High-throughput synthesis of CeO2 nanoparticles for transparent nanocomposites repelling Pseudomonas aeruginosa biofilms

Preventing bacteria from adhering to material surfaces is an important technical problem and a major cause of infection. One of nature’s defense strategies against bacterial colonization is based on the biohalogenation of signal substances that interfere with bacterial communication. Biohalogenation is catalyzed by haloperoxidases, a class of metal-dependent enzymes whose activity can be mimicked by ceria nanoparticles. Transparent CeO2/polycarbonate surfaces that prevent adhesion, proliferation, and spread of Pseudomonas aeruginosa PA14 were manufactured. Large amounts of monodisperse CeO2 nanoparticles were synthesized in segmented flow using a high-throughput microfluidic benchtop system using water/benzyl alcohol mixtures and oleylamine as capping agent. This reduced the reaction time for nanoceria by more than one order of magnitude compared to conventional batch methods. Ceria nanoparticles prepared by segmented flow showed high catalytic activity in halogenation reactions, which makes them highly efficient functional mimics of haloperoxidase enzymes. Haloperoxidases are used in nature by macroalgae to prevent formation of biofilms via halogenation of signaling compounds that interfere with bacterial cell–cell communication (“quorum sensing”). CeO2/polycarbonate nanocomposites were prepared by dip-coating plasma-treated polycarbonate panels in CeO2 dispersions. These showed a reduction in bacterial biofilm formation of up to 85% using P. aeruginosa PA14 as model organism. Besides biofilm formation, also the production of the virulence factor pyocyanin in is under control of the entire quorum sensing systems P. aeruginosa. CeO2/PC showed a decrease of up to 55% in pyocyanin production, whereas no effect on bacterial growth in liquid culture was observed. This indicates that CeO2 nanoparticles affect quorum sensing and inhibit biofilm formation in a non-biocidal manner.

A possible scenario could be as follows: In the reaction of cerium acetate hydrate (Ce(ac) 3 • × H 2 O) in benzyl ether an acid-base equilibrium ac − + HOCH 2 C 6 H 5 = ac-OCH 2 C 6 H 5 + OH − → precedes the formation of Ce(OH) 3 in an initial step at 120 °C (from Ce 3+ and OH − anions formed in the acid-base equilibrium). Ce(OH) 3 is sensitive to aerial oxygen and oxidized to CeO 2 in a topotactic condensation reaction, where water is eliminated 54 . In benzyl alcohol this water may be involved in a second acid-base reaction with oleylamine according to H 2 O + H 2 N-R = OH − + + NH 3 -R. In a competing reaction oleylamine acts as a capping agent in the surface complexation of the growing CeO 2 NPs. The hexagonal structure of the Ce(OH) 3 intermediate allows topotactic growth of anisotropic particles (e.g., nanorods) of CeO 2 (which adopts the cubic fluorite structure) 55 . The capping agent plays an important role for the size distribution of the NPs by binding to the growth species and controlling its growth rate [56][57][58] . Terminating the reaction at this stage leads to the formation of single-domain crystalline CeO 2 NPs with diameters of 2-7 nm, which could be dispersed in nonpolar solvents such as cyclohexane or toluene. Figure 1A shows commercially available CeO 2 NPs synthesized on an industrial scale. They have an ill-defined morphology and strong agglomeration. For the latter application small NPs with a high surface area are necessary, therefore the control of the morphology and size is mandatory. CeO 2 NPs from the batch (high-resolution transmission electron microscopy (HR)TEM images in Fig. 1B) are much smaller (two orders of magnitude) than their commercial counterparts and therefore suitable for application in nanocomposites. In the second step, the batch synthesis was transferred into a continuous setup. Fig. 1C shows that the CeO 2 NPs by segmented flow synthesis have essentially the same morphology as those obtained from the batch reaction (Fig. 1B). Compared to the commercial CeO 2 particles (Fig. 1A) the NPs from the segmented flow synthesis show a uniform morphology and size.

