**3. Conclusions**

Families of silver functionalized porous frameworks based around surfactant templated SBA-15 mesoporous silica were synthesized to investigate the impact of surface chemistry and pore architecture on antibacterial performance. Silver nanoparticles introduced by wet impregnation of AgNO3 and subsequent thermal processing comprised mixed metal and carbonate phases, with the concentration of silver carbonate (and dissolution rate for ionic Ag+) inversely proportional to particle size independent of the framework. The introduction of a conformal titania monolayer coating of SBA-15 improved the dispersion of silver nanoparticles, corresponding bulk and surface carbonate content, and hence ionic silver release kinetics. This observation likely reflects the ability of defective reducible metal oxides to act as nucleation centers for metals [38–40], thereby increasing the density of small nanoparticles, and amphoteric nature of titania and ability to capture atmospheric CO2 [41] which can subsequently react with (weakly basic) silver oxides [42,43] at the nanoparticle interface. Additional rate enhancements for ionic silver were observed for hierarchical microporous-mesoporous SBA-15 frameworks, presumably due to their stabilization of even smaller silver nanoparticles and hence higher carbonate concentrations. Silver release rates from the porous frameworks directly correlated with broad spectrum antibacterial activity, with a 16–38% decrease in decimal reduction times for *Staphylococcus aureus* and *Pseudomonas aeruginosa* observed following titania functionalization, and a further 65–89% reduction following the introduction of macroporosity. The quantitative structure-function relationship identified between the concentration of Ag2CO3 and antibacterial efficacy will guide the development of future nanocomposite architectures, notably optimization of the reducible metal oxide coatings and pore networks to further promote ionic silver release.

#### **4. Materials and Methods**

Polystyrene bead templates were synthesized using a method developed by Vaudreuil et al. [44] in which styrene, divinyl benzene (comonomer, Sigma Aldrich, 80%) and potassium persulphate (initiator, Sigma Aldrich, >99%) were the reagents. The reaction was performed on a large scale in a 2-liter jacketed Radleys Reactor-Ready system at 90 ◦C. Deionized water (1.5 L) was introduced to the reactor, along with a Leibig condenser, thermocouple, and a nitrogen line at 1.5 bar pressure. The reactor was stirred at 300 rpm overnight to outgas the solution. Styrene (140 mL, Sigma Aldrich, >99%) and divinylbenzene (27 mL) were washed with NaOH (0.1 M) three times in separate separating funnels and added to the reaction vessel. Potassium persulfate (Sigma Aldrich, 0.35 g) was dissolved in deionized water (20 mL) at 80 ◦C. After 30 min of stirring (300 rpm) in the reactor at 90 ◦C, the potassium persulfate solution was added. After stirring for 3 h, the solid particles were recovered as a concentrated solution and stored in a freezer overnight, then the product was allowed to warm before being filtered, washed with ethanol and the beads dried at 80 ◦C overnight.

SBA-15 was prepared using a cooperative self-assembly method [29], in which a 2.6 wt% solution of Pluronic P123 triblock copolymer (poly(ethylene glycol)-poly(propylene glycol)-poly(ethylene glycol) (Sigma Aldrich), in 1.6-M HCl solution was stirred (500 rpm) at 35 ◦C. Tetraethyl orthoxysilicate (Sigma Aldrich, 98%) was then added to the mixture, at a molar ratio of 60:1 [TEOS]:[P-123]. The mixture was aged at 80 ◦C for 24 h without stirring in a sealed container in an oven. The resultant solid material was filtered, then washed with ethanol before drying in air at 100 ◦C overnight. Removal of the P123 framework was performed by calcination at 500 ◦C in a muffle furnace for 6 h with a ramp rate of 1 ◦C/min. Macropores were introduced by the addition of polystyrene beads to the SBA synthesis (2.3.1) at a weight ratio of 5.3:1 [PS beads]:[TEOS] in the initial mixture.

