ACTAUNIVERSITATIS

UPSALIENSISUPPSALA

2015

Digital Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Science and Technology 1250

Antibacterial Strategies forTitanium Biomaterials

ERIK UNOSSON

ISSN 1651-6214ISBN 978-91-554-9241-0urn:nbn:se:uu:diva-249181

Dissertation presented at Uppsala University to be publicly examined in Siegbahn Hall,The Ångström Laboratory, Lägerhyddsvägen 1, Uppsala, Friday, 5 June 2015 at 09:15 forthe degree of Doctor of Philosophy. The examination will be conducted in English. Facultyexaminer: Professor Timo Närhi (University of Turku, Finland).

AbstractUnosson, E. 2015. Antibacterial Strategies for Titanium Biomaterials. Digital ComprehensiveSummaries of Uppsala Dissertations from the Faculty of Science and Technology 1250. 72 pp.Uppsala: Acta Universitatis Upsaliensis. ISBN 978-91-554-9241-0.

Titanium and titanium based alloys are widely used in dentistry and orthopedics to replacehard tissue and to mend broken bones. It has become a material of choice due to its lowdensity, high strength, good biocompatibility and its capacity to integrate closely with the bone.Today, modern materials and surgical techniques can enable patients to live longer, and aid inmaintaining or regaining mobility for a more fulfilling life. There are, however, instances whereimplants fail, and one of the primary causes for implant failure is infection.

This thesis deals with two possible ways of reducing or eliminating implant associatedinfections; TiO2 photocatalysis, where a surface can become antibacterial upon irradiation withUV light; and incorporation of silver, where a subsequent release of silver metal ions result inan antibacterial effect.

For the TiO2 photocatalysis strategy, a simple and cost effective chemical oxidation technique,using hydrogen peroxide (H2O2) and water, was used to create an active TiO2 surface ontitanium substrates. This surface was shown to effectively degrade an organic model substance(rhodamine B) by generating reactive oxygen species (ROS) under UV illumination. However,it was shown that Ti-peroxy radical species remaining in the surface after the H2O2-oxidationprocess, rather than generation of ROS from a heterogeneous photocatalytic process, wasresponsible for the effect. This discovery was further exploited in a TiO2/H2O2/UV system, whichdemonstrated synergy effects in both rhodamine B degradation tests and in antibacterial assays.

For the silver ion release strategy, a combinatorial materials science approach was employed.Binary Ag-Ti oxide gradients were co-deposited in a reactive (O2) environment using a custombuilt physical vapor deposition system, and evaluated for antibacterial properties. The approachenabled synthesis and composition-structure-property evaluation unlikely to have been achievedby traditional means, and the gradient coatings demonstrated antibacterial properties againstboth S. aureus and S. epidermidis according to silver ion release. The release was shown todepend more on structural features, such as surface area, crystallinity and oxidation state, thanon composition.

Ag-Ti oxide gradients were also evaluated under UV illumination, as Ag deposits oncrystalline TiO2 can enhance photocatalytic properties. In this work, however, the TiO2 wasamorphous and UV illumination caused a slight reduction in the antibacterial effect of silverions. This was attributed to a UV-induced SOS response in the S. epidermidis bacteria.

The results of this thesis demonstrate that both TiO2 photocatalysis, or UV induced activationof Ti-peroxy radical species, as well as incorporation of silver are viable antibacterial strategiesfor titanium biomaterials. However, their clinical applications are still pending risk-benefitanalyses of potential adverse host tissue responses.

Keywords: Titanium, silver, biomaterial, antibacterial, photocatalysis, hydrogen peroxide,reactive oxygen species, combinatorial materials science

Erik Unosson, Department of Engineering Sciences, Applied Materials Sciences, Box 534,Uppsala University, SE-75121 Uppsala, Sweden.

© Erik Unosson 2015

ISSN 1651-6214ISBN 978-91-554-9241-0urn:nbn:se:uu:diva-249181 (http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-249181)

To friends and family

“Truth is stranger than fiction” Probably Lord Byron

List of Papers

This thesis is based on the following papers, which are referred to in the text by their Roman numerals.

I Unosson, E., Persson, C., Welch, K., Engqvist, H. (2012) Photocata-lytic activity of low temperature oxidized Ti-6Al-4V. Journal of Materials Science: Materials in Medicine, 23(5):1173–1180.

II Unosson, E., Welch, K., Persson, C., Engqvist, H. (2013) Stability and prospect of UV/H2O2 activated titania films for biomedical use. Applied Surface Science. 285:317–323.

III Unosson, E., Tsekoura, E.K., Engqvist, H., Welch, K. (2013) Syn-ergetic inactivation of Staphylococcus epidermidis and Streptococ-cus mutans in a TiO2/H2O2/UV system. Biomatter, 3(4):1-6

IV Unosson, E., Rodriguez, D., Welch, K., Engqvist, H. (2015) Reac-tive combinatorial synthesis and characterization of a gradient Ag-Ti oxide thin film with antibacterial properties, Acta Biomaterialia. 11:503–510

V Unosson, E., Morgenstern, M., Engqvist, H., Welch, K. In vitro an-tibacterial properties and UV induced SOS response from Staphylo-coccus epidermidis on Ag/Ti oxide thin films. Submitted

Reprints were made with permission from the respective publishers.

Author’s Contributions

Paper I Major part of planning and evaluation. All experimental work and major part of writing.

Paper II Major part of planning and evaluation. All experimental work and major part of writing.

Paper III Part of planning and experimental work. Major part of eval-uation and writing.

Paper IV Major part of planning, evaluation, experimental work and writing.

Paper V Major part of planning and evaluation. Part of experimental work and major part of writing.

Also published

López, A., Unosson, E., Engqvist, H., Persson, C. (2011) Direct and in-teractive effects of three variables on properties of PMMA bone cement for vertebral body augmentation, Journal of Materials Science: Materials in Medicine. 22(6):1599–1606.

Engstrand, J., Unosson, E., Engqvist, H. (2012) Hydroxyapatite For-mation on a Novel Dental Cement in Human Saliva, ISRN Dentistry. 2012, Article ID: 624056, 7 pages.

Persson, C., Unosson, E., Ajaxon, I., Engstrand, J., Engqvist, H., Xia, W. (2012) Nano grain sized zirconia–silica glass ceramics for dental applica-tions. Journal of the European Ceramic Society, 32(16):4105–4110.

Unosson, E., Cai, Y., Jiang, X., Lööf, J., Welch, K., Engqvist, H. (2012) Antibacterial properties of dental luting agents: potential to hinder the development of secondary caries. International Journal of Dentistry, 2012, Article ID: 529495, 7 pages.

Contents

Introduction .................................................................................................. 13  

Aims and objectives ..................................................................................... 15  

Titanium as a biomaterial ............................................................................. 16  

Bacterial colonization and infection ............................................................. 19  

Photocatalysis on TiO2 ................................................................................. 22  

Chemical oxidation with H2O2 ..................................................................... 24  Photocatalytic activity ............................................................................. 25  Surface characteristics ............................................................................. 27  

Synergies in the TiO2/H2O2/UV system ....................................................... 28  Added benefits with added H2O2 ............................................................. 29  Stability of the system ............................................................................. 32  Effects against bacteria ............................................................................ 33  

Silver as an antibacterial agent ..................................................................... 36  Mechanism of action ............................................................................... 36  

Combinatorial materials science .................................................................. 38  Physical vapor deposition ........................................................................ 40  

Composition-structure-property relationships .............................................. 41  Gradient thin film characteristics ............................................................ 41  Resulting properties ................................................................................. 44  

UV-induced SOS response ........................................................................... 48  Photocatalysis on Ag-doped TiO2 ........................................................... 48  New gradient characteristics .................................................................... 49  Secondary effects of UV illumination ..................................................... 50  

Concluding remarks ..................................................................................... 54  

Future challenges .......................................................................................... 55  

Analytical techniques ................................................................................... 56  Rhodamine B degradation ....................................................................... 56  

Scanning electron microscopy (SEM) ..................................................... 57  Energy dispersive X-ray spectroscopy (EDS) ......................................... 57  Focused ion beam (FIB) .......................................................................... 58  Grazing incidence X-ray diffraction (GI-XRD) ...................................... 58  Surface roughness (Ra, Rq) ...................................................................... 59  Contact angle (CA) measurements .......................................................... 59  Bacterial viability assays ......................................................................... 59  Inductively coupled plasma–atomic emission spectroscopy (ICP-AES) 61  

Svensk sammanfattning ................................................................................ 62  

Acknowledgements ...................................................................................... 64  

References .................................................................................................... 66  

Abbreviations

AgNPs Silver nanoparticles CFU Colony forming unit cpTi Commercially pure titanium DCT Direct contact test DOE Design of experiments D-PBS Dulbecco’s phosphate buffered saline EDS Energy dispersive X-ray spectroscopy FIB Focused ion beam GI-XRD Grazing incidence X-ray diffraction HA Hydroxyapatite HT Heat treatment ICP-AES Inductively coupled plasma – atomic emission spec-

troscopy MRSA Methicillin resistant Staphylococcus aureus MRSE Methicillin resistant Staphylococcus epidermidis MIC Minimum inhibitory concentration NPs Nanoparticles �OH Hydroxyl radical PCA Photocatalytic activity PVD Physical vapor deposition ROS Reactive oxygen species SBF Simulated body fluid SEM Scanning electron microscopy Ti64 Ti-6Al-4V alloy TiO2 Titanium dioxide UV Ultraviolet

13

Introduction

Sometimes, accidents happen and vital parts of our bodies are in need of replacement or restoration. Other times, as increasingly seen on a global scale, the same parts are worn out by a lifelong service to its host. Thanks to the dedicated work of countless medical professionals, scientists and engi-neers, however, biocompatible materials, surgical procedures, medicines, and antibiotics have been developed to successfully mend this, significantly improving the quality of life for millions of people. Nevertheless, the com-plexity of our biological system and its intricate symbiotic relationship with bacteria can still cause materials, surgical procedures or treatments to fail. These challenges, along with longer life expectancies and higher demands from society, have resulted in a pressing need for improvements and the development of new advanced products and technologies.

Following surgical placement of an implant, a foreign body reaction oc-curs in which the human body tries to decide whether to accept or reject this alien object. If chosen wisely and the conditions are right, the material will be accepted and human cells will colonize the surface, integrating it with the surrounding tissue. Unfortunately, competing forces in the form of bacteria may also colonize the surface, prohibiting a proper integration and causing an infection. In such cases, the implant starts to lose its function and a num-ber of countermeasures have to be taken to save the implant. Antibiotic treatment has long been the go-to solution, but two major traits of bacteria make it difficult beat infection: their ability to form biofilm and their adapt-ability, resulting in the development of antibiotic resistance.

The biomaterial most commonly used for replacing missing teeth and mending fractured bones is titanium, or titanium based alloys, largely due to its capacity to bind closely with bone in a process called osseointegration.1 Modern implants are carefully engineered and the surface is usually modi-fied to improve the osseointegrative properties. Now, a critical next step in the evolution of implant coatings is combining osseointegration with anti-bacterial functionality, which could further help the body accept the implant and potentially reduce the use of antibiotic drugs. But as will be demonstrat-ed in this thesis, solutions to this task are very elusive and currently the topic of much research.2,3

With the aim of contributing to advancing the knowledge surrounding this challenging issue, this thesis focuses on the development of two antibac-

14

terial strategies potentially applicable on titanium based implant surfaces: TiO2 photocatalysis and the incorporation and release of silver ions. The former can be seen as a latent on-demand strategy, applicable on implants that penetrate soft tissue such as dental implants or bone-anchored prosthe-ses, which is activated by UV light. The latter is a more active approach, applicable also on totally internal orthopedic implants, where antibacterial silver ions will act to defend the implant site from infection.

The more specific Aims and objectives of the thesis are presented in the next section, followed by a more thorough description of Titanium as a bio-material. The cascade of events that occur during Bacterial colonization and infection is described thereafter.

To provide the reader with sufficient understanding on how bacteria can be killed by UV irradiation on a semiconductor surface, a section describing Photocatalysis on TiO2 is included. The subsequent sections, Chemical oxi-dation with H2O2 and Synergies in the TiO2/H2O2/UV system describe the surface modification technique utilized for a large portion of the work pre-sented in this thesis, along with key findings related to bacterial killing by reactive oxygen species. The section Silver as an antibacterial agent intro-duces the second strategy explored in this thesis. This is followed by a brief overview of Combinatorial materials science and Physical vapor deposition, the concept and method by which titanium and silver were combined in this work. In the sections Composition-structure-property relationships and UV-induced SOS response, the main findings related to bacterial killing via sil-ver ion release are presented. Finally, some Concluding remarks are drawn and Future challenges presented. A description of the main Analytical tech-niques used during the work is also included at the end.

15

Aims and objectives

The overall purpose of the work presented in this thesis was to explore, evaluate, and propose potential solutions to the problem of implant associat-ed infections. Regarding dental implants, this problem is usually manifested as peri-implantitis, which is a complex and all too common disease where bone resorption gradually leads to implant failure. For orthopedic implants, infections acquired during or after implant insertion can obstruct or disturb proper tissue integration, making it a leading cause for implant failure.

Understanding the cause of infection is instrumental to developing a treatment or a preventive measure, as well as to recognize the limited room to maneuver without inadvertently causing host cell damage or reduced os-seointegration. The aims and objectives were thus, from a materials science perspective and with a clear application in mind, to develop TiO2 surfaces with antibacterial functions that could complement the already biocompati-ble nature of the material. Inherent to the scientific approach, adjacent dis-coveries relevant to the advancement of knowledge in the field were also pursued.

More specifically, Papers I, II, and III were aimed at developing a cost effective surface modification technique that would result in a photocatalyti-cally active TiO2 surface. Applied on dental implants, abutments, or Ti based dental crowns and bridges, such a surface would add an on-demand antibacterial function activated by UV light.

By using a combinatorial materials science approach, Papers IV and V were aimed at exploring the composition-structure-property relationships in gradient coatings of silver and titanium. Incorporating the right amount, size, and structure of silver species in a TiO2 coating could render an implant surface actively antibacterial while maintaining its biocompatibility.

16

Titanium as a biomaterial

Depending on the intended biomedical application, certain criteria have to be met by a biomaterial to enable a successful outcome. For hard tissue re-placements, mechanical properties such as strength, Young’s modulus and toughness are critical for a good match with the surrounding bone. Further, the material has to be biocompatible, i.e. it should in essence be non-toxic and perform with an appropriate host response in a specific situation.4 It should also be highly corrosion resistant since leaching ions can provoke allergic and toxic reactions. For articulating joint replacements, a high wear resistance is also essential as wear debris particles are known to cause sever-al adverse responses, including aseptic loosening of the implant.5

In respect to the abovementioned criteria, titanium and titanium based al-loys are fairly good biomaterials. They are not as cheap as stainless steels or as fatigue and wear resistant as cobalt chromium alloys, but when it comes to interactions with bone, titanium and titanium based alloys have a set of characteristics making it the material of choice for dental implants. First of all, titanium is a low-density element, allowing for strong lightweight con-structions. Secondly, it has a comparably low modulus, which reduces the adverse effects of stress shielding when implanted in bone.5,6 The term stress shielding refers to a gradual reduction in bone density that can occur around implants when normal loads are transferred from the bone to the implant. Since bone is a living tissue that remodels according to mechanical stimuli, a reduced load will result in a weaker bone. Thirdly and perhaps most im-portantly, it spontaneously forms a thin passivation film on its surface when exposed to air. This native titanium dioxide (TiO2) film results in excellent corrosion resistance and has been deemed largely responsible for the out-standing biocompatibility of early titanium implants.7 The native film, how-ever, is usually too thin (≈5 nm) and non-uniform to meet the specific de-mands in modern clinical applications. A number of surface modification techniques have therefore been developed to enhance the properties required for, e.g. osseointegration and/or antibacterial properties,7 some of which will be described and discussed further on. The construct of a dental implant, placed in the bone and equipped with a tooth-mimicking crown, is illustrated in Figure 1.