Synthesis of CeO 2 NPs under segmented flow conditions. The HRTEM micrograph in
However, clogging of the reactor could not be prevented under continuous flow conditions. Therefore, the synthesis was carried out under segmented flow using water as carrier medium. Advantages of this setup are the low costs and the boiling temperature of water. Water evaporates when the flow passes through the heating element. The water vapor leads to additional pressure between the segmented droplets of the reaction medium and thus prevents-independent of the flow rate-clogging of the reaction channel. A constant flow rate of the reaction medium and the water through the PTFE tube was achieved with a peristaltic pump. Nucleation was triggered by warming the reactant stream in a heating element.
Two heating elements (120° and 180 °C) were used to simulate the temperature profile of the batch process. The setup is shown in Figure S1A. The suitability of the PTFE tubing for the thermally induced formation of CeO 2 NPs was assessed in preliminary experiments under continuous flow conditions. The residence times in each oil bath was set to 2.7 min to simulate the reaction conditions from the batch synthesis (vide supra). The reaction solution had a reddish color and turned purple after passing through the tubes. Two phases were obtained: the upper phase was an emulsion of water, benzyl alcohol and oleylamine, the lower phase contained the precipitated yellow NPs ( Figure S1B).
Importantly, the reactions could be carried out in automated form with much shorter reaction times. For the batch synthesis, the product yield was 0.7 mg/min, while 15.8 mg/min (i.e., 22 times as much) were produced at the same temperature in segmented flow ( Figure S1C). The average crystallite sizes of the NPs (bulk sample) were determined by powder X-ray diffraction (PXRD, Fig. 2 red line) and Rietveld refinement (blue line, difference green line) 59,60 . Details concerning the refinement are given in the Supporting Information (Table S1-2). The NPs are single-crystalline. The X-ray diffractograms of the CeO 2 NPs by batch synthesis (Fig. 2A1) are very similar to those obtained by segmented flow reaction (Fig. 2A2) and show the expected intensities for (bulk) CeO 2 (lattice parameter a = 5.42 Å). Reflection broadening allowed extracting the particle size (average 3.2 nm and 5.8 nm by Rietveld analysis of the PXRD patterns for CeO 2 NPs (batch/segmented flow synthesis). The morphologies of the NPs from both processes are similar, NPs from the segmented flow process are slightly larger than those Commercial CeO 2 is produced by a gas-phase condensation process in which particles that are still hot are collected in a container. In this process, the hot particles sinter together and form large agglomerates, which do not form stable dispersions of isolated particles. In addition, no monodisperse particles are obtained in this type of mass production. Dispersions of polydisperse particles and large aggregates show strong light scattering and lead to cloudy opaque films after embedding in polymers. Surface chemistry of CeO 2 NPs. The oleylamine surfactant required an analysis of the CeO 2 surface chemistry to assess the effect of the organic ligands on the catalytic activity. FT-IR spectroscopy was used to identify the surface ligands of the as-synthesized CeO 2 NPs. Figure 2B shows the spectrum of pure oleylamine. The spectra of the CeO 2 NPs from the batch (Fig. 2C) and the segmented flow synthesis (Fig. 2D) indicate the presence of surface-bound oleylamine. The broad band between 3500 cm −1 and 3000 cm −1 envelopes the symmetric and asymmetric O-H stretches. The bands around 2920 cm −1 and 2850 cm −1 are characteristic for the C-H stretch of the terminal CH 3 group 56 . The absorptions in the range of 1550 cm −1 to 1400 cm −1 and in the range from 1250 cm −1 and 1400 cm −1 and close to 1000 cm −1 are typical for the stretching and bending vibrations surface-bound bi-and polydentate carbonate groups 61 . The surfactant coverage was determined by TGA ( Fig. 2E,F). Figure 2E shows thermogravimetric traces of the CeO 2 NPs from the batch synthesis. The first step in the diagram is associated with a mass due to surface-bound H 2 O (∆m rel = 4.05%). This step is initiated between 50 °C and 200 °C. Between 200 °C and 650 °C there is another mass drop corresponding to the remaining organic residues and traces of carbonate being stripped off (∆m rel = 15.45%). The remaining solid material was CeO 2 . NPs obtained under segmented flow (Fig. 2F) contained a higher amount of surface-bound H 2 O (∆m rel = 6.06%) because they were prepared in aqueous environment, while the mass loss of organic residues and carbonate was lower (∆m rel = 13.89%) compared to the batch synthesis. The FT-IR spectra of the products obtained by batch and segmented flow synthesis (Fig. 2C,D) show a weaker intensity of the vibration at around 2920 cm −1 and 2850 cm −1 for the C-H stretch of the terminal CH 3 group. This indicates a lower share of surface-bound oleylamine. Consequently, a larger amount of H 2 O was bound onto the surface. Based on the spectroscopic and thermoanalytical data a temperature-dependent (185 °C, 500 °C, 800 °C) analysis of surface chemistry and the resulting catalytic activity of the NPs was carried out.