The grafting of titania onto the surface of the prepared silica materials was done using a modified procedure by Landau et al. [30] in which triethylamine is used to activate the surface silanols on the silica and allow the reaction to proceed at lower temperatures. To ensure a uniform coating of TiO2, the reaction must be performed under completely dry conditions, due to the facile hydrolysis of the titania precursor, which will readily form large titania particles in the presence of water. The synthetic procedure involves mixing titanium isopropoxide (Sigma Aldrich,) in anhydrous toluene

(Aldrich, water content <0.002%), adding triethylamine (Sigma Aldrich, >99%) and MM-SBA-15 or SBA-15 material whilst stirring at 85 ◦C for 6 h under nitrogen flow. The concentration of titanium isopropoxide was 145 g/L, the molar ratio between titanium isopropoxide and SBA-15 was fixed at 3.5 and the triethylamine: SBA-15 weight ratio at 1.5 on a scale of 5 g of SBA-15/MM-SBA-15. After the reaction, the solid was separated by filtration, washed with toluene (300 mL) and inserted in a 0.5 wt% water-ethanol solution (500 mL) under stirring for 24 h. The resultant solid was washed with ethanol, dried in air in an oven at 90 ◦C for 24 h, then calcined for 1 h at 250 ◦C, 1 h 400 ◦C and finally for 4 h at 500 ◦C all at 1 ◦C/min.

Ag NPs were deposited via wet impregnation using a solution of aqueous silver nitrate (99.9%, Sigma-Aldrich). A slurry of the silver precursor and support (10 mL of 5–25 μM+1g support) was stirred for 18 h at room temperature before heating to 50 ◦C. After 5 h, agitation was ceased and the solid aged at 50 ◦C for a further 24 h to yield a dry powder. Dried samples were calcined at 500 ◦C (ramp rate 1 ◦C/min) in static air for 3 h.

XRD patterns were recorded on either a PANalytical X'pertPro diffractometer fitted with an X'celerator detector and Cu Kα (1.54 Å) source or a Bruker D8 Advance diffractometer fitted with a LynxEye high-speed strip detector and Cu Kα (1.54 Å) source. Both instruments were calibrated against either Si (PANalytical, Malvern, UK) or SiO2 (Bruker, Billerica, MA, USA) standards. Low angle patterns were recorded for 2*θ* = 0.3–8◦ with a step size of 0.01◦. Wide angle patterns were recorded for 2*θ* = 10–80◦ with a step size of 0.02◦.

N2 adsorption isotherms were recorded using a Nova 4000 porosimeter (Quantachrome, Boynton Beach, FL, USA), before which the samples were thoroughly degassed under vacuum at 120 ◦C for 2 h. T-plot analysis was used to calculate microporosity. Data was analyzed using NOVAWin version 11 (Quantachrome, Boynton Beach, FL, USA).

XPS analysis was recorded using a Kratos Axis Hsi X-ray photoelectron spectrometer (Kratos Analytical Ltd., Manchester, UK) using a monochromated Al Kα (1486.6 eV) anode. Data was charge corrected to adventitious carbon at 284.6 eV and analyzed using CASAXPS version 2.3.15. Ag 3D peaks were fitted using a Doniach-Sunjic modified Gaussian-Lorentz peak shape and a doublet separation of 6 eV.

TEM analysis was performed using a JEOL 2100 microscope (JEOL Ltd., Tokyo, Japan) with a LaB6 source and HT of 180 kV. Particle sizing was performed using ImageJ 1.46r software (open source). Samples were prepared using a drop casting method in ethanol onto continuous carbon grids.

Dissolution rates were determined by stirring 10 mg of composite in 25 mL of 0.5 M NaNO3 at 37 ◦C with periodic aliquots of the analyte solution measured for silver content using an Agilent 6130B single Quad (ESI) ICP-MS (Agilent Technologies, Santa Clara, CA, USA) calibrated against a range of silver concentrations made by serial dilution of a 1000 ppm Ag in 1% HNO3 standard from Sigma-Aldrich.

