7]. However, since the nanoparticles are only bound by weak chemical or physical interactions, the mechanical stability of those layers is limited. A higher stability could be approached by embedding in a polymer melt. Such an embedding of gold nanoparticles was realised in the group of Composto by moulding in a hot press [8]. Here, the particles were embedded, partly completely, in the polymer melt. The authors explained this by the high surface energy of the particles and capillary forces that resulted in extensive wetting by the melt. The time of contact with the melt at high temperature was arbitrary in these experiments, so that conditions for the adjustment of a thermodynamic equilibrium prevailed. However, fabrication of polymer parts by hot pressing is not very efficient. Our previous work dealt with embedding of small gold nanoparticles using injection moulding [9,10]. Embedding occurred also here. However, some parts of the surface area remained uncovered by the melt and were still accessible for small molecules from solution.
⁎
This was used for catalytic reactions. Wetting of the particle surface by melt occurred despite the high cooling rate once the hot melt hit the cool surface. However, this could be a result of the small particle diameter of about 20 nm. In this paper, we used nanoparticles with equal size and different surface functionalisation. Pure silica particles are polar and exhibit silanole groups on their surface. Amino-modified particles are also polar and have, in addition, groups that can reactively bind onto polycarbonate [11,12]. Hydrocarbon-modified particles are typically non-polar and exhibit a low surface energy. The particle diameters were about 200 nm and, thus, larger than those of the discussed gold nanoparticles. Since the nanoparticles affect the temperature diffusion from melt to mould, we used particles with uniform size to realise equal cooling conditions. The aim of this paper is, to find out, whether there is an influence of the surface modification on wetting and embedding of particles in the polymer melt under the special conditions of injection moulding, or if the rapid cooling dominates the structure formation. The question is to be answered, whether the embedding of nanoparticles into the surface of a thermoplastic part can be controlled by surface functionalisation. 2. Material and methods Uniform silica nanoparticles with a diameter of about 200 nm were
Corresponding author. E-mail address: [email protected] (J. Nagel).
http://dx.doi.org/10.1016/j.colsurfa.2017.05.016 Received 27 February 2017; Received in revised form 5 May 2017; Accepted 8 May 2017 0927-7757/ © 2017 Elsevier B.V. All rights reserved.
Please cite this article as: Nagel, J., Colloids and Surfaces A (2017), http://dx.doi.org/10.1016/j.colsurfa.2017.05.016
Colloids and Surfaces A xxx (xxxx) xxx–xxx
J. Nagel et al.
obtained from Geltech (Orlando, USA). Octadecyltrimethoxysilane (OTMS), N-(3-Trimethoxysilylpropyl)-diethylenetriamine (DETA), sodium poly(ethylene sulphonate) (PES) and Poly(diallyldimethylammoniumchlorid) (PDADMAC) were ordered at Sigma-Aldrich, Germany. Polyethyleneimine (hyperbranched, Mw 750 kg mol−1, Sigma-Aldrich, Germany, PEI) was ordered as 50 wt.% aqueous solution. Polycarbonate (Makrolon LED 2245, Bayer Materials Science GmbH, now Covestro AG, Germany, PC) was used for moulding experiments. The plain silica nanoparticles were used as received. They exhibited a low porosity with a surface area of 4.5 m2 g−1, as measured by BET using nitrogen gas. This value equalled that given from the manufacturer. Surface modification with amino groups was carried out according to a protocol taken from literature in a 500 ml flask under nitrogen purging [13]. 0.15 g acetic acid, 250 ml water and 2.5 ml DETA were mixed. 0.6 g silica nanoparticles were added and stirred for 6 h at room temperature. After that time, they were centrifuged, and the supernatant solution was changed against water. This procedure was repeated three times. The yield was about 84%. Surface modification with OTMS was carried out in a 100 ml flask with nitrogen purging. 500 mg silica nanoparticles were dispersed in 10 ml toluene in an ultrasonic bath. Then a solution of 3 mmol OTMS in 10 ml toluene was added dropwise via a septum. The solution was stirred first 1 h at 80 °C and then 2 h under reflux. Finally, the mixture was centrifuged three times with exchange of the supernatant solution against toluene. The product was dried in vacuum. The yield of the hydrocarbon-modified nanoparticles was about 62%. A Zetasizer Nano ZS (Malvern Instruments GmbH, UK) was used for measurements of zeta potential using electrophoresis. Polyelectrolyte titration was carried out at pH 5.8 using a 702 SM Titrino (Metrohm GmbH & Co. KG, Germany) and with PDADMAC and PES as titrant, respectively. Glass microscope