X-ray analysis of the CeO 2 NPs from the batch preparation (Fig. 3A1) and segmented flow synthesis after annealing for 5 h at 185 °C (Fig. 3A2) showed no significant changes compared to their non-annealed counterparts ( Fig. 2A). CeO 2 NPs from segmented flow synthesis were also annealed for 5 h at 500 °C and 800 °C. The NP diameter increased from 5.8 nm (185 °C) to 16.2 nm (500 °C) and finally to 71.5 nm (800 °C). The sharp reflections at 500 °C and 800 °C (Fig. 3B,C) and the HRTEM images ( Figure S2A-C) show the increase in particle size. Since particle diameter correlated with surface area in an inverse fashion, the surface area was reduced at high annealing temperatures (Fig. 3D). The surface area for the NPs prepared at ambient temperature was 184 m 2 g −1 . It increased to 190 m 2 g −1 after annealing at 185 °C. Surface ligands of the CeO 2 NPs (ambient temperature) may cause errors in the BET measurements. Therefore, the surface area increased-formally-slightly after www.nature.com/scientificreports/ annealing, even though it is unfeasible physically. The surface areas decreased at higher temperatures to 78 m 2 g −1 (500 °C) and 17 m 2 g −1 (800 °C) due to grain growth. The FT-IR spectra of the CeO 2 NPs (batch preparation and segmented flow synthesis) showed after annealing for 5 h at 185 °C ( Fig. 3E) no organic ligand anymore, indicated by the absence of the bands around 2920 cm −1 and 2850 cm −1 , which are characteristic for the C-H stretch of the terminal CH 3 group 56 . The absorptions at 1570 cm −1 , between 1250 cm −1 and 1400 cm −1 and close to 1000 cm −1 are typical for the stretching and bending vibrations of surface-bound bi-and polydentate carbonate groups [61][62][63] . CeO 2 shows acid/base properties. Its acidic behavior is responsible for OAM surface binding. Its base properties are responsible for the binding of carbonate groups. This binding of carbonaceous species is responsible for the use of CeO 2 in gas-phase heterogeneous catalysis (e.g. in exhaust catalysts) 64,65 . It is also an important feature of hydroxylated and non-hydroxylated CeO 2 NPs in aqueous environment, where active surface sites are blocked by adsorbed species.
After calcination for 5 h at 800 °C all vibrational bands of the surface species vanished (Fig. 3E), and the ζ-potentials of the NPs before and after annealing changed from positive to negative values. For CeO 2 NPs (ambient temperature, segmented flow) an average value of 50.2 mV ± 1.5 mV was obtained. After annealing for 5 h at 185 °C the ζ-potential decreased slightly to 47.3 mV ± 0.12 mV. Annealing for 5 h at 500 °C reduced the ζ-potential to − 4.78 mV ± 0.17 mV, and after annealing for 5 h at 800 °C it was − 18.9 mV ± 0.59 mV (Fig. 3F).
Haloperoxidase activity. The haloperoxidase activity of CeO 2 NPs was determined by UV/Vis spectroscopy with the phenol red (PR) assay (Fig. 4A). The spectra in Fig. 4B,C show the time-dependent intensity change of the two principal absorption bands for PR for CeO 2 NPs (batch or segmented flow synthesis, after annealing at 185 °C) as haloperoxidation catalyst. The spectral changes and the slope of the absorbance correspond to the reaction rate. Figure 4D,E show the absorbance change for PR at 590 nm measured as a function of time for CeO 2 (catalyst, 25 µg/mL), H 2 O 2 (300 µM), PR (50 µM) and KBr (25 mM). The shift of the absorption bands is due to the fourfold oxidative bromination of PR to tetrabromophenol blue (TBPB).