17

Figure 1. Illustration of a dental implant with an accompanying crown, placed in the jawbone.

Granted that the bulk properties of a biomaterial are adequate, the fate of an implant or device is decided by the events taking place at the interface with biological tissue. The development of an intimate contact with bone, i.e. osseointegration, is essential for dental and orthopedic implants, and all as-pects of surface characteristics, such as surface chemistry, surface roughness and surface topography have to be considered. If the immune system finds the surface hostile or in some way unfavorable, it will respond with a in-flammation that may become chronic and encapsulate the implant in fibrous tissue rather than facilitating anchorage directly to the bone.8 Such an out-come usually requires revision surgery, which involves significant costs and an increased risk for infection.

The biocompatible nature of TiO2 allows for osteoprogenitor cells to be recruited to the implant site and differentiate into osteoblasts, which can make new bone at or around the implant surface. Although some bioactive implant coatings, primarily of hydroxyapatite (HA), can facilitate direct chemical bonding between bone and surface, the bond holding an integrated implant in place is primarily of physical nature.7 This is why surface rough-ness parameters and topography, both at the nano- and microscale play an important role.9 Small features or porosities at the surface allows for indi-vidual osteoblasts to adhere and lay down new bone, while larger features can allow ingrowth of bone for better interlocking and anchoring.

18

Surface modifications of titanium for improved osseointegration therefore strive to optimize properties related to surface chemistry, crystallinity, roughness and topography, all of which affect the surface energy. The sur-face energy in turn dictates the initial adsorption of water molecules and proteins directly upon implant insertion, which is followed by cell adhe-sion.5,10 Some of the more common techniques that are applied to modify or coat titanium surfaces include anodic oxidation,11 sol-gel coating,12 ion im-plantation,13 chemical or physical vapor deposition,14,15 and plasma spray-ing.16 Chemical treatments using acids, alkaline solutions or hydrogen per-oxide are also available and present a flexible and cost effective alterna-tive.17,18 The same basic techniques can also be applied to equip the surface with additional features such as photocatalytic properties, or more direct antibacterial properties via incorporation of antibiotics or noble metals. Both these approaches will be discussed in further detail ahead.

In summary, titanium and its alloys have been trialed and tested as a bio-material for the better part of 50 years, and it still presents the most depend-able choice for dental and orthopedic implants.5 There are, however, many areas in which its properties can and need to be improved. For instance, low-er stiffness and improved wear resistance are crucial objectives for better articulating and load bearing joint replacements. More urgently though, as is explained in the next section, strategies for equipping titanium based im-plants with antibacterial features need to be developed and applied.

19

Bacterial colonization and infection

Unfortunately, the same surface properties that make an implant or device hospitable for human cells can also provide a welcoming environment for pathogenic bacteria. In an illustrative phrase coined by Gristina in 1987, the events that take place upon insertion of an implant is described as a “race for the surface”,19 in which the host cells and opportunistic bacteria compete for occupation of the new real estate. An inert, biocompatible material poses no significant threat to bacteria. On the contrary, it presents a habitable sub-strate for which bacteria have evolved and adapted to for millions of years. Tissue cells are therefore at a significant disadvantage already from the start, and a foreign surface yet unguarded by the immune system can quickly be-come conquered and covered in biofilm, the stages of which are illustrated in Figure 2. Once a bacterial community is firmly attached to a comparably immense implant, the regular immune responses such as phagocytosis be-come effectively impossible. For a successful outcome, it is therefore of utmost importance to eliminate as many routes of infection as possible, and then aid the host tissue cells in broadly defending the implant surface from bacterial colonization.

Figure 2. The five stages of biofilm development. Stage 1: Planktonic (free floating) bacteria adhere to the biomaterial surface. Stage 2: Cells aggregate, form micro colonies and excrete extracellular polymeric substances (EPS), i.e. slime. The at-tachment becomes irreversible. Stage 3: A biofilm is formed. It matures and cells form multi-layered clusters. Stage 4: Three-dimensional growth and further matura-tion of the biofilm, providing protection against host defense mechanisms and anti-biotics. Stage 5: The biofilm reaches a critical mass and disperses planktonic bacte-ria, ready to colonize other surfaces. Figure adapted from Monroe20 under the crea-tive commons license. Image credit: D. Davis.

20

An implant-associated infection can originate from many sources, and one of the principal means of reducing the risk is rigorous precaution during sur-gery. Proper ventilation, clothing and sterile instruments, along with disci-pline in the operating theatre can significantly reduce the risk of periopera-tive contamination, but no surgical site or implant surface is truly sterile.2,21 Even after successful surgical placement of totally internal implants, such as hip, knee or heart valve replacements, an infection can arise by hematog-enous spread of bacteria having entered the bloodstream elsewhere. Percuta-neous or permucosal implants and devices, which penetrate through skin or a mucous membrane, are at constant risk of infection as they are exposed to the external environment. Such implants and devices include various cathe-ters, sutures, fixation pins and screws, voice prostheses, but also bone an-chored prostheses and dental implants.

According to the Swedish Hip Arthroplasty Register (2013), the infection incidence after primary hip replacement surgery can be as low as 0.5-1%.22 Other complications such as aseptic loosening, however, contribute to a failure rate of 5% within 10 years. Revision surgeries are generally associat-ed with higher risks of infection and other complications, and require signif-icantly more resources than the primary surgery. A revision hip arthroplasty for infection can take five hours instead of three, result in two liters of blood loss instead of one half, require a month in the hospital instead of one week, and cost $100,000 instead of $20,000, all without guaranteeing an implant free of infection.23

The primary causative agents responsible for orthopedic implant-associated infections have been identified as Staphylococcus aureus and Staphylococcus epidermidis, reported present in up to 70% of clinical iso-lates.24 Both of these species are common around the human body, with S. aureus often residing in the nares and S. epidermidis living on the skin, not posing any significant threat but rather keeping other pathogens away. Trou-ble begins first when the skin barrier is broken, or when the immune system is compromised. Both of these species are strong biofilm formers and have developed resistance to a range of antibiotic drugs. Unfortunately, the methicillin resistant strains of S. aureus (MRSA) and S. epidermidis (MRSE) are often acquired at the hospital, causing difficult-to-treat infec-tions that add costs and patient discomfort, or worse.

Concerning dental implants, which are currently being placed at an annu-al rate of ten million globally,25 an estimated 7% suffer complete failure within 10 years.26 A majority of these failures are related to peri-implantitis, a disease generally caused by biofilm-producing oral pathogens that lead to inflammation and a progressive bone loss around the implant.27 Peri-implantitis has been reported to affect around 10% of implants and 20% of patients27 (thus having multiple implants), and is a major source of concern for patients and dentists alike. Although S. aureus and S. epidermidis are

21

causative agents also for peri-implantitis, the etiology of biofilms formed in the oral cavity and on dental implants is usually different from biofilms found on orthopedic implants. Species that cause periodontitis (Porphy-romonas gingivalis, Aggregatibacter actinomycetemcomitans etc.), which is a common reason for tooth loss and prescription of a dental implant in the first place, are often also involved in the development of peri-implantitis.28 Streptococcus mutans, which is another common species around teeth and responsible for caries and plaque formation, have also been found to colo-nize dental implant surfaces.29 Peri-implantitis, as well as other implant-associated infections, do not only pose an immediate threat to the implant and the surrounding bone, but can also cause bacteremia if viable bacteria enter the bloodstream. This can induce sepsis, or cause infection and biofilm formation on other, perhaps more sensitive implanted devices such as artifi-cial heart valves.26

Monitoring and controlling the events that occur on the surface of an im-plant is thus of great importance, as down stream consequences of microbial colonization may result in serious complications. As implied in Figure 2, the further a biofilm is allowed to develop, the more difficult it becomes to elim-inate. A primary strategy to avoid implant-associated infection is thus to inhibit planktonic bacteria from adhering in the first place. This can be real-ized by applying hydrophilic anti-adhesive coatings, based on, e.g. polyeth-ylene oxide (or polyethylene glycol).30 However, such surfaces tend to also inhibit protein adhesion, which limits cellular attachment and makes for poor tissue integration. On the other hand, optimizing the nanotopography of a titanium surface has been shown to reduce bacterial adhesion, while still promoting certain protein adhesion.31 Another strategy is to equip the sur-face with its own active defenses, which can be done by doping or loading it with antimicrobial substances such as antibiotics, chitosan, or silver. These substances can either be leaching from the surface to kill bacteria preemp-tively, or be more or less immobilized on the surface to kill bacteria upon contact, prior to forming a biofilm. For antibiotic loading, however, the risk of eliciting bacterial resistance has to be considered. Silver ions, which will be discussed in further detail ahead, have a lower propensity to develop bac-terial resistance, and can also be effective against methicillin-resistant strains.32 A more passive approach, which can be viewed as an on-demand antibacterial strategy, is arming the surface with photocatalytic properties. This approach requires a certain set of surface characteristics, and is only applicable for UV-light accessible areas, i.e. percutaneous or permucosal implants. The theoretical background and basis for this approach is devel-oped in the next section.

22

Photocatalysis on TiO2

For a heterogeneous photocatalytic reaction to take place, two principal things are needed: (i) a material having a band gap, and (ii) a photon with sufficient energy. In essence, this translates to (i) a material where electrons cannot move freely without an input of energy, i.e. a semiconductor, and (ii) a lamp that emits light at short enough wavelengths.

All elements and materials consist of atoms, which are built on protons and neutrons that constitute the nucleus. Orbiting the nucleus is a set of neg-atively charged electrons, leveling the positive charge of the protons. Every orbital (there can be several) represents a specific energy level on part of the electron. Depending on the interatomic spacing in a material, neighboring atoms can either share electrons by having overlapping orbitals/electron states, as in metals, or be separated so an energy gap appears between the electron states, as in insulators and semiconductors. This energy level sepa-ration is called a band gap, and constitutes the forbidden energy region that extends from the top of the electron filled valence band to the bottom of the empty conduction band. In insulators, the band gap is wide, whereas in sem-iconductors it is narrower. This means that if electrons in a semiconductor are provided with a little energy, they can jump the band gap to occupy the conduction band, making the material conductive. Compared to pure silicon (Si), which has a band gap of 1.1 eV, crystalline anatase TiO2 is considered a wide band gap (3.2 eV) semiconductor, meaning that more energy is re-quired to excite its electrons to the conduction band.

In TiO2 photocatalysis, the objective is not to conduct electricity, but to generate electron-hole pairs. When generated, the electron-hole pair is at an energetically very unfavorable state, making it short lived (on the order of a nanosecond). In this short lifetime it can either recombine, to release the excess energy as heat, or transfer its charge to molecules adsorbed on the surface of the semiconductor, primarily oxygen (O2), water (H2O), or hy-droxyl groups (OH–). For efficient photocatalysis, the recombination rate should be low, and the charge transfer to adsorbed species should be high. Primary reactions in this process are given in Equations 1-5, and the prod-ucts superoxide (O2

–), hydroxyl radicals (�OH), and hydrogen peroxide (H2O2) are classified as reactive oxygen species (ROS), which can exert oxidative stress on nearby organisms, such as bacteria.

23

TiO2 + hv → e– + h+ (1) O2 + e– → O2

– (2) H2O + h+ → �OH + H+ (3) OH– + h+ →  �OH (4) �OH + �OH → H2O2 (5)

The entire process, which is illustrated by a simple model in Figure 3, is catalyzed by an input of energy in the form of UV light. The anatase form of TiO2 has a band gap of 3.2 eV, which corresponds to the photon energy of light at 385 nm. This wavelength is considered near-UV light, in the UV-A range, and lies close to the visible spectrum (400-700 nm).

The bactericidal activity of photocatalytic TiO2 has been attributed to li-pid peroxidation that damages the bacterial cell membrane, leading to lost respiratory activity and ultimately cell death.33 The lipid peroxidation, itself a free radical chain reaction, is initiated by �OH and propagates by reacting with polyunsaturated phospholipids, which are essential components of the cell membrane. Out of the ROS produced in the photocatalytic reaction, �OH is the most reactive, and therefore the most lethal for bacteria. Howev-er, due to its reactivity, it can never travel far from the surface where it was produced, and the bactericidal effect due to �OH production becomes very local. Other ROS, such as H2O2, are less reactive but have been shown to complement �OH by acting on bacteria at a distance.34

Figure 3. A simplified illustration of the photocatalytic disinfection process on ana-tase TiO2. Incoming UV irradiation promotes an electron (e–) to the conduction band, leaving a hole (h+) behind in the valence band. Adsorbed O2 can accept an electron and become reduced to O2

–, while H2O can donate an electron to fill the hole, thereby becoming oxidized to �OH. These, and other potential ROS, can react with organic molecules of nearby bacteria, mineralizing them to CO2 and H2O.

24

Chemical oxidation with H2O2

As previously mentioned, many surface modification techniques have been developed to enhance the osseointegrative properties of implants, and/or to deter implant-associated infections. In many instances, photocatalytic prop-erties can coincide with bioactivity. Anatase TiO2 coatings have been shown to promote initial osteoblast adhesion,35 but also to reduce bacterial adhesion when compared to native TiO2.36 Among the more cost effective methods to modify the surface of titanium is chemical oxidation, which is primarily done by treatment in H2O2 or in alkali solutions, resulting in an amorphous titania gel coating.17,18 Such wet chemical treatments also permit a uniform, controlled coating of implants or devices with complex shapes, where other techniques are often limited to a “line-of-sight” deposition or oxidation.

Chemical oxidation with H2O2 was used to modify the surface of com-mercially pure titanium (cpTi) and/or Ti-6Al-4V alloy (Ti64) in Papers I, II, and III. Previous work had shown that treatments with H2O2, followed by either heat treatments18,35,37,38 or aging in H2O39-42 could induce crystalliza-tion of the amorphous titania gel into anatase, resulting in improved bioac-tivity. For Paper I, it was therefore hypothesized that such treatments could also be used to produce a photocatalytic surface. The purpose of the study was to investigate if a low temperature (80°C) H2O2-oxidation and H2O-aging process of Ti64 could be optimized for adequate photocatalytic activi-ty within a limited treatment time. In Paper II, cpTi and Ti64 substrates were again H2O2-oxidized and H2O-aged according to previous optimiza-tion, but were also subjected to heat treatments to improve crystallinity and stability of the photocatalytic films. In Paper III, the optimized TiO2 coat-ing was tested against bacteria in conjunction with UV light and a small addition of H2O2.

The simple steps of the H2O2-oxidation and H2O-aging process, as well as an illustrative image of the resulting surface morphology, are given in Fig-ure 4.