slides with the dimensions 1 inch x 3 inch were used as substrates. They were cleaned in NoChromix solution and rinsed with water prior to use. Layer of amino-modified nanoparticles were prepared using cleaned plain glass slides. The adsorption was carried out using a stirred dispersion with a nanoparticle concentration of 2 g l−1 at pH 6. After an adsorption time of 30 min, the glass slides were rinsed for 15 min with water. For the adsorption of plain silica particles, PEI was adsorbed in advance onto the glass slides. The solution was prepared by dissolving 10 g of the concentrated PEI solution and 29 g NaCl in 400 ml water. Then, the pH was adjusted to 8 by adding HCl, and the volume was adjusted to 500 ml with water. The glass slides were rinsed in this solution for 15 min under stirring and then rinsed in water for 15 min. The adsorption of plain silica nanoparticles was carried out from a colloidal solution with a concentration of 1.2 g l−1 at pH 6. Hydrocarbon-modified nanoparticles were dispersed in toluene at a concentration of 52 mg ml−1. Layers were prepared by dispersion drying of a film of 32 μl of the suspension over the substrate. Substrates with a dense layer of nanoparticles on the surface were translucent. The substrates covered with nanoparticle layers were mounted in the mould cavity of a plunger injection moulding machine (construction made in-house). PC was dried before use at 110 °C for 5 h in vacuum. PC melt was injected at a melt temperature of 300 °C and with a mould temperature of 80 °C using typical processing parameters of PC (injection rate of 7.3 mm3 s−1, injection pressure of 100 bar, cooling time of 26 s). The specimens produced were round plates with a diameter of 50 mm and a thickness of about 2 mm. The substrates were removed from the formed part. They were clear after the moulding, but the surface of the parts fabricated were translucent, if the substrate were translucent before moulding, pointing to a successful transfer. In some cases, an adhesive tape with an adhesive strength of 5.4 N per 18 mm was mounted on the part surface and removed to characterise the adhesive strength of the particle immobilisation. Scanning electron images were taken with an Ultra Plus Gemini (Carl Zeiss SMT AG, Germany). The acceleration voltage was varied
Fig. 1. Zeta potentials of plain and amino modified silica nanoparticles as function of pH. Measured in 10−3 M KCl.
between 1 and 20 kV. The working distance was 3 mm 3 nm platinum were sputtered on the sample surfaces in advance using a Leica EM SCD 050 (Leica Mikrosysteme Vertriebs GmbH, Germany). Atomic force microscopy (AFM) images were taken with a Dimension 3100 Nanoscope (Veeco Instruments Inc., USA). The measurements were carried out in tapping mode with silicon SPM sensors (Budget Sensors, Bulgaria) having a spring constant of 4 N m−1 and a resonance frequency of 75 kHz. The radius of the tip was smaller than 10 nm. The dynamic contact angles were measured using an OCA35 XL (DataPhysics Instruments GmbH, Germany). An average was calculated from at least four measurements per sample. 3. Results and discussion 3.1. Characterisation of the particles Fig. 1 shows that the isoelectric point (IEP) of the plain silica nanoparticles was around a pH value of 2.5. The zeta potential decreased on increasing pH until a value of −58 mV was approached at pH 10. This is the typical behaviour for silica surface [14]. On the other side, the amino-modified nanoparticles showed an IEP of about 9.7, which is typical for basic functionalities. The plateau value of 50 mV is characteristic for complete protonated functional groups on the particle surface. This result is typical for amino-modified surfaces [15] and pointed to a stable binding of the amino silane onto the silica nanoparticle. The dispersions were titrated to zero potential with PES or PDADMAC, respectively, and the results are shown in Table 1. The consumption of ions was related to the total surface area, calculated from the particle diameter and the total amount of particles in the dispersion. The surface charge density of plain silica nanoparticles was −5.4∙10−2C m‐2. The alternation by treatment with the amino silane to 8.1∙10−2 pointed to binding of the amino silane onto the silica surface. The hydrocarbon-treated silica nanoparticles could not be dispersed in Table 1 Surface charge densities from polyelectrolyte titrations at pH value of 5.8.
2
Type of nanoparticle
Surface charge density/C m−2
Plain silica Amino-modified Hydrocarbon-modified
− 5.4∙10−2 + 8.1∙10−2 n.a.
Colloids and Surfaces A xxx (xxxx) xxx–xxx
J. Nagel et al.
Fig. 3. Heigth profiles of different types of silica nanoparticles on the surface of PC parts, measured by AFM. The arrow has a length equal to the particle diameter, 200 nm. The circle emphasises a peak with gaps at both sides.