As the haloperoxidase activity depends on the concentration of the H 2 O 2 substrate, the Michaelis-Menten kinetics was evaluated by varying the H 2 O 2 concentration. The kinetic data were fitted using the Hill equation. Table 1 shows the parameters derived from a fit to the Hill equation ( Figure S3). Evaluating the kinetic parameters for CeO 2 NPs obtained from the segmented flow synthesis (ambient temperature) was not possible, because their water dispersibility was low due to the OAM surface ligands.
Brunauer-Emmett-Teller (BET) surface areas were determined to normalize the reaction rates to surface areas between 190 m 2 g −1 (ambient temperature) and 17 m 2 g −1 (800 °C). Table 1 shows that the bromination rate increases with the annealing temperature (185 °C vs. 500 °C) for CeO 2 . Based on the surface areas (190 m 2 g −1 vs. 78 m 2 g −1 ) and the ζ-potentials (47.3 mV ± 0.12 mV vs. − 4.78 mV ± 0.17 mV) one may expect a higher activity www.nature.com/scientificreports/ for the sample treated at 185 °C. However, the bromination rate appears to depend not only on surface area and ζ-potential. Another factor might be the carbonate surface concentration of the CeO 2 sample annealed at 185 °C, as surface bound CO 3 2− groups may quench the reaction or block active sites. After annealing at 800 °C the bromination rate dropped significantly. The saturation rate (K m value) increased, i.e., the affinity for H 2 O 2 decreased. The Hill coefficient decreased initially from 185 to 800 °C. Thus, cooperative effects decrease by increasing the annealing temperature from 185 to 500 °C and finally to 800 °C. The catalytic activity described in terms of the turnover frequency (TOF) or the catalytic constant (k cat ) 66 depends on the catalyst surface area, time, and mass concentration. The rate of reaction (ROR) is defined by Eq. (1).
Here, the accessible surface area per mass equivalent of the catalyst is expressed through the BET surface (S BET in m 2 g −1 ). The turnover of PR to TBPB is surface specific because only the catalyst surface is accessible for the substrate. Equation (1) can be evaluated based on time (v max /M min −1 ), surface area (S BET /m 2 g −1 ) and mass concentration (β(Cat)/g L −1 ) and yields the molar amount of TBPB that was formed by reaction of HOBr with PR per unit surface and time. The calculated ROR value for CeO 2 (185 °C) is 0.031 µmol m −2 min −1 , while CeO 2 (500 °C) shows the highest ROR value (0.185 µmol m −2 min −1 ). For the CeO 2 annealed at 800 °C we obtained a value of 0.071 µmol m −2 min −1 . For the application of CeO 2 NPs on polycarbonate surfaces there are two important requirements. (i) A high surface area to volume ratio is essential for catalytic activity. (ii) High dispersibility of the NPs is necessary to ensure uniform distribution of the NPs on the polycarbonate surface during the dip coating process. In this way, we could fabricate a smooth and transparent nanocomposite. CeO 2 NPs annealed at 185 °C were the most successful candidate.  PC surface, the PC panels were exposed to an oxygen plasma for 20 min. Contact angle measurements of polycarbonate with water, showed a decrease from 86.75° before treatment to 26.88° after oxygen plasma treatment ( Figure S4), indicating a significant change in surface energy. The plasma treatment generated new oxygen containing groups such as C=O, O-C-C or O-C=O as reported by Shikova et al. 67 Dip-coating of the plasma-treated PC panels in a dispersion of CeO 2 particles in cyclohexane (ambient temperature) and water (185 °C) resulted in the