The antibacterial performance of all materials was evaluated against *Staphylococcus aureus* ATCC 6538, MRSA ATCC 33591, *Escherichia coli* NCTC 10418, *Bacillus subtilis* NCTC 8236, *Pseudomonas aeruginosa* ATCC 15442 and *Clostridium difficile* ATCC 9689, which are representative Gram-positive and Gram-negative problematic organisms found in hospital environments. Zone of inhibition (ZoI) tests were performed by inoculating the surface of a nutrient agar plate (Oxoid, Basingstoke, UK) with an excess volume (3 mL) of nutrient broth which had previously been inoculated and incubated to a cell density of ~108 colony-forming units (cfu)/mL as determined spectrophotometrically using a Perkin-Elmer Lambda 10 UV-Vis spectrophotometer. The liquid was manipulated by agitation to provide a confluent inoculum and the excess fluid removed to waste using a sterile pipette. Using a sterilized boring tool, 5 mm holes were then bored into the agar, and 100 μL of a solution of 10 mg of Ag nanocomposite in 5 mL of simulated body fluid (SBF, see the Supplementary Materials Table S2) dispensed into the borehole using a calibrated micropipette. SBF was prepared according to a method from Kokubo et al. [45] 750 mL of deionized water was stabilized at 37 ◦C with stirring, to this the following ions were added: NaCl (7.996 g, Sigma Aldrich >99%), NaHCO3 (0.35 g, Sigma Aldrich >99%), KCl (0.224 g, Sigma Aldrich >99%), K2HPO4·3H2O (0.228 g, Sigma Aldrich >99%), MgCl2 (0.305 g, Sigma Aldrich >99%), HCl (40 mL, 1 kmol/L, Fisher scientific 37%), CaCl2 (0.278 g, Sigma Aldrich >99%), Na2SO4 (0.071 g, Sigma Aldrich >99%) and (CH2OH)3CNH2 (6.057 g, Sigma Aldrich 99%). Finally, the pH was adjusted to 7.35 using HCl solution (1 kmol/L, Fisher scientific 37%). Plates were then incubated at 37 ◦C overnight, photographed, and calibrated zone areas determined using ImageJ software (open source).

Quantitative antimicrobial activity was determined by logarithmic reduction [13,14]. Here, 5 mg of Ag nanocomposite material was added to an Eppendorf tube kept in dark conditions containing 1 mL of either *S. aureus* or *P. aeruginosa* in a nutrient broth at concentrations of 107 cfu·mL−1. 100 μL aliquots of the resulting suspensions were subsequently removed at 0, 60, 240 min and 24 h, and added to a 1 mL solution of Tween 20 (Fisher, 1%), sodium dodecyl sulphate (Sigma-Aldrich, 0.4%) and sodium chloride (Sigma-Aldrich, 0.85%) in deionized water to neutralise any soluble silver species [13,14].