Fig. 2. Schematic structure of the used silica particles.
water. Moreover, FTIR spectra of their products after pyrolysis exhibited hydrocarbon fragments (see Supplementary material). Since the OTMS particles could not be dispersed in water, no zeta potential measurement and titration could be carried out. Nevertheless, this behaviour revealed a hydrophobic surface of the particles, i.e., that binding of the hydrocarbon chains was successful. This is in contrast to the other particles, which could be easily dispersed in water and are, thus, hydrophilic. Consequently, the particles shown in Fig. 2 have formed.
established, although the receding angle should be increased for real applications [17]. This could probably be realised by a more controlled structuring of the particles on the substrate surface. The PC surfaces with immobilised silica nanoparticles were investigated by different microscopic approaches. SEM images (see figure SM-4, Supplemental material) as well as The AFM topography images (see figure SM-5, Supplemental material) showed a distribution similar to that on a glass surface with single particles and relatively dense islands of particles, but no complete coverage. In these investigations, the immobilisation of single particles was in focus. Fig. 3 shows the height profiles of the PC surface for the different types of particles. The curve of the immobilised plain silica nanoparticles partly exhibited a flat base line with some peaks. Most of the peaks were close to other, pointing to the formation of islands, as already suggested from the AFM image. The peaks had a width at the base of less than 200 nm. Moreover, the peak height was in a range between 10 and 40 nm. This suggests that the main parts of the particles were surrounded by the PC melt and, thus, were effectively embedded. However, a considerable part of the surface area of a particle was uncovered by PC and, thus, potentially useful for further interactions. The surface with immobilised amino-modified particles exhibited a straight line with only some peaks. The peaks had, again, a width of less than 200 nm. The peak height was about 30–40 nm above the base line, pointing to an effective embedding, as in the case of the plain silica particles. Moreover, there were negative peaks on both sides of each peak (see circle in Fig. 3), suggesting that there was a small gap between the particle and the PC, which increased the accessible surface area. The gap had an apparent depth of less than 20 nm, but for a detailed analysis, the curvature of the AFM tip should be taken into account. The profile of the surface with transferred hydrocarbon-modified particles showed peaks with a width of more than 200 nm. This is more than the particle diameter and was a result of the curvature of the AFM tip. The peaks were mostly single, but close to other peaks. However, the height of the peaks was about 200 nm, too, as shown by the arrow in Fig. 3. This suggested that the particles lay on top of the PC surface rather than being embedded. In some cases, even particles were immobilised on top of the first layer. The SEM image of a PC surface with immobilised silica particles in Fig. 4 exhibited random distribution of the particles, with some formation of clusters. The particle surface density was estimated to be similar to that on the substrate. This pointed to high transfer ratio, but it should be investigated more in detail, if the collective properties of the layer are coming more into focus. The plain particles in Fig. 4A are mostly wetted by the melt and permanently immobilised on the PC surface. This conclusion agrees to that drawn from the AFM profiles.
3.2. Nanoparticles immobilised on PC Layers on glass slides were formed by adsorption or dispersion drying. This resulted in random layers with loosely packed nanoparticles (see Supplementary material). The diameter of the particles as found by analysis of the SEM images on a glass slide, see figure SM-3 in the Supplemental materials, was about 200 nm. These layers were mounted in an injection moulding machine and brought in contact with the PC melt during injection. The surfaces of the formed parts were investigated, see Table 2. The advancing contact angle of plain PC was 83°. Immobilisation of silica particles resulted in a slightly lower advancing angle. The receding angle was largely reduced to 31°. The reason for that could be the interaction between water and the surface of the immobilised silica particles, which resulted in a structured surface. However, PEI was also immobilised by covalent bonding to the PC surface, since it was used as adsorption layer during the formation of the silica particle layer on the substrate surface [12,16]. The adsorption of aminomodified silica particles was carried out without the use of PEI, because it adsorbed spontaneously on glass surface. Nevertheless, the contact angles of the PC surface with amino-modified nanoparticles were lower than those of plain PC. This pointed to an accessibility of the surface of the hydrophilic particles immobilised on the PC surface. On the other hand, the contact angles of the PC surface with immobilised hydrocarbon-modified silica particles had higher advancing and receding angles than that of plain PC. A trend to super hydrophobic behaviour is Table 2 Dynamic contact angles of the PC surfaces against water. Surface
Advancing angle/°
Receding angle/°
PC PC-silica (PEI) PC-silica amino-modified PC-silica hydrocarbon-modified
83 78 72 150
66 31 50 112
3
Colloids and Surfaces A xxx (xxxx) xxx–xxx
J. Nagel et al.
Fig. 4. SEM images of different types of silica nanoparticles on the surface of PC parts, after removing of adhesive tapes. A) Plain, B) hydrocarbon-modified, C) amino-modified nanoparticles, D) enlargement of an area with amino-modified particles.