binding of the particles to the polycarbonate surface by condensation reaction between the polar groups of polycarbonate surface and the free hydroxyl groups of the particles leading to a very dense and stable coating of CeO 2 NPs on the PC surface (10 mg/mL, immersion speed of 60 mm/min, deposition time of 5 s and withdrawal speed 60 mm/min). The coated plates were dried in air (110 °C), excess CeO 2 was removed by washing with ultra-pure MilliQ (QMQ) water, and the plates were dried at 80 °C overnight. Surface coating prevented embedding of CeO 2 particles in the PC matrix without having an active surface area for catalytic reactions. It also circumvented the problem of producing stable dispersions from a mixture of CeO 2 NPs and organic binders which often result in hazy films 68 due to particle aggregation. The coating with CeO 2 NPs was demonstrated by SEM/EDX (Fig. 5A,B, not annealed, annealed at 185 °C). The binding of the CeO 2 NPs to the surface proved successful, as both CeO 2 /PC panels showed high catalytic activity (demonstrated by conversion PR to TBPB in the PR assay, Fig. 5C,D). Importantly, coating of polycarbonate panels with CeO 2 NPs did not lead to any color change. Highly transparent CeO 2 /PC sample panels were obtained, which showed no difference in transmission compared to uncoated PC panels ( Fig. 5E; UV-vis absorption spectra in Figure S5). Similar Young's moduli / hardness values were obtained for the CeO 2 /PC samples. The Young's moduli for CeO 2 /PC with annealed NPs slightly (3.74 GPa ± 0.06 GPa) increase compared to those of CeO 2 /PC with non-annealed NPs (3.73 GPa ± 0,12 GPa), while the hardness value slightly decreased for the former ones from 274.81 MPa ± 8.76 MPa (ambient temperature) to 265.80 MPa ± 9.86 MPa (annealed at 185 °C). These values are much higher than those which were obtained for pure PC 2.32 ± 0.02 GPa (2.4 GPa ISO 527-1,-2) and 178.77 MPa ± 2.20 MPa 69 . As expected, the Young's modulus increased with surface-bound CeO 2 content.  CeO 2 /PC nanocomposites inhibit biofilm formation and affect pyocyanin production in P. aeruginosa. CeO 2 /PC composites were tested on their potential to inhibit bacterial biofilm formation and on the ability to interfere with bacterial quorum sensing with P. aeruginosa, a Gram-negative soil bacterium, recognized for its ubiquity, its antibiotic resistance mechanisms, which is known to contaminate drinking water distribution systems. Furthermore, P. aeruginosa is a potent biofilm producer and therefore widely used as an indicator strain 73 . Additionally, P. aeruginosa is known to be responsible for opportunistic infections in situations with limited barriers, such as skin defects in wounds, or epithelial impairment in advanced stages of chronic pulmonary diseases or cystic fibrosis 74 .
To test biofilm formation on CeO 2 coated PC surfaces, the bacteria were grown for 72 h at 30 °C on different composites placed in 24-well plates. After removing the planktonic cells, the attached bacteria were stained with crystal violet. In comparison to the uncoated PC plate, both nanoparticle composites were able to decrease bacterial biofilm formation (Fig. 6). CeO 2 particles annealed at 185 °C resulted in a 65% reduction in of P. aeruginosa biofilm formation, showing a stronger effect than untreated CeO 2 particles, where the reduction was only 30%. While the biofilm on the untreated PC plate was distributed equally all over the surface, the distribution was inhomogeneous on both CeO 2 RT/PC composites and the CeO 2 185 °C/PC nanocomposites (Fig. 6).