25

Figure 4. Illustration of the simple H2O2-oxidation and H2O-aging process of titani-um samples, which was used to produce TiO2 films in Papers I, II, and III. The resulting surface morphology shown to the right was obtained after 18 h in 30 wt% H2O2 followed by 54 h in H2O, all held at 80°C.

Photocatalytic activity The optimization of the process, carried out in Paper I, was guided by a design of experiments (DOE) software package (MODDE, Umetrics AB, Umeå, Sweden), where the input variables were limited to oxidation time in H2O2 (t1) and aging time in H2O (t2). The output was a quantification of each resulting surfaces’ photocatalytic activity (PCA). The quantification method is elaborated on in the Analytical techniques section, under Rhodamine B degradation. Briefly, the photocatalytically-induced degradation of an or-ganic dye was monitored over time, yielding a degradation rate constant (k) reflecting the PCA of each surface. From the DOE, 16 experimental runs were carried out, varying t1 from 0.5 to 18 h, and t2 from 1.5 to 54 h. Un-treated samples were used as control, while samples with t1 = 24 h and t2 = 72 h were used to validate the statistical model.

As time is always of the essence, an explicit goal of the study in Paper I was to develop a simple yet quick modification method of Ti64 for high PCA. One of the intended areas where photocatalytic disinfection can be of use is on Ti-based dental crowns and bridges, which are always custom made to fit a specific patient. Extended oxidation and aging was therefore unwanted, as it would result in longer production lead times. As was shown, however, the PCA depended almost linearly on the total treatment time.

Figure 5 shows the degradation curves of rhodamine B in presence of se-lected samples under UV illumination. Quantification of the slope of each curve gives the rate constant (k), which was used to plot the contour graph in Figure 6a, as well as the relationship with total treatment time in Figure 6b. It was shown that PCA increased with longer H2O2-oxidation times, as well as with longer H2O-aging times, but neither direct (t1 and t2) nor interactive (t1 × t2) effects were shown statistically significant at p < 0.05. However,

26

when adding t1 and t2 together for a total treatment time, the PCA showed a statistically validated linear dependence. In conclusion, an extended treat-ment was needed for a satisfactory result.

Figure 5. Graph showing the photocatalytic degradation of organic dye rhodamine B in presence of Ti-6Al-4V samples, oxidized in H2O2 and aged in H2O for durations (hours) indicated in the legend.

Figure 6. The contour plot in (a) shows a qualitative increase in PCA with both increasing time in H2O2 (t1) and in H2O (t2). However, the direct (t1 and t2) and in-teractive (t1 " t2) effects on PCA could not be statistically validated at p < 0.05. On the other hand, plotting PCA against the total treatment time (t1 + t2), as done in (b), yielded a linear regression with p < 0.05.

27

Surface characteristics The structure and morphology of a Ti surface after extended H2O2-oxidation and H2O-aging, compared to the native oxide is illustrated in Figure 7. Thickness of the titania gel layer has been shown to depend almost linearly on treatment time,18 which results in an increasingly porous structure with greater surface area. The coatings produced in Paper I were found to be amorphous or poorly crystalline, even after extended aging in H2O. This led to the conclusion that an increase in porosity and surface area, rather than anatase content, was responsible for the increase in PCA. The true nature of the coatings in relation with PCA would be further revealed in Paper II.

Figure 7. Illustration of the native oxide film formed naturally on Ti (above), and the thicker, more porous film formed after 24 h chemical oxidation in H2O2 and aging for 72 h in H2O (below). The Ti/O ratio and density of the oxide increases with proximity to the Ti substrate. Illustration adapted from Variola et al.43

28

Synergies in the TiO2/H2O2/UV system

As was noticed in Paper I, the H2O2-oxidized and H2O-aged Ti64 samples showed photocatalytic properties even though they were essentially amor-phous. From the Photocatalysis on TiO2 section, we know that for a true photocatalytic reaction to take place, there has to be a band gap and genera-tion of electron-hole pairs, which in turn requires a crystalline structure. What then, caused the PCA on these samples?

Some experiments, where the samples were tested multiple times to check repeatability, also revealed a staggering decline in PCA with repeated use. This decline is displayed in Figure 8, where four consecutive curves repre-sent the samples’ capacity to degrade rhodamine B in a cyclic fashion. The inset in Figure 8 shows the rate constant k for each repetitive run, which demonstrated that the PCA of the same samples was quickly reduced to ap-proximately one tenth of its initial value. Considering the samples were not crystalline, and that longer H2O2-oxidation times had yielded higher PCA, a reasonable explanation was that Ti-peroxy species left from the H2O2 treat-ment contributed to the initial activity, but eventually became consumed. The details of this photo-activated reaction, however, were not known.

Figure 8. Repeated tests on the same samples, H2O2-oxidized and H2O-aged for 24 h and 72 h respectively, revealed a gradual decline in PCA due to consumption of titanium superoxide radical species associated to the surface.

29

Added benefits with added H2O2 Previous work by Tengvall et al.44,45 had shown that stable titanium superox-ide radicals (Ti(IV)O2

–) can reside in a Ti-peroxy gel formed by Ti-H2O2

interactions. A later study by Wu et al.46 suggested that such superoxide groups could be responsible for an initially high PCA on H2O2-oxidized titanium, but as the superoxide radicals became consumed, the PCA dimin-ished. These findings support the abovementioned theory, but do not fully explain the process by which superoxide or other radical species are activat-ed by UV light. In order to tackle the issue of diminishing PCA, yet trying to make use of the radicals residing in the titania gel layer after H2O2-oxidation, Paper II focused on stability and potential synergy effects in the TiO2/H2O2/UV system.

Having H2O2 present during a photocatalytic reaction, which can either be residual from the oxidation process, or actively added prior to the reaction, may be beneficial as H2O2 can scavenge conduction band electrons (e–) to yield �OH and OH–,47,48 as shown in Equation 6. H2O2 can also absorb UV light directly to yield two �OH,49 as shown in Equation 7. The reaction in Equation 6 requires generation of e– from a photocatalytic process, while the photolytic splitting of H2O2 in Equation 7 generally requires high energy UV light (λ < 300 nm).50 Having an excess of H2O2, however, may also inhibit PCA according to Equation 8, which produces the comparatively less reac-tive perhydroxyl radical (HO2�).

H2O2 + e– → �OH + OH– (6) H2O2 + hv → 2�OH (7) H2O2 + �OH → H2O + HO2� (8)

For improved stability and PCA of the surfaces, crystallization heat treat-ments (HT) at 400 – 600°C were conducted on H2O2-oxidized and H2O-aged cpTi and Ti64 samples in Paper II. At or above 500°C, HT demonstrated to induce anatase formation on cpTi, while both anatase and rutile polymorphs of TiO2 was formed on Ti64. The PCA of the surfaces, quantified via degra-dation of rhodamine B and the rate constant k in Figure 9, was shown to increase with increasing HT temperature, and could be correlated with de-gree of anatase formation on the samples. The highest noted value, for cpTi H2O2-oxidixed and H2O-aged for 24 h and 72 h, respectively, then HT at 600°C, came close to matching the initial value of the non-heat treated sam-ples in Figure 8. The PCA of the heat treated samples were also shown to remain stable over several test cycles. It was further noted that the surface morphology of the samples remained essentially unchanged after HT, alt-hough a densification of similar coatings have been reported in the litera-

30

ture.46 A full treatment, i.e. H2O2-oxidation, H2O-aging and HT at 600 °C, was also found to yield a higher PCA than when skipping the H2O-aging, which indicates a significance of the surface events taking place during this step.

Figure 9. PCA of heat treated samples as a function of heat treatment (HT) tempera-ture. The legend indicates substrate material and the duration (hours) of H2O2-oxidation and H2O-aging.

The H2O-aging step has previously been shown to generate an abundance of hydroxyl groups (Ti-OH) on the surface, which improves bioactivity.51 Here, the concept of bioactivity implies that bone-like apatite forms spontaneously on the surface when it is immersed in a simulated body fluid (SBF). It has been shown that Ti-OH enhances the apatite deposition process by attracting Ca2+ ions present in the SBF, which in turn attracts HPO4

2- ions to form an amorphous calcium phosphate layer, later transforming into crystalline apa-tite.52,53 This surface property is an advantage for any material meant to inte-grate with bone. Moreover, having Ti-OH present can benefit the photocata-lytic reaction by increasing oxygen adsorption,54 which results in an in-creased production of O2

– according to Equation 2. Having more electron scavengers (O2, H2O2 etc.) present at the surface reduces the electron-hole recombination process, which then increases the amount of h+ available to produce �OH, e.g. via Equation 3.55

The consequence of adding H2O2 during the photocatalytic reaction is shown in Figure 10. Compared to the previously reported PCA (UV/TiO2: k ≈ 5×10-3 min-1), the degradation rate was approximately five times faster with 3 wt% H2O2 present (UV/H2O2/TiO2: k ≈ 25×10-3 min-1). This clearly demonstrated the capacity and potential synergy effects in the TiO2/H2O2/UV system. Removing the TiO2 samples to investigate photolytic

31

degradation (UV/H2O2) also revealed a substantial effect, indicating that splitting of H2O2 according to Equation 7 occurred at least to a certain de-gree, even though the incoming wavelength was > 300 nm. One should however consider that the photolytic reaction occurs throughout the entire volume of the rhodamine B solution, whereas the photocatalytic reaction is confined to the surface of the TiO2 samples.

The considerably lower degradation rates noted for TiO2, H2O2, and H2O2/TiO2 when removing the UV illumination, also in Figure 10, demon-strated that the UV light activation contributes significantly to the capacity of the system. Nevertheless, a notable increase in the degradation rate was seen for the H2O2/TiO2 system compared to H2O2 or TiO2 alone, indicating an autocatalytic reaction at the surface.

Figure 10. Normalized degradation curves of rhodamine B under a variety of condi-tions, as indicated by the legend. Inset shows the quantified PCA of each condi-tion/system. According to increasing value, the PCA ranked as follows: TiO2 < H2O2 < H2O2/TiO2 < UV < UV/TiO2 < UV/H2O2 < UV/H2O2/TiO2. Non-heat treated TiO2 surfaces that were H2O2-oxidized and H2O-aged for 24 h and 72 h, respective-ly, were used for the measurements.

32

Stability of the system To investigate the stability of both heat treated and non-heat treated samples in the TiO2/H2O2/UV system, cyclic tests, like the ones presented in Figure 8, were conducted. The results are shown in Figure 11, where the solid lines represent repeated tests with Ti64 samples H2O2-oxidized and H2O-aged for 24 h and 72 h, respectively. The dashed lines represent identically H2O2-oxidized and H2O-aged samples, but which were also heat treated for 1 h in air at 600°C.

Figure 11. Repetitive tests of H2O2-assisted photocatalysis on heat treated (600 °C) and non-heat treated TiO2 samples. The PCA, as quantified by the rhodamine B degradation rate constant k (×10–3min–1), is indicated for each run.

Both heat treated and non-heat treated samples demonstrated high PCA in the TiO2/H2O2/UV system, and the activity was essentially maintained through four repeated runs. However, the non-heat treated samples still out-performed their heat treated counterparts.

It is known that at temperatures above 300 °C, Ti-OH and radical species stabilized in the Ti-peroxy gel are effectively decomposed and crystalliza-tion of the layer starts to occur.56 Hence, PCA of the heat treated samples was likely due to a heterogeneous photocatalytic reaction on crystalline ana-tase, enhanced by H2O2 which is a more efficient electron scavenger than O2.57,58 On the other hand, PCA of the non-heat treated samples relied less on crystalline anatase, and more on photo-activation of already present radi-cal species. The abundance of Ti-OH would also aid any photocatalytic reac-tion at the surface. In both cases, photolysis of H2O2 occurred in the back-ground and contributed to the total effect.

33

In summary, the experiments in Paper II revealed interesting synergies be-tween TiO2, H2O2, and UV light, which were not considered in Paper I. The fact that less crystalline samples, if equipped with Ti-OH and Ti-peroxy radicals, generated a greater effect in the TiO2/H2O2/UV system than sam-ples with anatase did, demonstrates an interesting phenomenon. Granted that the incoming photon energy is sufficient, similar systems have shown the photolytic contribution (UV/H2O2) to dominate the degradation process, while an excess of H2O2 can inhibit pathways of heterogeneous photocataly-sis (UV/TiO2) according to, e.g. Equation 8.57 As the heat treated samples were more dependent on heterogeneous photocatalysis, they were more sus-ceptible to inhibitory effects of H2O2. The non-heat treated samples, on the other hand, could benefit from surface bound Ti-peroxy radicals46 and poten-tially have them restored when adding 3% H2O2 for the assisted reaction.

The studies in Papers I and II utilized an organic dye (rhodamine B) as a model substance to evaluate PCA, and although the same radical species that degrade the dye would be responsible for oxidation and mineralization of bacteria, the in vitro performance of the surfaces and systems could not be known for certain. Hence, the work in Paper III was directed against bacte-ria.

Effects against bacteria After concluding in Paper II that the highest PCA on Ti64 substrates was obtained after 24 h H2O2-oxidation and 72 h H2O-aging, and that this activi-ty could be significantly enhanced and stabilized by adding a small amount of H2O2, these were the conditions tested further in Paper III.

As mentioned in the Bacterial colonization and infection section, one of the most common pathogens causing implant-associated infections is S. epi-dermidis. Another bacterial species, common in plaque and a leading cause for caries, is S. mutans. Both of these species are Gram-positive, meaning that they lack an outer membrane present in Gram-negative bacteria. Addi-tionally, S. mutans is catalase negative, i.e. missing the enzyme responsible for decomposing hazardous H2O2, which makes it more sensitive to H2O2 exposure.

Both S. epidermidis and S. mutans are target species for the intended ap-plication of the system, i.e. debridement/cleaning of a dental or otherwise tissue penetrating implant surface, which could be at-risk or already afflicted with bacterial infection. Testing the TiO2/H2O2/UV system against these clinically relevant strains would therefore give an interesting look into the different modes of bactericidal action of the system.

One of the challenges in evaluating a system where different modes of ac-tion are possible is designing a test that takes all these actions into account,

34

while making it possible to discern one bactericidal action from another. Such a test was developed for Paper III, and later adapted for bacteria tests in Papers IV and V. Termed a direct contact test (DCT), the idea was to optimize an adhesion-type test for short-range, UV-activated bactericidal surfaces. Briefly, a small amount of bacterial suspension (5-10 µl) is added to the sample surface, and the entity is incubated at 37 °C for a limited time to allow planktonic bacteria to settle on the surface while some of the sus-pension liquid is evaporated. Depending on the antibacterial function of the surface (contact killing or UV activated), the bactericidal action is initiated either immediately or when turning on the light. Quantification of adhered and viable bacteria can then be done after a certain time point, either by measuring fluorescence, as done in Papers III and V, or by counting colo-ny-forming units (CFUs), as done in Paper IV. Further details on the DCT are given in the Analytical techniques section, under Bacterial viability as-says.

For Paper III, the direct effect of H2O2 on S. epidermidis and S. mutans was first evaluated, and the result is shown in Figure 12. It was shown that both species are sensitive to H2O2 exposure but, at levels ≥ 0.1 wt%, S. mu-tans suffered greater losses. This was attributed to the fact that S. mutans lacks the catalase enzyme, which makes it unable to decompose H2O2. At 3 wt% H2O2, which is considered safe for human use and is routinely applied as an antiseptic in dentistry and on wounds,59 some 90% reduction in viabil-ity was seen for both species after 15 min exposure. In order to resolve syn-ergy effects in the TiO2/H2O2/UV system, however, the intermediate dose (0.1 wt%) was used for further tests.