Each of the resulting neutralized solutions was serially diluted with phosphate buffered saline (PBS) prior to plating onto agar and incubation at 37 ◦C for 24 h. The experiments were all run with positive and negative controls of silver nitrate and without any nanocomposite respectively. After incubation, the number of colonies present on the agar was counted by visual inspection, and normalized relative to the initial colony count in the negative control at time t = 0 min to determine the logarithmic reduction of bacteria. All experiments were performed in triplicate, with mean values and standard deviations reported.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2079-6382/7/3/55/s1, Figure S1: (left) N2 adsorption-desorption isotherms, and BJH pore size distributions for parent SBA-15, TiO2/SBA-15 and TiO2/MM-SBA-15. Type-IV isotherms with H1 hysteresis is characteristic of mesoporous SBA-15, Table S1: Elemental analysis of Ag/SBA-15, Ag/TiO2/SBA-15 and Ag/TiO2/MM-SBA-15, Figure S2: XRD patterns for the three prepared support materials, SBA-15, TiO2/SBA-15 and TiO2/MM-SBA-15, Figure S3: XRD patterns for (a) Ag/SBA-15; (b) Ag/TiO2/SBA-15; and (c) Ag/TiO2/MM-SBA-15 as a function of silver loading. All reflections associated with fcc silver metal., Figure S4: Particle size distributions from TEM for (a) 0.3 wt%; (b) 0.95 wt% and (c) 2.4 wt% Ag nanoparticles deposited on SBA-15, Figure S5: Particle size distributions from TEM (a) 0.3 wt%; (b) 0.9 wt% and (c) 2.4 wt% Ag nanoparticles deposited on TiO2/SBA-15, Figure S6: Particle size distributions from TEM (a) 0.25 wt%; (b) 0.75 wt% and (c) 1.2 wt% Ag nanoparticles deposited on TiO2/MM-SBA-15, Figure S6: Particle size distributions from TEM (a) 0.25 wt%, (b) 0.75 wt% and (c) 1.2 wt% Ag nanoparticles deposited on TiO2/MM-SBA-15, Figure S7: Mean particle size and standard deviation for Ag/SBA-15, Ag/TiO2/SBA-15, and Ag/TiO2/MM-SBA-15 as a function of bulk silver loading, Figure S8: Fitted Ag 3d XP spectra of (a) Ag/SBA-15; (b) Ag/TiO2/SBA-15; and (c) Ag/TiO2/MM-SBA-15 as a function of Ag loading. All spectra fitted to carbonate and metal spin-orbit doublets possessing common lineshapes and fixed binding energies, Figure S9: XANES profiles for (a) 0.3 wt%; (b) 0.95 wt% and (c) 2.4 wt% Ag/SBA-15, Figure S10: XANES profiles for (a) 0.3 wt%; (b) 0.9 wt% and (c) 2.4 wt% Ag/TiO2/SBA-15, Figure S11: XANES profiles for (a) 0.25 wt%; (b) 0.75 wt% and (c) 1.2 wt% Ag/TiO2/MM-SBA-15, Figure S12: Fitted dissolution kinetics for Ag/SBA-15, Ag/TiO2/SBA-15, and Ag/TiO2/MM-SBA-15 as a function of particle size, Figure S13: Zone of Inhibition assays for *Staphylococcus aureus*, MRSA, *Escherichia coli*, *Bacillus subtilis*, *Pseudomonas aeruginosa* and *Clostridium difficile* normalized to the mass of silver for (a) Ag/SBA-15; (b) Ag/TiO2/SBA-15; and (c) Ag/TiO2/MM-SBA-15, Figure S14: Logarithmic reductions for SBA-15, Ag/TiO2/SBA-15, and Ag/TiO2/MM-SBA-15 against (a) *Staphylococcus aureus* and (b) *Pseudomonas aeruginosa* after 24 h incubation, Figure S15: Logarithmic reductions for (a) Ag/SBA-15; (b) Ag/TiO2/SBA-15; and (c) Ag/TiO2/MM-SBA-15 against *Staphylococcus aureus* and *Pseudomonas aeruginosa*, Figure S16: Logarithmic colony forming units of (a) *S. aureus*; and (b) *P. aeruginosa* as a function of time for SBA-15, TiO2/SBA-15, and TiO2/MM-SBA-15, Table S1: Elemental analysis of Ag/SBA-15, Ag/TiO2/SBA-15 and Ag/TiO2/MM-SBA-15, Table S2: Ion concentrations in SBF solution.

**Author Contributions:** Formal analysis—M.A.I., A.C.H. and A.F.L.; Funding acquisition—K.W. and A.F.L.; Investigation—M.A.I.; Methodology—M.A.I., B.B., A.C.H. and L.O.; Supervision—K.W. and A.F.L.; Visualization—L.J.D. and C.M.A.P.; Writing—original draft, M.A.I.; Writing—review & editing, A.F.L.

**Acknowledgments:** We thank the Knowledge Economy Skills Scholarship (K.E.S.S.) program and Polymer Health Technology for providing funding for this work.

**Conflicts of Interest:** The authors declare no conflicts of interest.