Furthermore, we analyzed biofilm formation on CeO 2 coated PC surfaces by quantifying single live and dead cells on the surfaces. For that purpose, the P. aeruginosa PA14 was grown for 72 h at 30 °C on the different PC samples placed in 24-well plates. After removing the planktonic cells, the attached bacteria were stained with SYTO 9 (total and live cell count) and propidium iodide (dead cell count). In comparison to the blank plate (Fig. 7A), both CeO 2 /PC composites showed a decrease in bacterial cell attachment. While the biofilm on the blank plate was equally distributed all over the surface, less biofilm clusters were visible on the CeO 2 RT/PC (Fig. 7B) and much less as well as smaller clusters on the CeO 2 185 °C/PC (Fig. 7C) composites. Although both nanoparticle composites were able to reduce the bacterial attachment, resulting in a reduced number of cell clusters on the surface, the most promising effect could be achieved with the CeO 2 185 °C/PC composites. In comparison to the blank plate, approximately 60% of the cells were attached to the CeO 2 RT/PC surfaces, and only 15% on the CeO 2 185 °C/PC composites (Fig. 7D). Furthermore, the composites had no effect on the life/ dead cell-ratio, as the portion of dead cells on the treated surfaces were comparable to that on the blank plate (approximately 15-20%) (Fig. 7D). These data clearly show the inhibitory and non-biocidal effect of the composites on bacterial biofilm formation. This could be also demonstrated in liquid culture. The addition of CeO 2 to a liquid culture of P. aeruginosa PA14 had no effect on the growth of the bacteria at 30 °C (Fig. 7E). Therefore, CeO 2 must interfere with bacterial biofilm formation on another than a toxic way. www.nature.com/scientificreports/ For the organization within a biofilm, bacteria usually communicate via small diffusible molecules sensing their cell count, a process called quorum sensing 75 . The CeO 2 NPs promote the production of halogenated compounds, which are similar to quorum sensing molecules of Gram-negative bacteria. Therefore, we assumed that the non-toxic biofilm inhibiting effect of CeO 2 composites might be due to a putative interference with the bacterial cell-cell communication, an effect that is generally referred to as quorum quenching 76 .
To get first evidence on the molecular mechanism of how CeO 2 /PC inhibit bacteria biofilm formation, we analyzed its effect on the production of the virulence factor pyocyanin in P. aeruginosa PA14. Besides biofilm formation, the biosynthesis of pyocyanin is also under control of the entire quorum sensing systems in these bacteria 77 . For that purpose, the bacteria were cultivated for 72 h at 37 °C on blank plates and on CeO 2 /PC composites. Then, the PC plates were removed, and the remaining culture fluid was analyzed on the presence of pyocyanin as described under Materials and methods. Both nanocomposites had a negative effect on pyocyanin production (Fig. 8). However, CeO 2 /PC composite annealed at 185 °C showed the highest decrease of up to 55% compared to the control cultures. Since P. aeruginosa is known to communicate via a distinct cell-cell signaling scheme 78 , it can be assumed that the NPs act on at least one of the three known quorum sensing systems. The results regarding inhibition of biofilm formation and pyocyanin production in P. aeruginosa finally showed the applicability of CeO 2 /PC nanocomposites to inhibit bacterial communication-most likely-and therefore bacterial biofilm formation as a non-toxic strategy.

Conclusion
CeO 2 NPs were coated on polymer surfaces to fabricate highly transparent and hard polymer/nanoparticle composites, which display the intrinsic haloperoxidase-like characteristics of the CeO 2 NPs as demonstrated by the bromination of PR and the formation of the virulence factor pyocyanin, which is regulated by the quorum sensing system. Since P. aeruginosa is known to communicate via a distinct cell-cell signaling scheme, it can be assumed that the biomimetic activity of the CeO 2 NPs acts on at least one of the three known quorum sensing systems. An effective segmented flow synthesis of CeO 2 NPs was devised. It allows the continuous synthesis of non-agglomerated material with a single mixer to fabricate nanomaterials for polycarbonate displays in large quantities. The yield can easily be enhanced by parallelization. The highly reproducible (automated) synthesis of CeO 2 NPs in copious amounts allowed to systematically analyze the role of particle size, ion concentration and surface chemistry on their (bio)catalytic properties. www.nature.com/scientificreports/ An important point is that ceria does not leach because of its extremely low solubility (solubility product log K L = − 60) 79,80 . This is also the reason why cerium-despite its abundance (comparable to copper) 81 -is involved in biological processes only under extreme conditions in volcanic mudpots 82 . The low solubility makes ceria nanoparticles also highly biocompatible: little or no cytotoxic effects were observed (depending on the uptake pathway) 83 . Ceria nanoparticles have been proposed for biomedical applications (e.g. in sun creams) 54 , and it was demonstrated with in vivo experiments that ceria NPs can even protect cells against irradiation and oxidative stress 55,84 .