Results with H2O2-oxidized and H2O-aged Ti64 samples, 0.1 wt% H2O2, and UV light at 365 ± 10 nm in different combinations are presented in Fig-ure 13. The control, denoted Ti in the figure, was a non-oxidized Ti64 sur-face. In these experiments, 5 µl of bacterial suspension, with or without a supplement of 0.1 wt% H2O2, was spread onto sample surfaces and immedi-ately irradiated with UV light while being incubated at 37°C for 10 or 20 min. The UV light, intensity adjusted to 1.5 mW/cm2, was applied right away since the action of H2O2 starts immediately upon mixing with bacteria.

Compared to the control surface, 20 min exposure to the TiO2/H2O2/UV system caused 99.7% and 98.9% viability reductions of S. epidermidis and S. mutans, respectively. Significant reductions were also seen with photoca-talysis (TiO2/UV) and H2O2 exposure (TiO2/H2O2) alone. Interestingly, the H2O2-oxidized and H2O-treated surface (TiO2) also showed antibacterial properties without UV activation, indicating that surface bound Ti-peroxy radicals remaining from the oxidation process had an effect on nearby bacte-ria.

35

Figure 12. Direct effect of different H2O2 concentrations on bacterial viability, ana-lyzed after 15 min incubation at 37 °C. 100% viability corresponded to 7.8×105 CFU/ml S. epidermidis, and 7.3×105 CFU/ml S. mutans.

Figure 13. Viability of S. epidermidis and S. mutans bacteria after 10 and 20 min exposure to individual and synergetic effects of TiO2 discs, UV light and 0.1 wt% H2O2. In the assay, 100% viability corresponded to 1.04×106 CFU/ml S. epidermid-is, and 1.28×106 CFU/ml S. mutans. The asterisk (*) indicates a significant differ-ence from Ti at p < 0.05. The double asterisk (**) indicates a significant difference from both Ti and TiO2 at p < 0.05.

The results obtained from the bacteria tests in Paper III, shown in Figure 13, correlate well with the rhodamine B degradation data recorded in Paper II. It is obvious, however, that bacteria are more sensitive than rhodamine B to H2O2 exposure. Nevertheless, using an organic dye rather than bacteria suspensions to optimize the TiO2 surface and devise a system for adequate ROS production was shown successful.

The challenge that followed, i.e. developing a rather simple bacterial via-bility assay that accounts for multiple short-range bactericidal actions, was also quite successful. Through the assay, it was demonstrated that a TiO2 surface, at-risk of becoming colonized by biofilm-forming bacteria, could be effectively disinfected by a low concentration H2O2 and low intensity UV light treatment within an acceptable timeframe. Increasing the H2O2 concen-tration (up to 3 %), as well as the UV light intensity, especially with a device designed to shield sensitive host tissue, could also significantly increase the efficiency of the treatment in a clinical setting.

36

Silver as an antibacterial agent

Silver nitrate (AgNO3) has been applied to treat wounds, burns and infec-tions for centuries, but its use diminished significantly with the introduction of penicillin in the 1940s.60 Twenty years later, it returned in the form of silver sulfadiazine, a combined silver/antibiotic cream, to manage burns. More recently, however, silver wound dressings have returned yet again to counter the emergence of antibiotic resistance.61 Along the way, some more questionable medicinal uses of nanoparticulate or colloidal silver have also been exploited.62,63

The primary use of silver has thus far been as a topical agent, but incorpo-ration of silver nanoparticles (AgNPs) in implant coatings or on medical devices is emerging as a viable antibacterial strategy. Silver-hydrogel coat-ings on urinary catheters are already in clinical use and have significantly reduced the occurrence of urinary tract infections.64,65 Orthopedic and dental implant coatings containing silver or AgNPs are yet to be introduced in a wide clinical setting, but early trials and intense research efforts suggest that it is only a matter of time.66-68

One of the major hurdles to be overcome is the question of dose. For both implant coatings and wound dressings, there has been a lack of standard procedures for determining efficient release rates and minimum inhibitory concentrations (MIC), largely due to the broad scope of technologies ap-plied.60,69 It is also worth noting that the ideal Ag+ release rate differs widely between different applications. For wound dressings, Ag+ concentrations up to 100 ppm are applicable,60 whereas a suitable Ag+ release from an implant surface lies in the ppb range,70,71 i.e. three to four orders of magnitude lower.

Mechanism of action In its ground state, silver is a rather stable noble metal. Nevertheless, small amounts of ions are constantly released in aqueous solutions, and it is the Ag+ ion that is responsible for the bactericidal effect.72

In a study by Feng et al.,73 it was shown that Ag+ interferes with bacterial DNA molecules of Escherichia coli and S. aureus, causing DNA condensa-tion that result in lost replication abilities. The same study also suggested that Ag+ interacts with essential thiol groups, inducing the inactivation of

37

proteins and critical enzymes for the respiratory chain. Similar results have been shown for a wide range of bacterial species and strains, including MRSA, making silver a broad-spectrum antibacterial agent. 32,63 These bac-tericidal effects are enhanced by the use of AgNPs, which, by means of their small size can release more Ag+ due to an increased surface to volume ratio, and can also disturb or penetrate the bacterial cell membrane to induce fur-ther damage.74,75

Aside from the bactericidal effects of Ag+, nanoparticles (NPs) in general have been reported toxic to bacteria.76 Electrostatic interactions enable NPs to attach to, and disrupt the integrity of the bacterial cell membrane, and similarly to photocatalytic disinfection, NPs can induce oxidative stress on bacteria by triggering the production of ROS.76,77

In suitable doses, silver presents little or no cytotoxic effects.78-80 It can also quite easily be deposited on titanium or other biomaterials via a range of established techniques, such as sol-gel, 81,82 plasma immersion ion implanta-tion,68,83 or various physical vapor deposition (PVD) techniques such as magnetron sputtering.71,84,85 As an antibacterial agent with low propensity to develop resistance in bacteria, silver has therefore gained momentum against antibiotics when developing implant coatings to reduce infection.86 Ques-tions still remain, however, on what is a suitable dose and how silver is best incorporated to provide an adequate Ag+ release for antibacterial effect, while still maintaining good osseointegration properties and avoiding poten-tial cyto- or nanotoxicity.

In the work leading up to this thesis, silver and titanium was co-deposited in a reactive environment by means of a combinatorial synthesis approach. The concept of combinatorial materials science is further explored in the next section. The resulting Ag-Ti oxide gradients were characterized and evaluated for antibacterial properties in Papers IV and V.

38

Combinatorial materials science

Conventional materials research and development is typically conducted by synthesizing and testing a series of samples within a predetermined design space. Further optimization to meet specific property demands is then done by iterative steps where the input parameters are altered individually. This “trial-and-error” approach is a scientifically sound model, but the process can be tedious and the knowledge gained does not necessarily extend beyond our predictive abilities.

An alternative approach to discovery and optimization of new com-pounds, structures, alloys or molecules is combinatorial materials science. The concept revolves around rapid synthesis and high-throughput screening of different material combinations, which helps to build materials libraries and to find qualitative trends in composition-structure-property relation-ships.87,88 Further exploration or optimization can then be done in areas of interest.

Considering the complexity of just binary or ternary inorganic systems, which can be complemented with oxides and nitrides, it becomes clear that our understanding of possible material combinations is far from complete. The same holds true for organic molecules. Hence, if the goal is to discover new materials or efficiently optimize the properties of known ones, the methods of doing so have to cover more parameters and a larger design space than any conventional method.89

This approach can be applied by a number of means, such as automated fabrication and high-throughput screening of polymeric microarrays for cell studies, stem cell growth,90,91 or comprehensive surface characterization.92 Various sputtering or deposition techniques can also be used to create com-positional or structural gradients between inorganic materials,93,94 which have been applied to produce materials libraries for a wide range of applica-tions.95,96

39

An illustrative example of a binary concept system, synthesized and evaluat-ed by traditional and combinatorial means, is given in Figure 14. The obvi-ous advantage of the combinatorial approach is that only a single sample is used, saving both time and resources. Another feature is that, with a contin-uous gradient any number of points can be evaluated, and potential unin-tended variations between multiple samples due to process parameters can be avoided. Creating combinatorial arrays of materials or molecules also significantly enhances the probability of making new discoveries, which has been shown for both organic and inorganic materials, as well as for pharma-ceutical drugs.95,97,98

Figure 14. Illustration of two different approaches to materials research and devel-opment in a binary system: traditional and combinatorial. The combinatorial ap-proach strives to produce a “library” sample that contains the compositional or structural material varieties of interest, rather than producing multiple samples with discrete compositions or structures.

A binary gradient between silver and titanium oxide, like the system illus-trated in Figure 14, was synthesized by PVD, characterized and evaluated for antibacterial properties in Paper IV. Further development and evaluation of how incorporation of silver could potentially affect the photocatalytic properties of TiO2 led to the results presented in Paper V, which demon-strated some intriguing secondary effects of UV illumination on multifunc-tional coatings.

40

Physical vapor deposition PVD involves the physical removal of atoms from a source material, called a target, and the deposition of said atoms as a coating on a designated sub-strate. A common PVD technique is magnetron sputtering, in which a plas-ma is generated from ionization of inert gas atoms, usually argon (Ar). The Ar+ ions in the plasma are then accelerated towards the target by a magnetic field, and physically eject atoms from the target material. The ejected (sput-tered) atoms then journey through the plasma to condense on the substrate, as well as on any other surface in the sputtering chamber, to form a coating. Upon depositing on a surface, there is very little energy remaining to form a thermodynamic equilibrium, meaning that atomic migration to form a crys-talline structure is limited. Crystallization of the coating can however be done by heating the substrate, either during or after the deposition.

Aside from the sputtering gas (Ar), the process is performed under vacu-um. If non-metallic compounds such as carbides, nitrides or oxides are to be deposited, an additional gas (CH4, N2 or O2) is added. The process is then termed reactive magnetron sputtering, and the reactive gas is ionized along with Ar in the plasma and can react with metal atoms sputtered from the target to form, e.g. an oxide coating on the substrate.

In Papers IV and V, a custom-built combinatorial magnetron sputtering system was used to co-deposit silver and titanium in a reactive (O2) envi-ronment. By means of the geometrical set-up illustrated in Figure 15, binary, combinatorial samples containing a compositional and structural gradient between silver and titanium oxides were produced. To facilitate uniform depositions, and handy characterization and analysis, the coatings were de-posited on silicon wafers.

Figure 15. Illustration of the combinatorial magnetron sputtering system used to deposit binary gradients between silver and titanium. The geometrical set-up, with opposing targets facing the substrate at a 45° angle, allows for deposition of a con-tinuous gradient with enrichment in Ag and Ti close to each respective target.

41

Composition-structure-property relationships

As previously established, titanium is a biocompatible material suitable for hard tissue replacement, whereas silver is intrinsically antibacterial due to leaching of Ag+ ions. These characteristics are also enhanced for their re-spective oxides. Hence, in an attempt to map and characterize the composi-tion-structure-property relationships that govern the bioactive features of these materials, while demonstrating the benefits of a combinatorial ap-proach, a single Ag-Ti oxide gradient sample was synthesized and evaluated for Paper IV. By co-sputtering silver and titanium in a reactive environ-ment, rather than, e.g. depositing AgNPs on a prefabricated TiO2 surface, the potential for discovering new structures or features would also be improved.

Gradient thin film characteristics The crystalline structures at extreme positions of the Ag-Ti gradient sample produced in Paper IV are given in Figure 16. Spectrums intermediate of these were found along the gradient.

Figure 16. Grazing incidence X-ray diffraction patterns of the gradient Ag-Ti oxide thin film, showing presence of metallic Ag and traces of rutile TiO2.

The majority of crystalline content on both ends of the gradient was metallic silver, while signs of rutile TiO2 was found towards the titanium end. Ag-doping of TiO2 has previously been shown to inhibit the growth of anatase in

42

favor of rutile.99 The significant difference between metallic content of silver and titanium is explained by the oxygen affinity of titanium, causing it to deposit as amorphous or poorly crystalline oxide. The noble silver, however, is less prone to oxidation and could deposit in its metallic form. By refrain-ing from a crystallization heat treatment, the as-deposited metastable struc-ture could be preserved to reveal any potentially new materials with benefi-cial properties.

An example of a structure unlikely to have remained after heat treatment is shown in Figure 17, which depicts the morphology and cross-sectional view of the gradient coating towards the Ag-side. Here, a rough and porous, yet rather homogeneous surface appeared. Enrichment in Ag could only be seen as slightly more electron dense regions (appearing brighter) in the cross-sectional scanning electron microscopy (SEM) images, which also indicated a potential presence of intermetallic AgTi phases or corresponding oxides. A gradual phase separation, coupled with nucleation and growth of metallic Ag particles, was seen when moving towards the center and Ti-rich end of the gradient, shown in Figure 18. At the Ti-side, silver was present as AgNPs with diameter ranging from 20 to 200 nm, scattered across a consid-erably denser coating. Smaller AgNPs, having been lodged in the TiO2 struc-ture during deposition, were seen when examining the cross-section.

Figure 17. SEM images showing surface morphology at the Ag-side of the gradient coating in (a) and (b). The cross-sectional view in (c) and (d) was cut using a fo-cused ion beam (FIB), and show (from the bottom): Si wafer, Ag-Ti oxide coating (approx. 600 nm thick), and two layers of protective Pt film deposited at different currents.

43

Figure 18. SEM images of the surface structure and morphology at the center (a) and Ti-side (b) of the gradient coating. Protruding crystals in (a) and bright particles in (b) were verified as pure Ag. Cross-sectional views of the center in (c), and of the Ti-side in (d).

Composition along the gradient was dominated by silver, largely due to its softer nature that gives it a considerably higher sputtering rate. As shown in Figure 19, the compositional spread ranged from 62 wt% Ag, 21 wt% Ti, and 17 wt% O on the Ag-side, to 35 wt% Ag, 35 wt% Ti and 30 wt% O on the Ti-side. In atomic terms, the corresponding content was 27-11 at% Ag, 20-27 at% Ti, and 38-53 at% O, counting from the Ag-side to the Ti-side.

Considering the Ti:O ratio, which is thermodynamically most stable at 1:2, forming compounds of TiO2, it ranged from 1:2.65 on the Ag-side to 1:2.3 on the Ti-side. This implies that more oxygen was bound to silver at the Ag-side than at the Ti-side, further supporting the claim that the silver content shifted from oxide to metallic form along the gradient.

On the right hand side in Figure 19, contact angle measurements along the gradient are shown, which demonstrated hydrophilic properties on the Ag-side, but an increase in contact angle towards the Ti-side. Illuminating the sample with UV-light to probe for photocatalytic or superhydrophilic properties resulted in a general decrease of the contact angle, indicating a slight degree of photoactivity. The lower contact angles on the Ag-side were in line with its more porous and rough appearance, contributing to a larger specific surface area and higher surface energy.100

44

Figure 19. To the left: composition (wt%) at three different points along the gradi-ent, obtained by energy dispersive X-ray spectroscopy (EDS). To the right: sessile drop contact angle measurements along the gradient. Trendlines included for both measurement series.