4. Conclusions
The wetting is based on the high surface energy of the silica nanoparticles. Thus, they were mechanically fixed. They withstand even the adhesion of the adhesive tape. Only some particles, which most likely were not deeply embedded during the melt contact, were removed on treatment with the tape. On the other hand, the hydrocarbon-modified particles (see Fig. 4B) were removed almost completely on treatment with the adhesive tape. The reason for that can be explained by the non-polar character of the particle surfaces. In case of a higher surface energy of the melt (advancing angle against water < 90°) compared to the surface energy of the solid, only a partial wetting can be expected. Consequently, the particles were not really embedded but lay only on top of the PC surface. This agrees with results of the profile measurements. The amino-modified nanoparticles in Fig. 4C, in contrast, were almost completely embedded and almost completely wetted, like in the case of the plain silica particles. Unlike them, there is always a small gap between the particle and the melt. This is particularly visible in the enlargement on the right side, Fig. 4D. This supports the assumption from the appropriate profile in Fig. 3. Small molecules from the environment can enter this gap. Therefore, the assemblies of aminomodified nanoparticles on the PC surface exhibited obviously the highest free surface area of the assembly types. The particles could not be removed by treatment with the adhesive tape, and, thus, were strongly bound to the PC. Probably, the particles were fixed by chemical bonds between the amino groups and the carbonate groups of PC under formation of urethane linkages, in addition to a mechanical fusing [16]. Recently, the thermal processes in nanometre dimensions at the interface during a contact of a hot thermoplastic melt and a mould or a thin film on a mould was investigated by numerical simulations. It was found, that the melt was cooled down rapidly and freezes usually instantly due to the low temperature and high thermal conductivity of the mould. Using the processing conditions for PC as applied in this paper, the simulations revealed a time at a temperature above the glass transition temperature of PC of only about 1 μs [18]. From this point of view, an instant freezing of the melt on contact with the nanoparticle surface should be expected. However, the presented experiments suggest that the interface interactions played a dominant role for the wetting process even in nanometre dimensions, despite the rapid cooling.
Different types of functional groups, i.e. amino groups and hydrocarbon chains, were bound to silica nanoparticles. Layers of these particles were deposited onto the surface of a glass substrate and transferred to a PC melt by injection moulding. The embedding and the bond strength of the immobilised particles were dependent on the type of surface functionalisation of the nanoparticles. Particles with high surface energy, i.e. plain and amino-modified silica particles, were almost completely embedded according to their high surface energy. This resulted in a form-fit immobilisation as shown in Fig. 5A. The main parts of the particles were covered by PC, but a certain amount of surface area was still accessible for molecules from the environment. In addition to that, a chemical bonding on PC is suggested for amino-
Fig. 5. Scheme of different types of wetting and embedding of surface-modified silica nanoparticles with PC melt. A) Almost complete wetting of the hydrophilic silica particles (plain, amino-modified), B) Partly wetting of the hydrophobic silica particles (hydrocarbon-modified).