In contrast, various metal NPs (e.g., Cu or Ag) preventing biofouling are known to be harmful to the environment 85 . V 2 O 5 NPs showed excellent performance to prevent biofouling on surfaces in marine environments 43 , but leaching leading to soluble polyvanadates in combination with reported carcinogenic and mutagenic effects make any practical application unlikely [86][87][88][89] . TiO 2 and SnO 2 nanoparticles show antibiofouling properties, but their mode of biocidal activity TiO 2 is associated with the induction of oxidative stress due to photogenerated reactive oxygen species under UV irradiation. The production costs of TiO 2 and SnO 2 dispersible nanoparticles are moderate 90,91 . TiO 2 particles can be incorporated into various polymers, such as polyamide 10 poly(lactic-co-glycolic acid) 92 or polytetrafluorethylene 93 with antibiofouling properties. Titania nanoparticles have a broad activity spectrum against microorganisms, and they have been considered environmentally friendly with non-contact biocidal action 94 . Their recent classification as a "suspected carcinogen" by the EU, however, may lead to restrictions or even a ban on their chemical use in consumer products 95 . CeO 2 offers a sustainable alternative to TiO 2 , with non-toxic properties and high abundance, more importantly catalytic activity does not depend on UV irradiation.
Because of their non-toxicity and environmental sustainability polycarbonate/CeO 2 nanocomposites may find broad application as green antimicrobials on plastics for automotive, aircraft, or railway components, for drinking bottles, glasses and food containers, where traditional biocides (toxines, antibiotics, etc.) are strictly prohibited because of toxicity, non-target impacts or the evolution of resistance by micro-organisms. Other applications could be touch panels, sporting helmets, glasses, fountain pens, portable devices for consumer electronics like smart phones, audio player cases, computer cases, durable, lightweight luggage, musical instruments, and toys. Especially in the health sector this may prevent infections that are becoming increasingly troublesome with the emergence of drug resistant bacteria and reduce the use of antibiotics.

Experimental details
Materials. Ce(ac) 3  Batch synthesis. Before the synthesis Ce(ac) 3 • × H 2 O was dried for 10 min at 130 °C in an oven. Subsequently 0.5 mmol (150 mg) of dried Ce(ac) 3 was transferred into a 100 mL three-necked flask, furthermore 7 mL of benzyl alcohol and 3 ml of oleylamine were added. The solution was heated up to 120 °C (5 °C/min) and kept at the temperature for 20 min. Then it was heated up to 180 °C (5 °C/min) for 30 min with a stirring speed of 400 rpm. The reaction solution was subsequently cooled down to room temperature. For the purification the product was washed three times with ethanol and cyclohexane and dried under vacuum at room temperature. oven. Subsequently, 1.4 mmol of dried Ce(ac) 3 was transferred into a beaker, 21 mL of benzyl alcohol and 9 mL of oleylamine were added, and the solution was stirred at room temperature. Deionized water was added in a second beaker. The overall tube length was 9 m. The first segment (3 m) was put in an oil bath with a temperature of 120 °C, the second segment (3 m) was heated in an oil bath to 180 °C, the last segment (3 m) was kept at room temperature. The two solutions were pumped separately, the flow rate for the metal precursor solution was 2 mL/min, the flow rate for the deionized water was 1.5 mL/min. The two solutions were connected through a T-fitting. The nanoparticles were washed three times with ethanol and cylclohexane and dried under vacuum at room temperature. The reaction mixture was also scaled up to the tenfold and even 20-fold of the amount of the original batch synthesis.
Powder X-ray diffraction. X-ray diffraction patterns were recorded on a STOE Stadi P diffractometer equipped with a Dectris Mythen 1 k detector in transmission mode using Mo Kα 1 radiation. Crystalline phases were identified according to the PDF-2 database using Bruker AXS EVA 10.0 software 96 . Full pattern profile fits (Pawley/Rietveld) were performed with TOPAS Academic 6.0 applying the fundamental parameter approach 59,60 .

Transmission electron microscopy. Transmission Electron Microscopy (TEM) mages for determin-
ing the size and morphology were acquired on a FEI Tecnai G2 Spirit microscope operating at 120 kV (LaB 6 filament), equipped with a Gatan US1000 CCD-camera (16-bit, 2048 × 2048 pixels), using the Gatan Digital Micrograph software 97  Brunauer-Emmet-Teller (BET). The specific surface area was determined using a 3P micro 300 adsorption-desorption device from 3P instruments. Measurements were conducted with an appropriate amount of sample (150 mg) at 77 K with nitrogen as analysis gas.