Resulting properties The antibacterial properties of the Ag-Ti oxide gradient were tested against S. aureus in a bacterial adhesion test, quantified by CFU counting. Briefly, 10 µl of S. aureus suspension was spread onto 1 cm2 samples cut from three parts along the gradient (Ag-side, center, and Ti-side), and incubated at 37°C for 2 h. Adhered bacteria were then removed by sonication, plated on agar, and incubated overnight before counting viable colonies. The test was per-formed twice with duplicates for each position on the gradient, resulting in n = 4.

The result, showed in Figure 20, demonstrated a clear antibacterial effect compared to the control, with statistically significant reductions in survival rate for all positions tested along the gradient (p < 0.05). The differences, however, were great along the gradient with 99.6% reduction at the Ag-side compared to 17% reduction at the Ti-side. This was correlated with the Ag+ release curves, obtained via inductively coupled plasma-atomic emission spectroscopy (ICP-AES), shown to the right in Figure 20.

The Ag+ ion release study demonstrated a significantly higher dissolution rate of the coating at the Ag-side than at the Ti-side. This difference was much greater than the reported difference in composition (see Figure 19), illustrating that the structural variety along the gradient had a major impact on the resulting properties.

45

Figure 20. On the left: calculated survival rate (%) and number of colonies remain-ing after 2 h contact with different parts of the Ag-Ti oxide gradient coating and plating on agar. On the right: cumulative Ag+ ion release from the Ag-side and Ti-side of the gradient.

The Ag+ ion release study was performed in Dulbecco’s phosphate buffered saline (D-PBS), containing MgCl2 and CaCl2. Since the D-PBS has an ion concentration somewhat similar to that of blood plasma, it can be likened with an SBF and the release study doubled as an assay for in vitro apatite formation on the Ag- and Ti-sides of the sample. In short, if HA deposits on the surface when submerged in D-PBS, it can be considered an indication of bioactivity.101

Both the Ag-side and Ti-side of the gradient sample demonstrated bioac-tive properties, but with varying degree. Figure 21 shows the precipitated HA layer on the Ag-side in (a), and the scattered HA particles found on the Ti-side in (b). The GI-XRD spectra in (c) verified the precipitated species as HA, but also demonstrated that they contained traces of silver.

The greater coverage of HA on the Ag-side than on the Ti-side goes hand in hand with the contact angle measurements shown in Figure 19, which indicated a higher surface energy towards the Ag-side. Increased surface hydration and HA deposition also indicates a greater surface hydroxylation, as was discussed in relation to the H2O2-oxidized and H2O-aged titanium in Paper II.

46

Figure 21. SEM images showing precipitated HA on the Ag-side in (a) and on the Ti-side in (b) after being submerged in D-PBS for 7 days. GI-XRD spectrum in (c) verifying the layers as HA with traces of silver.

The deposited HA layer was poorly adhered and easily removed. In Figure 22, the structural appearance and composition of the underlying Ag- and Ti-side surfaces are shown. Summarizing the above-mentioned findings and looking at these surfaces, one can appreciate the difference in corrosion and consequent release of Ag+ ions, governed by structural features such as crys-tallinity and oxidation state. On the Ti-side, previously adhered AgNPs were no longer present, but the underlying structure and composition was nearly intact. This indicated poor adhesion of the formed AgNPs, and that the re-leased Ag+ ions primarily came from this source. The Ag-side, on the other hand, demonstrated a visibly corroded surface with considerably less silver than before, indicating more structural dissolution of the coating. To draw any conclusions regarding biocompatibility, further in vitro cell studies would have to be done. It is likely that the high dose of Ag+ released from the Ag-side would have caused adverse effects also in mammalian cells, but a more acceptable release is also plausible within the confines of the gradi-ent. What is more important, however, is that the study showed clearly how structure trumps composition in terms of Ag+ release and resulting antibacte-rial effect, which is a key aspect to consider.

47

Figure 22. SEM images of the Ag-side (a) and Ti-side (b) surfaces after performing the Ag+ ion release study in D-PBS. Surface composition quantified by EDS in (c). The precipitated HA layer was removed prior to analysis.

As a tool to explore potentially new compounds and structures, or efficiently screen materials for new applications, the combinatorial approach has sever-al advantages over more traditional methods. As demonstrated in Paper IV, a single gradient sample can be produced to have very distinct properties from one end to another. Producing the same results by traditional means would require a series of coatings and extremely tight control over process parameters. A true combinatorial (or high-throughput) process, however, should preferably be complemented with automated or otherwise highly efficient characterization and property screening. To this end, the current approach is still in need of some development. Nevertheless, the composi-tion-structure-property relationships displayed in Paper IV can very well shape the development of future silver-based antibacterial implant coatings, where structure can be tailored to provide a therapeutic but non-toxic release of Ag+.

48

UV-induced SOS response

Similarly to the study in Paper IV, a combinatorial approach was used to deposit Ag-Ti oxide gradients in Paper V. The aim was to further evaluate antibacterial properties along the gradient, this time against S. epidermidis, and investigate potential photocatalytic properties of Ag-doped TiO2.

The photocatalytic disinfection process on TiO2, as well as the antibacte-rial function of Ag+, has already been laid out. What remains to explore, however, is how these two strategies can be combined to further improve the antibacterial effect in multifunctional coatings.

Photocatalysis on Ag-doped TiO2 Noble metal doping, or deposition of noble metals on TiO2 coatings have been shown to enhance the photocatalytic antibacterial effect in two ways; by providing electron traps that reduce electron-hole recombination, and by interfering with the band-gap to increase visible light absorption.81,99,102,103 The proposed process is illustrated in Figure 23. The trapping of electrons, excited across the band-gap, occurs according to scheme (a) and can in-crease the incidence of charge transfer to adsorbed O2 and H2O. Sensibility to visible light is enabled by the surface plasmon resonance of adhered AgNPs, which can render an injection of e- to the TiO2 conduction band, as illustrated in (b).

Again, a number of technologies are available to Ag-dope or deposit Ag particles on TiO2, and the reactive magnetron sputtering used in Paper IV is a good and versatile alternative. However, the coating was deposited at room temperature to enable characterization and property evaluation of metastable phases, resulting in poor TiO2 crystallinity. Heat treating the amorphous TiO2 layer, e.g. like was done for the H2O2-oxidized samples in Paper II, would likely induce crystallization for improved PCA, but would most cer-tainly also induce phase separation and a loss of some beneficial properties discovered in Paper IV. Hence, the samples produced for Paper V were still sputtered at room temperature and tested without any crystallization heat treatment. Additionally, similar structures of Ag-doped TiO2, deposited via magnetron sputtering at low temperature, have previously shown an ena-blement or enhancement of photocatalytic reactions.104,105

49

Figure 23. Illustration of how Ag-doping can aid or enhance photocatalytic reac-tions on TiO2. In (a), Ag deposits can act as electron traps; reducing electron-hole recombination and thereby enhancing the charge transfer to adsorbed O2 and H2O. In (b), surface plasmon resonance effects in adhered AgNPs, responsive to visible light, can render an injection of electrons to the conduction band of TiO2.

New gradient characteristics

Using the same set-up and similar deposition parameters as in Paper IV, a series of identical Ag-Ti oxide gradient coatings, along with Ag and Ti oxide reference coatings, were produced for Paper V. Three areas along the gradi-ent (A, B, and C as depicted on top in Figure 24) were characterized and tested. The elemental composition and crystalline structure of these areas, compared to Ag and Ti reference samples that were also produced via mag-netron sputtering, are also presented in Figure 24. A thin Ti biding layer was first deposited on all substrates, explaining the presence of Ti in the Ag ref-erence sample. This presence, however, should be considered an artifact on part of the EDS quantification, which is related to the excitation volume created by the incoming electron beam, in this instance set to 7 kV. As seen in the cross-section images at the bottom of Figure 24, the Ag reference coating was thin (70 nm) compared to the Ag-Ti gradient (310–360 nm) and the Ti reference (200 nm). Hence, no Ti was present at the very top of the Ag reference, and the Ti content in the other samples might be slightly over-estimated.

Compared to the gradient produced in Paper IV, where the Ag content varied from 11 to 27 at%, the gradient in Paper V had an Ag content be-tween 9 and 34 at%. The structural variety, however, was not as distinct as before. Area “A”, close to the Ti-side on samples produced in Paper V, bore more resemblance to the center of the gradient produced in Paper IV.

50

Moreover, no columnar TiO2 structure laden with AgNPs could be seen along the gradient in Figure 24, indicating a general shift towards Ag oxide in samples produced for Paper V. The GI-XRD data also indicated that crystalline Ag2O was present on the Ag reference, and that the Ti reference was essentially amorphous.

Figure 24. Summary of sample characteristics from Paper V: composition, crystal-line structure and cross-sectional appearance of Ag-Ti oxide gradient and reference coatings. Illustration on top shows geometry of the Ag-Ti gradient and the position of areas A (Ti-rich), B (center), and C (Ag-rich).

Secondary effects of UV illumination

The bacterial viability assay in Paper V was developed from the methods used in Papers III and IV, and the target bacterium was S. epidermidis. In brief, 10 µl of bacteria suspension was spread onto 1 cm2 samples cut from each area of interest, then incubated for 30 min at 37°C and 100 % relative humidity. One set of samples was then irradiated with UV light (365 nm, 3.0 mW/cm2) for 10 min, while another set was kept in darkness. All composi-tions and conditions were tested in triplicates to allow for statistical evalua-tion. The remaining viable bacteria were then quantified by fluorescence

51

measurements, and compared to two negative controls representing 0 % and 100 % viability.

The results, displayed in Figure 25, demonstrated a significant antibacte-rial effect along the entire gradient, and near complete reduction (97%) for area C. However, the results also revealed another intricate effect: for the Ti oxide reference (denoted Ti), area C, and the Ag oxide reference (denoted Ag), the bacterial viability was greater after UV irradiation. For area C and the Ag oxide reference, these differences were significant at p < 0.05.

Figure 25. Viability of bacteria after contact with the Ti oxide reference sample, areas A, B, and C along the Ag-Ti oxide gradient, and with the Ag oxide reference sample, with and without UV irradiation. The asterisk (*) indicates a significant difference in test for multiple comparisons, while the double asterisk (**) indicates a significant difference in independent sample t-tests, both at p < 0.05. In the assay, 100% bacterial viability corresponded to 2.6×106 CFU/ml of S. epidermidis.

Since there were no appreciable amounts of crystalline anatase in the coat-ings, the photocatalytic properties were poor even with the Ag dopant. Hence, little or no ROS were produced by electron-hole generation, and the reduction in bacterial viability could be attributed entirely to Ag+ release. With a non-crystalline TiO2 contribution, the effect of UV irradiation was thus opposite from what was hypothesized.

To investigate the matter further, S. epidermidis was again seeded on Ti and Ag oxide reference samples, which showed the most diverse properties, and irradiated with UV light or kept in darkness as previously described. Adhering bacteria were then fixated for inspection in the SEM, and the re-sults are shown in Figure 26 and Figure 27.

The visual inspection corroborated the results shown in Figure 25, dis-playing healthy bacteria and some early biofilm formation on the Ti oxide

52

reference, while remnants of lysed cells were scattered across the Ag oxide reference. Adding UV light caused a debatable reduction in biofilm for-mation on the Ti oxide reference, which otherwise hosted a similar amount of bacteria. UV irradiation of bacteria on the Ag oxide reference, however, yielded a greater difference since all bacteria appeared intact. These findings suggested a defensive reaction within the bacteria in response to UV irradia-tion, especially visible in those also subjected to Ag+ ions.

Figure 26. Visual inspection of S. epidermidis on the Ti oxide reference; kept in darkness (Ti) in (a) and (b), and irradiated with UV light for 10 min (Ti-UV) in (c) and (d). The SEM images were color enhanced after acquisition, highlighting bio-film in yellow and bacteria in pink.

Figure 27. Visual inspection of S. epidermidis on the Ag oxide reference; kept in darkness (Ag) in (a), (b), and (c), and irradiated with UV light for 10 min (Ag-UV) in (d) and (e). The SEM images were color enhanced after acquisition to highlight traces of biofilm (yellow), intact bacteria (pink), and remnants of lysed bacteria (blue).

53

High energy UV-C is known to cause irreparable DNA damage and is used as a disinfection method.106,107 Sublethal doses of UV, however, can trigger an SOS response in E. coli and other bacteria,108,109 which pauses the DNA replication process while the damage is being assessed and potentially re-paired.

In Paper V, it was thus stipulated that Ag+ induced damage to bacteria could be limited by a UV-induced SOS response. Since both UV irradiation and Ag+ ions affect the bacterial DNA, a pause in replication and hindering of premature cell division could, if only momentarily, help survival.

By surveying the viability of bacteria 2 and 4 h after treatment, it was noted that it increased with time for Ti and Ti-UV samples, suggesting that bacterial communities thrived after contact with these surfaces. However, for the Ag and Ag-UV samples, as well as for samples along the gradient, the viability, declined steadily with time, which supports the claim of reduced replication abilities after Ag+ exposure. Moreover, since all bacterial com-munities exposed to Ag+ demonstrated a gradual decline in viability with time, the relatively higher levels seen after UV irradiation should be inter-preted as a slightly slower reduction rate, rather than a recovery in viability. Hence, the suggested UV-induced SOS response could not stop the damage caused by Ag+, only slow the process down.

The qualitative reduction in biofilm after UV irradiation observed with the SEM inspection, most notable on the Ti oxide reference, could partly explain the increase in viability, as more bacteria would still be in a metabol-ically active planktonic state. For the Ag oxide reference, a UV-induced pause in replication, or rendering of an active but nonculturable state by Ag+ exposure, as suggested by Jung et al.,110 could explain the numerous bacteria seen in Figure 27d-e, compared to the poor viability seen for Ag-UV sam-ples in Figure 25.

In summary, the Ag-Ti oxide gradient thin film produced in Paper V was clearly antibacterial due to Ag+ ion release, but it lacked photocatalytic properties since the deposited TiO2 was amorphous. These characteristics led to some interesting findings when testing antibacterial properties under UV illumination. Contrary to what was hypothesized, UV light did not gen-erate electron-hole pairs to yield ROS, but rather provoked an SOS response in the bacteria that limited its reduction rate. For the development of multi-functional coatings with antibacterial properties that wholly or partially de-pend on UV activation, this is a secondary effect to keep in mind.

54

Concluding remarks

Developing and evaluating antibacterial surfaces is not a trivial matter, cer-tainly not when potential adverse effects on friendly cells have to be mitigat-ed. The extent of the work presented in this thesis was limited to the manu-facturing, characterization, photocatalytic property and in vitro antibacterial property evaluation of two principal surfaces; H2O2-oxidized and H2O-aged Ti, and co-deposited Ag-Ti oxide. As for the antibacterial properties of these surfaces, they were both demonstrated effective against planktonic cells of clinically relevant species, encouraging further development and evaluation to ensure safety aspects and maintained efficiency in an in vivo situation.

Concerning the H2O2-oxidized and H2O-aged Ti and Ti-6Al-4V in Pa-pers I, II, and III, it was concluded that an extended treatment time (24 h + 72 h) was necessary to obtain satisfactory results. It was further revealed that the ROS generation on these surfaces was not necessarily reliant on photoca-talysis, but rather on an activation of Ti-peroxy radical species. Both the degradation of rhodamine B, which served as an indicator of PCA, and the actual antibacterial activity was enhanced by adding a small amount of H2O2 to the reaction, demonstrating synergy effects in the TiO2/H2O2/UV system.