4
Colloids and Surfaces A xxx (xxxx) xxx–xxx
J. Nagel et al.
detection, ACS Appl. Mater. Interfaces. 5 (2013) 2800–2814. [2] L.M. Rossi, N.J.S. Costa, F.P. Silva, R. Wojcieszak, Magnetic nanomaterials in catalysis: advanced catalysts for magnetic separation and beyond, Green Chem. 16 (2014) 2906–2933. [3] Y. Lvov, F. Caruso, Biocolloids with ordered urease multilayer shells as enzymatic reactors, Anal. Chem. 73 (2001) 4212–4217. [4] Z.J. Deng, S.W. Morton, E. Ben-Akiva, E.C. Dreaden, K.E. Shopsowitz, P.T. Hammond, Layer-by-Layer nanoparticles for systemic codelivery of an anticancer drug and siRNA for potential triple-Negative Breast cancer treatment, ACS Nano 7 (2013) 9571–9584. [5] J.J. Richardson, M. Bjoernmalm, F. Caruso, Technology-driven layer-by-layer assembly of nanofilms, Science 348 (2015) 2491. [6] G. Decher, J. Hong, J. Schmitt, Buildup of ultrathin multilayer films by a selfAssembly process .3. consecutively alternating adsorption of anionic and cationic polyelectrolytes on charged surfaces, Thin Solid Films 210 (1992) 831–835. [7] T.R. Andersen, T.T. Larsen-Olsen, B. Andreasen, A.P.L. Bottiger, J.E. Carle, M. Helgesen, E. Bundgaard, K. Norrman, J.W. Andreasen, M. Jorgensen, F.C. Krebs, Aqueous processing of low-band-gap polymer solar cells using roll-to-roll methods, ACS Nano 5 (2011) 4188–4196. [8] R.D. Deshmukh, R.J. Composto, Direct observation of nanoparticle embedding into the surface of a polymer melt, Langmuir 23 (2007) 13169–13173. [9] J. Nagel, P. Chunsod, C. Zimmerer, F. Simon, A. Janke, G. Heinrich, Immobilization of gold nanoparticles on a polycarbonate surface layer during molding, Mater. Chem. Phys. 129 (2011) 599–604. [10] F. Kroschwald, J. Nagel, A. Janke, F. Simon, C. Zimmerer, G. Heinrich, B. Voit, Gold nanoparticle layers from multi-step adsorption immobilised on a polymer surface during injection molding, J. Appl. Polym. Sci. 133 (2016) 43608. [11] J. Nagel, M. Bräuer, B. Hupfer, K. Grundke, S. Schwarz, D. Lehmann, Investigations on the reactive surface modification of polycarbonate by surface-reactive injection molding, J. Appl. Polym. Sci. 93 (2004) 1186–1191. [12] C. Zimmerer, J. Nagel, G. Steiner, G. Heinrich, Nondestructive molecular characterization of polycarbonate-Polyvinylamine composites after thermally induced aminolysis, Macromol. Mater. Eng. 301 (2016) 648–652. [13] M. Qhobosheane, S. Santra, P. Zhang, W. Tan, Biochemically functionalized silica nanoparticles, Analyst 126 (2001) 1274–1278. [14] G. Petzold, R. Rojas, M. Mende, S. Schwarz, Application relevant characterization of aqueous silica nanodispersions more, J. Dispers. Sci. Technol. 30 (2009) 1216–1222. [15] M. Mihai, I. Stoica, S. Schwarz, pH-sensitive nanostructured architectures based on synthetic and/or natural weak polyelectrolytes more, Colloid Polym. Sci. 289 (2011) 1387–1396. [16] M. Jang, S. Park, N.Y. Lee, Polycarbonate bonding assisted by surface chemical modification without plasma treatment and its application for the construction of plastic-based cell arrays, Sens. Actuator A-Phys. 206 (2014) 57–66. [17] M. Ma, R.M. Hill, Superhydrophobic surfaces, Curr. Opin. Colloid Interface Sci. 11 (2006) 193–202. [18] J. Nagel, G. Heinrich, Temperature transitions on the surface of a thermoplastic melt during injection moulding and its use for chemical reactions, Int. J. Heat Mass Trans. 55 (2012) 6890–6896.
modified particles. Because of the controlled increase of effective surface area compared to a flat surface, those surfaces with immobilised particles have potential for use in sensing applications. If the layers on the substrate were ordered, e.g. in photonic crystals, then it might be possible to create also permanently immobilised ordered structures on PC. To address the different fields of application, particle properties like size and surface functionalities should be properly adjusted. Future work will deal with those problems. Silica nanoparticles with hydrocarbon-modified surfaces were only slightly embedded by the PC melt. The situation is sketched in Fig. 5B. The particles could be removed by an adhesive tape and, thus, were not permanently immobilised. The remaining PC surfaces exhibited imprints according to the particle size and shape. Those cavities may be used as reaction sites or, according to a structuring, as negative photonic crystals. An approach was presented in this paper for the fabrication of nanostructured polymer surface that could probably be realised for different polymer-nanoparticle combinations and for arbitrary size of parts and surface areas. The embedding of the nanoparticles in the polymer matrix can be controlled by the surface properties. Wetting by the melt is a dominant factor, despite the fast freezing of the melt on contact with the cold objects. Acknowledgements This work was supported by the German Research Foundation (HE 4466/17-1). The authors thank Mrs. Martina Priebs and Mrs. Anja Caspari for measuring of the contact angles and the zeta potentials, respectively. Covestro AG is thanked for supplying the Polycarbonate. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.colsurfa.2017.05.016. References [1] Z. Yue, F. Lisdat, W.J. Parak, S.G. Hickey, L. Tu, N. Sabir, D. Dorfs, N.C. Bigall, Quantum-Dot-Based photoelectrochemical sensors for chemical and biological
5