Preparation of CeO 2 /polycarbonate nanocomposites. Polycarbonate (PC) plates (ca. 1 × 1 cm) were washed with MQ-water and ultrasonicated for 10 min in isopropanol to remove contaminants. After that, PC plates were dried in air at 80 °C overnight. PC substrates were exposed to oxygen plasma for 20 min to produce hydroxyl groups onto the surface. The oxygen plasma was created at a pressure = 0.1 mbar and a power = 300 W. was added stepwise at a final concentration of 0.8 mM every 24 h. Then, the plates were rinsed with water to remove the planktonic cells. After drying for 5 min, 1 ml of 1% (w/v) crystal violet (Merck, Darmstadt) was added to the wells containing the plates. After 30 min incubation at room temperature, unbound crystal violet was removed by gently submerging the plates for two times in water. The plate was then air-dried over-night at room temperature. For quantification, 1 ml of 30% (v/v) acetic acid (Roth, Karlsruhe) was added to the plates to solubilize the crystal violet from the biofilm. After 15 min of incubation at room temperature, absorbance was quantified in a plate reader (Tecan, Salzburg) at 575 nm.
Pyocyanin analyses. For pyocyanin production, an alternate method of already published protocols were used 100 . P. aeruginosa was cultivated in LB medium over night at 30 °C. The cultures were then diluted in LB, in a volume of 1 ml per well of a 24-well polystyrene microtiter plate (Sarstedt, Nürnbrecht) at a final OD 600 of 0.5. Additionally, KBr (Roth, Karlsruhe, Germany) and H 2 O 2 (Roth, Karlsruhe, Germany) at a final concentration of 32 mM and 0.8 mM respectively were added to the wells. The non-coated plates and CeO 2 /PC composites were then added to the wells, respectively. The microtiter plate was then incubated for 72 h under shaking conditions (350 rpm) at 37 °C. Every 24 h H 2 O 2 was added at a final concentration of 0.8 mM. Afterwards, the composites were taken out of the wells and the remaining supernatant inside the wells was collected in a microreaction tube (Eppendorf, Germany). Cells were separated from culture fluids via centrifugation at 16 × 1000 g for 15 min. Then, the supernatant was passed through 0.22 μm filters (Merck, Darmstadt). Cell-free culture fluids were then analysed for pyocyanin in a plate reader (Tecan, Salzburg) at 695 nm.
Epifluorescence microscopy and fluorescence quantification. P. aeruginosa PA14 was cultivated in LB medium over night at 30 °C. The cultures were then diluted in LB, in a volume of 1 ml per well of a 24-well polystyrene microtiter plate (Sarstedt, Nürnbrecht) at a final OD 600 of 0.5. Additionally, KBr (Roth, Karlsruhe, Germany) and H 2 O 2 (Roth, Karlsruhe, Germany) at a final concentration of 32 mM and 0.8 mM, respectively, were added to the wells. The non-coated blank plates and nanoparticle composites were then added to the wells. www.nature.com/scientificreports/ cells. Afterwards, the plates were placed in 1 ml of a combined SYTO 9 and propidium iodide solution (Thermo Fisher, Pittsburgh, USA) and incubated for 30 min at 30 °C. Then, the plates were rinsed again with water and mounted on microscopy slides. Biofilm samples were microscopically analyzed on an Axio Imager 2 fluorescence microscope, which is especially designed for material surface analysis (Carl Zeiss, Jena, Germany). The fluorophore SYTO 9 was visualized with an excitation of 470 nm while propidium iodide was visualized with an excitation of 558 nm. The emission settings for the used filters were 509 nm for SYTO 9 and 570 nm for propidium iodide. The images were then processed using the ZEN 3.3 blue software 101 . The green and red fluorescence intensities within the different field of views were then quantified and set into correlation using ImageJ software (Version: 2.0.0-rc-69/1.52n) 98 .