For the Ag-Ti oxide gradients in Papers IV and V, a combinatorial approach was applied by means of concurrent magnetron sputtering from two opposing targets in a reactive environment. This enabled deposition of continuous, binary gradients containing a range of compositions and structures that would have been impossible or very cumbersome to produce by traditional means. The ad-vantages of the combinatorial approach were highlighted in Paper IV, where the entire study was based on a single Ag-Ti oxide gradient sample. In Paper IV, the amount of Ag+ released from parts of the coating likely exceeded ac-ceptable levels in a normal case scenario, but the compositional and structural properties affecting the release were clarified. This led to the conclusion that future implant coatings could be better tailored for clinical needs by altering structure of incorporated silver species rather than amount.

While trying to combine and improve on the two antibacterial strategies explored in this thesis, i.e. photocatalysis and Ag+ release, it was found in Paper V that on amorphous TiO2, UV illumination caused an SOS response in bacteria that limited the reduction rate caused by Ag+ exposure. Hence, multifunctional implant coatings that rely on UV activation should be devel-oped with such secondary effects in mind.

55

Future challenges

Translation of scientific findings from the lab bench to a successful clinical application, which would be the end goal of the work presented in this the-sis, is a long endeavor. The natural next step for the surfaces presented here would be mammalian cell studies and thorough evaluation of potential cyto-toxicity, followed by in vivo animal trials given the biocompatibility is not compromised. In short, all surfaces and systems presented would require further development before being considered clinically.

In its present state, the H2O2-oxidized and H2O-aged Ti surface could likely find success on abutments, a component of the dental implant con-struct that is susceptible to infection, but also easily accessible by UV light. It is also a component with fewer requirements for direct tissue integration, making potential adverse effects of the surface less impactful. For applica-tion as an implant coating, in vivo trials to confirm osseointegration and/or form basis for a risk-benefit analysis would be required. For controlled UV-activation of the surface, a hand-held device delivering the right dose of light would also have to be developed.

Concerning the Ag+ releasing surfaces, much research and development remains. The presented studies form more of a basis for understanding com-position-structure-property relationships than demonstrating a prototype surface for antibacterial applications. Continued development of the combi-natorial deposition and analysis approach, including more streamlined prop-erty evaluation is needed. Once the right balance between biocompatibility and antibacterial properties have been struck by further tweaking the compo-sition and structure, other fabrication methods than PVD might also benefit.

For the combined Ag+ releasing and photoactivated surfaces, further de-velopment is obviously needed to optimize both Ag+ release and photocata-lytic properties, all while keeping potential SOS responses in mind.

56

Analytical techniques

Rhodamine B degradation In Papers I and II, degradation of the organic dye rhodamine B was used to evaluate PCA of H2O2-oxidized and H2O-aged titanium samples. ROS gen-erated from the photocatalytic process will degrade the dye molecules, and the resulting loss of color is easily monitored by measuring UV/Vis absorp-tion.

For each test, four identical samples were placed in a flat-bottomed quartz cuvette with 1 cm optical path length, containing 2.5 ml of 5 µM rhodamine B solution, as shown in Figure 28. A UV transparent glass lid covered the cuvette to avoid evaporation, and a small well drilled in the bottom of the cuvette allowed magnetic stirring for homogenization and oxygenation dur-ing the tests. The cuvette was placed in a UV/Vis spectrophotometer, fitted with a UV light source (λ = 365±10 nm) irradiating the samples and solution with pulsed light (100 Hz) at 6.7 mW/cm2 intensity. The UV light was mo-mentarily blacked out for automated stepwise absorbance measurements of the solution, taken at 554 nm every 5 min for up to 8 h.

Converting the absorbance values to rhodamine B concentration then al-lowed quantification of PCA via the function ln(C/C0) = –kt, where C/C0 expresses the normalized concentration, k is the reaction rate constant re-flecting the PCA, and t is the reaction time.

Figure 28. Illustration of the method used to quantify PCA. Degradation of rhoda-mine B was monitored over time in presence of TiO2 samples and UV light.

57

Scanning electron microscopy (SEM) SEM is a widely used technique for high resolution imaging of surfaces. It also features a large depth of field, giving images a three-dimensional ap-pearance. Rather than using visible light photons, like in an optical micro-scope, the SEM uses a focused beam of electrons to scan the surface, which have significantly shorter wavelengths than visible light.

The incident beam of electron interacts with the surface in various ways, giving rise to a number of detectable signals. Secondary electrons are pro-duced during inelastic scattering of the incident beam, i.e. when incident electrons knock out K-shell electrons from specimen atoms. The secondary electrons generated close to the surface have a greater chance of escape to reach the detector, resulting in topographical information about the surface. Backscattered electrons are electrons that originate from the electron beam, but have been elastically scattered by coming close to atomic nuclei in the specimen. Heavier elements generate more backscattered electrons, making the signal an indicator of composition.

SEM was used to image surfaces in Papers I, II, IV, and V, primarily us-ing the InLens detector (secondary electrons) in a Zeiss LEO 1550 operated a 3-7 kV. For Paper IV, a Zeiss crossbeam Neon 40 operated at 5 kV was also used. For non-conducting surfaces, e.g. the HA deposits in Paper IV and the bacteria imaged in Paper V, a thin layer of Au/Pd was sputter de-posited on top of the samples to avoid charge build-up.

Energy dispersive X-ray spectroscopy (EDS) In the SEM, inelastic scattering or excitation of electrons creates atomic inner shell vacancies. This state of ionization forces outer shell electrons to drop down in energy to fill the vacancy, resulting in release of excess energy in the form of an X-ray photon or an Auger electron. The energy differences between electron shells in different elements are well defined, making each X-ray photon constitute a characteristic X-ray of its chemical element. Hence, by detecting the energy and amount of emitted X-ray photons, the chemical composition of a specimen can be determined.

The surface sensitivity and accuracy of EDS is related to the acceleration voltage used in the SEM. A higher acceleration voltage creates a larger (deeper) excitation volume, generating more X-ray photons from bulk mate-rial that may suppress the surface signal, but a lower acceleration voltage may not be sufficient to excite electrons with distinguishable energies in all elements present.

EDS was used to determine the elemental compositions in gradient Ag-Ti oxide thin films in Papers IV and V, using detectors fitted to above-mentioned SEMs and the INCA AZtec software.

58

Focused ion beam (FIB) The FIB, which can be incorporated into the SEM, uses an ion beam source (usually Ga+) to sputter or mill the surface of a specimen. It is commonly used to prepare exquisitely thin samples for transmission electron microsco-py (TEM), which have to be electron transparent. It can also be used for micromachining or milling of surfaces, e.g. to analyze a cross-section of a coating.

In relation to this thesis, a FIB (fitted in a Zeiss crossbeam Neon 40) was used to mill a cross-section for analysis at the Ag-side of the gradient coat-ing produced in Paper IV.

Grazing incidence X-ray diffraction (GI-XRD) XRD uses an incident beam of monochromatic X-rays, usually Cu-Kα (λ = 0.154 nm), to determine the phase composition of a crystalline material. The analysis is done by enumeration of the amount and position of diffracted X-rays, and result in a spectrum displaying intensity vs. angular position.

Crystalline materials have a long-range periodical arrangement of atoms, built on specific crystal lattices that constitute a number of crystallographic planes. In XRD, constructive interference between X-rays diffracted from different parallel planes, according to Bragg’s law (nλ = 2dhkl sinθ), will generate a signal, as described in Figure 29. Scanning a range of angles will thus create an XRD pattern, which can be matched with reference patterns to determine the crystalline composition of the sample.

GI-XRD is based on the same principle, but the incident beam is fixed at a low angle, typically 1°. This allows for analysis of crystalline structures in thin films, since the X-ray penetration depth at low angles is shallow (order of nanometers). GI-XRD was used on thin films in Papers I, II, IV, and V.

Figure 29. The wavelength of incident X-rays (λ), along with the distance between crystallographic planes (dhkl) and the angle θ constitute Bragg’s law of diffraction, which is the basis for XRD. Scanning θ, a spectrum peak occurs when the diffracted beams are “in phase”.

59

Surface roughness (Ra, Rq) Surface roughness is, as previously discussed, an important parameter for dental and orthopedic implants. For this thesis, the average roughness (Ra) and root mean square roughness (Rq) of sample surfaces in Papers I and IV were determined by white light interferometry.

This optical, non-contact method uses light that is separated into two beams; one directed at the sample and one directed at a reference mirror. When the reflected light is recombined, interference patterns according to sample topography appear. By scanning a surface, the vertical position of each point can thus be determined at the nanometer level by measuring the level of interference with the reference beam. A WYKO NT1100 optical profiler was used for the measurements.

Contact angle (CA) measurements Water CA measurements are often used to determine the wettability of a surface. The CA is defined as the angle between the solid surface and the tangent line of the liquid phase at the solid-liquid-gas interphases, as shown in Figure 30. Surfaces with a CA above 90° are considered hydrophobic, while a CA below 90° signifies a hydrophilic surface. If the CA is close to 0° or above 150°, the surface is considered superhydrophilic or superhydro-phobic, respectively.

In Paper IV, the sessile drop method was used to determine the CA along the Ag-Ti oxide gradient.

Figure 30. Contact angle (θ) measured by the sessile drop method.

Bacterial viability assays There are a number of techniques available to evaluate the antibacterial or bacteriostatic properties of surfaces and materials. However, depending on the antibacterial function, or the materials mode of action against bacteria, all techniques are not suitable, and careful planning and sterile execution is required to obtain reliable results.

60

A quite common and simple method is to evaluate bacterial adhesion, in which the sample of interest is generally submerged in a single strain sus-pension of bacteria. After some time of incubation, the sample is removed and washed from non-adhering bacteria. The remaining, adhering bacteria are then suspended in solution by sonication or vigorous vortex shaking. Quantification of the adhered bacteria can then be done by, e.g. optical den-sity measurements (rough estimate), serial plating on agar followed by CFU counting (good estimate), mixing with a fluorescence marker sensitive to bacterial viability (good estimate, but requires calibration), or visual inspec-tion by light microscopy or SEM (variable estimate).111

For the surfaces evaluated in this thesis, which were photoactivated and/or deposited on Si-wafers, submerging the entire sample in bacterial solution would constitute a significant disadvantage, primarily since only bacteria adhered to the active (UV-illuminated or Ag+ releasing) surface would be affected. Hence, a DCT method was devised where only the active surface is evaluated. In effect by allowing a small volume (5-10 µl) of bacte-ria in suspension adhere to the surface of interest, then testing while losses due to other factors such as desiccation can be controlled. The approach, as applied with UV light and TiO2 in Paper III, is illustrated in Figure 31.

In Papers III and V, bacterial viability was quantified by mixing sus-pended bacteria with culture medium and the metabolic activity indicator resazurin. Resazurin is a blue non-fluorescent dye, which in presence of metabolic intermediates is reduced to pink, fluorescent resorufin. As such, the fluorescence becomes a sensitive indicator of viable bacteria. To corre-late the fluorescence readings with known amounts of bacteria, however, a standard curve has to be prepared in parallel with every experiment. With the aid of standard curves and negative controls, the viability of bacteria after exposure to various surfaces and conditions could be expressed in per-cent compared to a control surface.

In Paper IV, surviving bacteria detached from the sample surfaces were quantified by CFU counting. From the suspensions, ten-fold serial dilutions were plated on brain heart infusion (BHI) agar, and colonies were counted the following day.

Figure 31. Illustration of the DCT method used to evaluate photocatalytic samples.

61

Inductively coupled plasma–atomic emission spectroscopy (ICP-AES) ICP-AES is an analytical technique used to detect trace amounts of metals in solution. In the equipment, a radio frequency coil ionizes Ar gas to generate a high temperature plasma. When the sample to be analyzed is introduced, it is evaporated and broken down into atoms, which collide with electrons and Ar+ ions in the plasma to repeatedly lose and gain electrons, giving off char-acteristic radiation in the process. The wavelength and intensity of each emission is then used to identify and quantify the elements present in the sample.

In Paper IV, ICP-AES was used to quantify the amount of Ag+ that was released in D-PBS from different parts of the Ag-Ti oxide gradient over time.

62

Svensk sammanfattning

Dentala och ortopediska implantat är allt som oftast tillverkade i titan, eller i någon titanlegering där de mekaniska egenskaperna förbättrats ytterligare. Anledningarna till detta är flera: Titan är starkt, men har också låg densitet, vilket gör att man kan tillverka starka lättviktskonstruktioner. Det är också mindre styvt än många andra tänkbara medicinska metaller, vilket gör att matchningen mot benvävnad blir aningen bättre. Framförallt är det dock biokompatibelt och har visat sig binda väldigt bra till ben i en process som kallas osseointegration, en term som myntades av svensken Per-Ingvar Brå-nemark på 1950-talet.

Möjligheten att ersätta förlorade tänder eller utnötta leder med hjälp av metalliska implantat har i modern tid bidragit till att miljontals människor kan leva längre, mer aktivt och förhoppningsvis även lite lyckligare. I många fall håller implantaten livet ut, men beroende på en lång rad faktorer kan de i vissa fall behöva bytas ut. En starkt bidragande orsak till att implantat falle-rar och man tvingas till en ny operation är infektioner. De kan härstamma från första operationstillfället, då bakterier alltid finns närvarande oavsett hur väldisciplinerat läkarteamet är eller hur väl steriliserade instrumenten och omgivningarna är. Bakterier kan även komma att kolonisera ytan av ett implantat långt efter operationen, då de kan färdas med blodet från en infekt-ion någon annanstans i kroppen. Mest utsatta är dock implantat som passerar någon av kroppens naturliga barriärer, såsom tandimplantat eller proteser förankrade direkt i benet. I dessa fall har bakterier en naturlig väg in, och det är i princip omöjligt att förhindra någon form av infektion.

Denna avhandling syftar till att utveckla tekniker för att förhindra eller begränsa spridningen av infektioner på titanimplantat, och två huvudspår har följts upp: fotokatalys på titandioxidytor (TiO2), samt inkorporering och frisättning av silverjoner (Ag+).

Fotokatalys på kristallin TiO2 sker under belysning av UV ljus, vilket skapar fria radikaler på materialets yta som kan oxidera organiska förening-ar. Om dessa organiska föreningar utgör byggstenarna i en bakterie kan denna komma att utplånas.

I avhandlingen utvecklades och användes en enkel och kostnadseffektiv metod för att oxidera titanytor till TiO2. Den baserades på oxidering i väte-peroxid (H2O2) följt av en åldringsprocess i vatten, och visade sig generera ytor som var aktiva mot organiska föreningar under UV ljus. Däremot visade

63

sig dessa ytor vara amorfa (icke-kristallina) och mindre stabila vid uppre-pade försök, vilket föranledde misstankar om att aktiviteten berodde på nå-got annat än fotokatalys. Vidare undersökningar visade att aktiviteten till stor del berodde på UV-aktivering och förbrukning av Ti-peroxy föreningar som skapats vid H2O2-oxideringen. För att gå vidare och på ett smart sätt utnyttja denna kunskap så utvecklades ett system där TiO2 ytorna, UV ljus och en svag H2O2 lösning kombinerades och testades mot bakterier. Det visade sig vara ett effektivt system, där cirka 99 % av två kliniskt relevanta bakteriearter kunde elimineras på 20 min.

Gällande inkorporering och frisättning av silverjoner så utnyttjades en kombinatorisk materialutvecklingsmetod, där silver och titan deponerades tillsammans på ett substrat i syreatmosfär, men kommandes olika från källor placerade mittemot varandra. Detta möjliggjorde deponering av en kompo-sitions- och strukturmässig gradient, innehållandes en silverdominerad och en titandominerad ände. De antibakteriella egenskaperna hos denna gradient följde tydligt dess benägenhet att frisätta silverjoner. Den silverdominerade änden visade starkast antibakteriella egenskaper, men mer på grund av att ytan var mer porös och troligen innehöll med silveroxid än vad som fanns på den titandominerade ytan, där silvret var representerat i form av metalliska nanopartiklar.

De två spåren för att utveckla antibakteriella ytor kombinerades även i den sista studien, där fotokatalys på silverdopad TiO2 utvärderades. För att behålla de intressanta ytegenskaper som hade påträffats tidigare, med tydlig skillnad i frisättning av silverjoner beroende på skillnader i ytstruktur, gjor-des ingen värmebehandling för att kristallisera ytan. Detta ledde, återigen, till en amorf TiO2 utan tydlig fotokatalytisk aktivitet. I test med bakterier visades en distinkt effekt på grund av frisättning av Ag+, men mot förmodan så begränsades effekten då ytor och bakterier belystes med UV ljus. Detta ledde till slutsatsen att UV ljuset, i avsaknad av kristallin TiO2 som kunde generera fria radikaler, istället framkallade ett försvar hos bakterierna som till viss mån bromsade den skadliga effekten av Ag+.

64

Acknowledgements

Först och främst vill jag tacka fisken som krälade upp på land – utan hen vore inget av det här möjligt.

Skämt åsido, det finns många som bär ansvar för att denna avhandling har blivit till. Först Håkan och Cecilia, som plockade in mig som ex-jobbare när jag ville mjölka det sista av studentlivet i Sveriges mesta studentstad. Att rekrytera nämnd ex-jobbare till blivande doktorand genom att per omgående skicka honom till Leipzig för att provköra Porsches senaste modellutbud gjorde karriärvalet enkelt. Ni har båda ett driv och intresse för forskning, utveckling och materialvetenskap som är fantastiskt. Att sedan handleda likt Smiths osynliga hand men alltjämt ha stenkoll för väl avvägda punktinsatser är avundsvärt. Tillslut blev dock bollarna i luften för många och Ken tog över. Sedan starten har vi bollat idéer och du har alltid funnits till hands för vägledning i detaljerna. Jag är imponerad över din nyfikenhet och uppfin-ningsrikedom. Utan dina hemmabyggda manicker hade jag bara varit halv-vägs. Det är även få förunnat att ha ett flygaress till handledare. Stort tack till er alla tre!

Sedan vill jag uttrycka ett stort tack till Stiftelsen för Strategisk Forskning (SSF) för finansiering och satsningen på svensk produktutveckling genom ProViking-programmet. Det har varit ett rent nöje att delta i forskarskolan och åka på doktorandträffar. Jag vill även tacka Jukka på SP, Anders på Arcam och Reza på Dentware för samarbetet inom EBM projektet. Ni har alla bidragit på olika sätt med kunskap och idéer som drivit mitt arbete framåt. Joakim, som doktorandkollega i projektet har du varit min vapendra-gare på resor och möten, tack för trevligt sällskap!

Thank you Yanling for helping me get familiar with bacteria assays, and thank you to my protégés Eleni, and then Matthias for you hard work. I hope you learned as much as I did. Unfortunately, the bacteria will always win!

Tack Andreas för allt jobb och all hjälp med Sleipnir. Thanks also to Fang for helping with the depositions.

Muchas gracias a Daniel y todo el grupo BIBITE en la UPC a Barcelona. Fue una experiencia molt bé y aprendí mucho, espero qua la colaboración puede continuar.

Thank you all past and present members of the MiM research group. Jo-nas, Kathryn, Carl, Ale, Mirjam, and of course Johanna, you all helped out in various projects and made the lab a fun place to be. Thank you Wei for

65

keeping the GC project alive. Thank you Maryam for bringing me out of my own bubble in the office from time to time. Thank you Ingrid for being a good friend and for entertaining Stina. Thank you Oscar for all the trials and tribulations with the nanoparticles. Thank you Maria for also working with metals and coatings. Thanks Torbjörn for your business mindset. Thank you everybody else for making the group a fun and dynamic mix of people!

Fortunately, not everyone at the department works with biomaterials. The endless discussions during fika and mombasa are usually fruitless, but an absolute necessity all the same. While there is progress in biotribology and microchip diagnostics, I’m still waiting for proper implant propulsion in space! Past and present co-workers in the Tribology, ElMiN, and MST groups have made the last five years truly enjoyable. Fredrik, you’ve been a partner in crime on many occasions. Thanks for challenging me in the slopes, in the streets and on the tennis course. Still can’t believe your favor-ite brew is Newcastle though. Thanks to Peter F, Benny, Johanna, Janna, Matthias, Martina, Bagge, Timo, Petra, Harald, Jill, Anders, Ville, Peter S, Pontus and a bunch of others for help on various occasions and for all the good times. Thanks also to all seniors for your wisdom and encouragement.

Thanks to everyone in the Nano group for bearing with me when stealing your space, consumables or attention, and for keeping me in shape through the Monday innebandy sessions.

Thanks to the supporting and administrative staff. Ingrid R, Maria S och Sara, ni är alltid hjälpsamma och har koll på typ allt. Tack Jocke på verkstan för svarvande och fräsande av titanstycken. Thanks to Jan-Åke and the MSL staff for keeping everything in the cleanroom and elsewhere up and running.

Jag vill även tacka Anders, Marja och Axel för ert stöd, er omtanke och allt trevligt umgänge. Tack även alla övriga Engstrands.

Till mina föräldrar, Jan och Kajs-Marie, som stöttat i vått och torrt vill jag ägna ett särskilt tack. Utan er vore verkligen inget av det här möjligt. Till min syster Klara och till min bror Jon, ni har utan tvekan format mig och fått mig att sikta högre än jag hade vågat på egen hand. Tack för allt!

Johanna, min kära, den här är till dig och Stina. Jag ser fram emot varje dag jag får dela med er.

Lastly, I know it’s always great to see your own name on these pages, so if it hasn’t been mentioned already, thank you ______________ for your unwa-vering support, encouragement and/or interest. It has been most appreciated, cheers!

66

References

1. Albrektsson T, Brånemark PI, Hansson HA, Lindström J. Osseointegrated titanium implants. Requirements for ensuring a long-lasting, direct bone-to-implant anchorage in man. Acta Orthop Scand 1981; 52:155–70.

2. Busscher HJ, van der Mei HC, Subbiahdoss G, Jutte PC, van den Dungen JJAM, Zaat SAJ, Schultz MJ, Grainger DW. Biomaterial-Associated Infec-tion: Locating the Finish Line in the Race for the Surface. Sci Transl Med 2012; 4:153rv10–0.

3. Campoccia D, Montanaro L, Arciola CR. A review of the biomaterials tech-nologies for infection-resistant surfaces. Biomater 2013; 34:8533–54.

4. Williams DF. On the mechanisms of biocompatibility. Biomater 2008; 29:2941–53.

5. Geetha M, Singh AK, Asokamani R, Gogia AK. Ti based biomaterials, the ultimate choice for orthopaedic implants – A review. Prog Mater Sci 2009; 54:397–425.

6. Long M, Rack H. Titanium alloys in total joint replacement - a materials science perspective. Biomater 1998; 19:1621–39.

7. Liu X, Chu P, Ding C. Surface modification of titanium, titanium alloys, and related materials for biomedical applications. Mater Sci Eng R 2004; 47:49–121.

8. Anderson JM, Rodriguez A, Chang DT. Foreign body reaction to biomateri-als. Semin Immunol 2008; 20:86–100.

9. Brånemark R, Emanuelsson L, Palmquist A, Thomsen P. Bone response to laser-induced micro- and nano-size titaniumsurface features. Nanomedicine: NBM 2011; 7:220–7.

10. Gittens RA, Scheideler L, Rupp F, Hyzy SL, Geis-Gerstorfer J, Schwartz Z, Boyan BD. A review on the wettability of dental implant surfaces II: Biolog-ical and clinical aspects. Acta Biomater 2014; 10:2907–18.

11. Yang B. Preparation of bioactive titanium metal via anodic oxidation treat-ment. Biomater 2004; 25:1003–10.

12. Peltola T, Pätsi M, Rahiala H, Kangasniemi I, Yli-Urpo A. Calcium phos-phate induction by sol-gel-derived titania coatings on titanium substrates in vitro. J Biomed Mater Res 1998; 41:504–10.

13. Sawase T, Wennerberg A, Baba K, Tsuboi Y, Sennerby L, Johansson CB, Albrektsson T. Application of oxygen ion implantation to titanium surfaces: effects on surface characteristics, corrosion resistance, and bone response. Clin Implant Dent Rel Res 2001; 3:221–9.

14. Casaletto M, Ingo G, Kaciulis S, Mattogno G, Pandolfi L, Scavia G. Surface studies of in vitro biocompatibility of titanium oxide coatings. Appl Surf Sci 2001; 172:167–77.

15. Unosson E, Rodriguez D, Welch K, Engqvist H. Reactive combinatorial synthesis and characterization of a gradient Ag-Ti oxide thin film with anti-bacterial properties. Acta Biomater 2015; 11:503–10.

67

16. Lee B-H, Kim JK, Kim YD, Choi K, Lee KH. In vivo behavior and mechani-cal stability of surface-modified titanium implants by plasma spray coating and chemical treatments. J Biomed Mater Res 2004; 69A:279–85.

17. Kim H, Miyaji F, Kokubo T, Nakamura T. Preparation of bioactive Ti and its alloys via simple chemical surface treatment. J Biomed Mater Res 1996; 32:409–17.

18. Wang X, Hayakawa S, Tsuru K, Osaka A. Bioactive titania gel layers formed by chemical treatment of Ti substrate with a H2O2/HCl solution. Biomater 2002; 23:1353–7.

19. Gristina AG. Biomaterial-centered infection: microbial adhesion versus tis-sue integration. Science 1987; 237:1588–95.

20. Monroe D. Looking for Chinks in the Armor of Bacterial Biofilms. Plos Biol 2007; 5:e307.

21. Loftus RW, Koff MD, Burchman CC, Schwartzman JD, Thorum V, Read ME, Wood TA, Beach ML. Transmission of pathogenic bacterial organisms in the anesthesia work area. Anesthesiology 2008; 109:399–407.

22. Svenska Höftprotesregistret Årsrapport 2013. http://shpr.se/Libraries/Docu-ments/Arsrapport_2013_3_WEB.sflb.ashx; 2014.

23. Bozic KJ, Ries MD. The impact of infection after total hip arthroplasty on hospital and surgeon resource utilization. J Bone Joint Surg Am 2005; 87:1746–51.

24. Campoccia D, Montanaro L, Arciola CR. The significance of infection relat-ed to orthopedic devices and issues of antibiotic resistance. Biomater 2006; 27:2331–9.

25. Machtei EE, Frankenthal S, Levi G, Elimelech R, Shoshani E, Rosenfeld O, Tagger-Green N, Shlomi B. Treatment of peri-implantitis using multiple ap-plications of chlorhexidine chips: a double-blind, randomized multi-centre clinical trial. J Clin Periodontol 2012; 39:1198–205.

26. Campoccia D, Montanaro L, Arciola CR. A review of the clinical implica-tions of anti-infective biomaterials and infection-resistant surfaces. Biomater 2013; 34:8018–29.

27. Mombelli A, Müller N, Cionca N. The epidemiology of peri-implantitis. Clin Oral Implants Res 2012; 23 Suppl 6:67–76.

28. Pye AD, Lockhart DEA, Dawson MP, Murray CA, Smith AJ. A review of dental implants and infection. J Hosp Infect 2009; 72:104–10.

29. Kumar PS, Mason MR, Brooker MR, O'Brien K. Pyrosequencing reveals unique microbial signatures associated with healthy and failing dental im-plants. J Clin Periodontol 2012; 39:425–33.

30. Neoh KG, Kang ET. Combating Bacterial Colonization on Metals via Poly-mer Coatings: Relevance to Marine and Medical Applications. ACS Appl Mater Interfaces 2011; 3:2808–19.

31. Puckett SD, Taylor E, Raimondo T, Webster TJ. The relationship between the nanostructure of titanium surfaces and bacterial attachment. Biomater 2010; 31:706–13.

32. Necula BS, Fratila-Apachitei LE, Zaat SAJ, Apachitei I, Duszczyk J. In vitro antibacterial activity of porous TiO2–Ag composite layers against methicil-lin-resistant Staphylococcus aureus. Acta Biomater 2009; 5:3573–80.

33. Maness P, Smolinski S, Blake D, Huang Z, Wolfrum E, Jacoby W. Bacteri-cidal activity of photocatalytic TiO2 reaction: Toward an understanding of its killing mechanism. Appl Environ Microbiol 1999; 65:4094–8.

68

34. Kikuchi Y, Sunada K, Iyoda T, Hashimoto K, Fujishima A. Photocatalytic bactericidal effect of TiO2 thin films: dynamic view of the active oxygen species responsible for the effect. J Photoch Photobio A 1997; 106:51–6.

35. Yang X-F, Chen Y, Yang F, He F-M, Zhao S-F. Enhanced initial adhesion of osteoblast-like cells on an anatase-structured titania surface formed by H2O2/HCl solution and heat treatment. Dent Mater 2009; 25:473–80.

36. Del Curto B, Brunella MF, Giordano C, Pedeferri MP, Valtulina V, Visai L, Cigada A. Decreased bacterial adhesion to surface-treated titanium. Int J Ar-tif Organs 2005; 28:718–30.

37. Karthega M, Nagarajan S, Rajendran N. In vitro studies of hydrogen perox-ide treated titanium for biomedical applications. Electrochim Acta 2010; 55:2201–9.

38. Ferraris S, Spriano S, Pan G, Venturello A, Bianchi CL, Chiesa R, Faga MG, Maina G, Vernè E. Surface modification of Ti–6Al–4V alloy for biomineral-ization and specific biological response: Part I, inorganic modification. J Ma-ter Sci: Mater Med 2011; 22:533–45.

39. Wu J, Hayakawa S, Tsuru K, Osaka A. Low-temperature preparation of ana-tase and rutile layers on titanium substrates and their ability to induce in vitro apatite deposition. J Am Ceram Soc 2004; 87:1635–42.

40. Wu J-M, Wang M, Li Y-W, Zhao F-D, Ding X-J, Osaka A. Crystallization of amorphous titania gel by hot water aging and induction of in vitro apatite formation by crystallized titania. Surf Coat Tech 2006; 201:755–61.

41. Sun T, Wang M. Low-temperature biomimetic formation of apatite/TiO2 composite coatings on Ti and NiTi shape memory alloy and their characteri-zation. Appl Surf Sci 2008; 255:396–400.

42. Sun T, Wang M. A comparative study on titania layers formed on Ti, Ti-6Al-4V and NiTi shape memory alloy through a low temperature oxidation pro-cess. Surf Coat Tech 2010; 205:92–101.

43. Variola F, Yi J-H, Richert L, Wuest JD, Rosei F, Nanci A. Tailoring the surface properties of Ti6Al4V by controlled chemical oxidation. Biomater 2008; 29:1285–98.

44. Tengvall P, Elwing H, Lundström I. Titanium Gel Made From Metallic Tita-nium and Hydrogen-Peroxide. J Colloid Interf Sci 1989; 130:405–13.

45. Tengvall P, Wälivaara B, Westerling J, Lundström I. Stable titanium super-oxide radicals in aqueous Ti-peroxy gels and Ti-peroxide solutions. J Colloid Interf Sci 1991; 143:589–92.

46. Wu J-M. Photodegradation of Rhodamine B in Water Assisted by Titania Nanorod Thin Films Subjected to Various Thermal Treatments. Environ Sci Technol 2007; 41:1723–8.

47. Li X, Chen C, Zhao J. Mechanism of Photodecomposition of H2O2 on TiO2 Surfaces under Visible Light Irradiation. Langmuir 2001; 17:4118–22.

48. Hirakawa T, Nosaka Y. Properties of O2*- and OH* Formed in TiO2 Aqueous Suspensions by Photocatalytic Reaction and the Influence of H2O2 and Some Ions. Langmuir 2002; 18:3247–54.

49. Lopez A, Bozzi A, Mascolo G, Kiwi J. Kinetic investigation on UV and UV/H2O2 degradations of pharmaceutical intermediates in aqueous solution. J Photoch Photobio A 2003; 156:121–6.

50. Stefan MI, Hoy AR, Bolton JR. Kinetics and Mechanism of the Degradation and Mineralization of Acetone in Dilute Aqueous Solution Sensitized by the UV Photolysis of Hydrogen Peroxide. Environ Sci Technol 1996; 30:2382–90.

69

51. Cui X, Kim H-M, Kawashita M, Wang L, Xiong T, Kokubo T, Nakamura T. Effect of hot water and heat treatment on the apatite-forming ability of titania films formed on titanium metal via anodic oxidation in acetic acid solutions. J Mater Sci: Mater Med 2007; 19:1767–73.

52. Takadama H, Kim HM, Kokubo T, Nakamura T. An X-ray photoelectron spectroscopy study of the process of apatite formation on bioactive titanium metal. J Biomed Mater Res 2001; 55:185–93.

53. Takadama H, Kim HM, Kokubo T, Nakamura T. TEM-EDX study of mech-anism of bonelike apatite formation on bioactive titanium metal in simulated body fluid. J Biomed Mater Res 2001; 57:441–8.

54. Boonstra AH, Mutsaers CAHA. Relation between the photoadsorption of oxygen and the number of hydroxyl groups on a titanium dioxide surface. J Phys Chem 1975; 79:1694–8.

55. Yu JC, Lin J, Lo D, Lam SK. Influence of Thermal Treatment on the Adsorp-tion of Oxygen and Photocatalytic Activity of TiO2. Langmuir 2000; 16:7304–8.

56. Takemoto S, Yamamoto T, Tsuru K, Hayakawa S, Osaka A, Takashima S. Platelet adhesion on titanium oxide gels: effect of surface oxidation. Bio-mater 2004; 25:3485–92.

57. Wang Y, Hong C-S. Effect of hydrogen peroxide, periodate and persulfate on photocatalysis of 2-chlorobiphenyl in aqueous TiO2 suspensions. Water Res 1999; 33:2031–6.

58. Fernández J, Kiwi J, Baeza J, Freer J, Lizama C, Mansilla HD. Orange II photocatalysis on immobilised TiO2. Appl Catal B Environ 2004; 48:205–11.

59. Linley E, Denyer SP, McDonnell G, Simons C, Maillard JY. Use of hydro-gen peroxide as a biocide: new consideration of its mechanisms of biocidal action. J Antimicrob Chemoth 2012; 67:1589–96.

60. Chopra I. The increasing use of silver-based products as antimicrobial agents: a useful development or a cause for concern? J Antimicrob Chemoth 2007; 59:587–90.

61. Rai M, Yadav A, Gade A. Silver nanoparticles as a new generation of anti-microbials. Biotechnol Adv 2009; 27:76–83.

62. Fung MC, Bowen DL. Silver Products for Medical Indications: Risk-Benefit Assessment. J Toxicol Clin Toxicol 1996; 34:119–26.

63. Chernousova S, Epple M. Silver as Antibacterial Agent: Ion, Nanoparticle, and Metal. Angew Chem Int Ed 2012; 52:1636–53.

64. Rupp ME, Fitzgerald T, Marion N, Helget V, Puumala S, Anderson JR, Fey PD. Effect of silver-coated urinary catheters: Efficacy, cost-effectiveness, and antimicrobial resistance. Am J Infect Control 2004; 32:445–50.

65. Lederer JW, Jarvis WR, Thomas L, Ritter J. Multicenter Cohort Study to Assess the Impact of a Silver-Alloy and Hydrogel-Coated Urinary Catheter on Symptomatic Catheter-Associated Urinary Tract Infections. J Wound Ostomy Continence Nurs 2014; 41:473–80.

66. Svensson S, Suska F, Emanuelsson L, Palmquist A, Norlindh B, Trobos M, Bäckros H, Persson L, Rydja G, Ohrlander M, et al. Osseointegration of tita-nium with an antimicrobial nanostructured noble metal coating. Nanomedi-cine: NBM 2013; 9:1048–56.

67. Wang H, Cheng M, Hu J, Wang C, Xu S, Han CC. Preparation and Optimi-zation of Silver Nanoparticles Embedded Electrospun Membrane for Implant Associated Infections Prevention. ACS Appl Mater Interfaces 2013; 5:11014–21.

70

68. Mei S, Wang H, Wang W, Tong L, Pan H, Ruan C, Ma Q, Liu M, Yang H, Zhang L, et al. Antibacterial effects and biocompatibility of titanium surfaces with graded silver incorporation in titania nanotubes. Biomater 2014; 35:4255–65.

69. Simchi A, Tamjid E, Pishbin F, Boccaccini AR. Recent progress in inorganic and composite coatings with bactericidal capability for orthopaedic applica-tions. Nanomedicine: NBM 2011; 7:22–39.

70. Lan M-Y, Liu C-P, Huang H-H, Lee S-W. Both Enhanced Biocompatibility and Antibacterial Activity in Ag-Decorated TiO2 Nanotubes. PLoS ONE 2013; 8:e75364.

71. Gao A, Hang R, Huang X, Zhao L, Zhang X, Wang L, Bin Tang, Ma S, Chu PK. The effects of titania nanotubes with embedded silver oxide nanoparti-cles on bacteria and osteoblasts. Biomater 2014; 35:4223–35.

72. Xiu Z-M, Zhang Q-B, Puppala HL, Colvin VL, Alvarez PJJ. Negligible Par-ticle-Specific Antibacterial Activity of Silver Nanoparticles. Nano Lett 2012; 12:4271–5.

73. Feng QL, Wu J, Chen GQ, Cui FZ, Kim TN, Kim JO. A mechanistic study of the antibacterial effect of silver ions on Escherichia coli and Staphylococcus aureus. J Biomed Mater Res 2000; 52:662–8.

74. Kim JS, Kuk E, Yu KN, Kim J-H, Park SJ, Lee HJ, Kim SH, Park YK, Park YH, Hwang C-Y, et al. Antimicrobial effects of silver nanoparticles. Nano-medicine: NBM 2007; 3:95–101.

75. Morones JR, Elechiguerra JL, Camacho A, Holt K, Kouri JB, Ramírez JT, Yacaman MJ. The bactericidal effect of silver nanoparticles. Nanotechnology 2005; 16:2346–53.

76. Hajipour MJ, Fromm KM, Ashkarran AA, Jimenez de Aberasturi D, de Lar-ramendi IR, Rojo T, Serpooshan V, Parak WJ, Mahmoudi M. Antibacterial properties of nanoparticles. Trends Biotechnol 2012; 30:499–511.

77. Soenen SJ, Rivera-Gil P, Montenegro J-M, Parak WJ, De Smedt SC, Braeckmans K. Cellular toxicity of inorganic nanoparticles: Common aspects and guidelines for improved nanotoxicity evaluation. Nano Today 2011; 6:446–65.

78. Bosetti M, Massè A, Tobin E, Cannas M. Silver coated materials for external fixation devices: in vitro biocompatibility and genotoxicity. Biomater 2002; 23:887–92.

79. Hardes J, Ahrens H, Gebert C, Streitbuerger A, Buerger H, Erren M, Gunsel A, Wedemeyer C, Saxler G, Winkelmann W, et al. Lack of toxicological side-effects in silver-coated megaprostheses in humans. Biomater 2007; 28:2869–75.

80. Zhao L, Wang H, Huo K, Cui L, Zhang W, Ni H, Zhang Y, Wu Z, Chu PK. Antibacterial nano-structured titania coating incorporated with silver nano-particles. Biomater 2011; 32:5706–16.

81. Page K, Palgrave RG, Parkin IP, Wilson M, Savin SLP, Chadwick AV. Tita-nia and silver-titania composite films on glass-potent antimicrobial coatings. J Mater Chem 2006; 17:95.

82. Mai L, Wang D, Zhang S, Xie Y, Huang C, Zhang Z. Synthesis and bacteri-cidal ability of Ag/TiO2 composite films deposited on titanium plate. Appl Surf Sci 2010; 257:974–8.

83. Li J, Liu X, Qiao Y, Zhu H, Ding C. Antimicrobial activity and cytocompati-bility of Ag plasma-modified hierarchical TiO2 film on titanium surface. Col-loids Surf, B 2014; 113:134–45.

71

84. Uhm S-H, Song D-H, Kwon J-S, Lee S-B, Han J-G, Kim K-N. Tailoring of antibacterial Ag nanostructures on TiO2 nanotube layers by magnetron sput-tering. J Biomed Mater Res 2013; 102:592–603.

85. Ewald A, Glückermann SK, Thull R, Gbureck U. Antimicrobial titani-um/silver PVD coatings on titanium. BioMed Eng OnLine 2006; 5:22.

86. Zhao L, Chu PK, Zhang Y, Wu Z. Antibacterial coatings on titanium im-plants. J Biomed Mater Res 2009; 91B:470–80.

87. Xiang XD, Sun X, Briceño G, Lou Y, Wang KA, Chang H, Wallace-Freedman WG, Chen SW, Schultz PG. A combinatorial approach to materi-als discovery. Science 1995; 268:1738–40.

88. McFarland E, Weinberg W. Combinatorial approaches to materials discov-ery. Trends Biotechnol 1999; 17:107–15.

89. Amis EJ, Xiang X-D, Zhao J-C. Combinatorial materials science: What's new since Edison? MRS bulletin 2002; 27:295–300.

90. Hook AL, Anderson DG, Langer R, Williams P, Davies MC, Alexander MR. High throughput methods applied in biomaterial development and discovery. Biomater 2010; 31:187–98.

91. Mei Y, Saha K, Bogatyrev SR, Yang J, Hook AL, Kalcioglu ZI, Cho S-W, Mitalipova M, Pyzocha N, Rojas F, et al. Combinatorial development of bi-omaterials for clonal growth of human pluripotent stem cells. Nat Mater 2010; 9:768–78.

92. Urquhart AJ, Anderson DG, Taylor M, Alexander MR, Langer R, Davies MC. High Throughput Surface Characterisation of a Combinatorial Material Library. Adv Mater 2007; 19:2486–91.

93. Takeuchi I, van Dover RB, Koinuma H. Combinatorial synthesis and evalua-tion of functional inorganic materials using thin-film techniques. MRS bulle-tin 2002; 27:301–8.

94. Gebhardt T, Music D, Takahashi T, Schneider JM. Combinatorial thin film materials science: From alloy discovery and optimization to alloy design. Thin Solid Films 2012; 520:5491–9.

95. Potyrailo R, Rajan K, Stoewe K, Takeuchi I, Chisholm B, Lam H. Combina-torial and High-Throughput Screening of Materials Libraries: Review of State of the Art. ACS Comb Sci 2011; 13:579–633.

96. Green ML, Takeuchi I, Hattrick-Simpers JR. Applications of high throughput (combinatorial) methodologies to electronic, magnetic, optical, and energy-related materials. J Appl Phys 2013; 113:231101.

97. Lowe G. Combinatorial chemistry. Chem Soc Rev 1995; 24:309. 98. Thompson LA, Ellman JA. Synthesis and Applications of Small Molecule

Libraries. Chem Rev 1996; 96:555–600. 99. Xin B, Jing L, Ren Z, Wang B, Fu H. Effects of Simultaneously Doped and

Deposited Ag on the Photocatalytic Activity and Surface States of TiO2. J Phys Chem B 2005; 109:2805–9.

100. Rupp F, Gittens RA, Scheideler L, Marmur A, Boyan BD, Schwartz Z, Geis-Gerstorfer J. A review on the wettability of dental implant surfaces I: theoret-ical and experimental aspects. Acta Biomater 2014; 10:2894–906.

101. Forsgren J, Svahn F, Jarmar T, Engqvist H. Formation and adhesion of bio-mimetic hydroxyapatite deposited on titanium substrates. Acta Biomater 2007; 3:980–4.

102. Sung-Suh HM, Choi JR, Hah HJ, Koo SM, Bae YC. Comparison of Ag dep-osition effects on the photocatalytic activity of nanoparticulate TiO2 under visible and UV light irradiation. J Photoch Photobio A 2004; 163:37–44.

72

103. Seery MK, George R, Floris P, Pillai SC. Silver doped titanium dioxide na-nomaterials for enhanced visible light photocatalysis. J Photoch Photobio A 2007; 189:6–6.

104. Horakova M, Martin N, Aubry E, Spatenka P. Photocatalysis of Ag doped TiOx films prepared at room temperature. Catalysis letters 2009; 132:244–7.

105. Limage H, Tichelaar FD, Closset R, Delvaux S, Cloots R, Lucas S. Study of the effect of a silver nanoparticle seeding layer on the crystallisation tem-perature, photoinduced hydrophylic and catalytic properties of TiO2 thin films deposited on glass by magnetron sputtering. Surf Coat Tech 2011; 205:3774–8.

106. Lindenauer KG, Darby JL. Ultraviolet disinfection of wastewater: Effect of dose on subsequent photoreactivation. Water Res 1993; 28:805–17.

107. United States Environmental Protection Agency. Ultraviolet disinfection guidance manual for the final long term 2 enhanced surface water treatment rule. http://www.epa.gov/ogwdw000/disinfection/lt2/pdfs/guide_lt2_uvguid-ance.pdf; 2006.

108. Crowley DJ, Courcelle J. Answering the Call: Coping with DNA Damage at the Most Inopportune Time. J Biomed Biotechnol 2002; 2:66–74.

109. Erill I, Campoy S, Barbé J. Aeons of distress: an evolutionary perspective on the bacterial SOS response. Fems Microbiol Rev 2007; 31:637–56.

110. Jung WK, Koo HC, Kim KW, Shin S, Kim SH, Park YH. Antibacterial Ac-tivity and Mechanism of Action of the Silver Ion in Staphylococcus aureus and Escherichia coli. Appl Environ Microbiol 2008; 74:2171–8.

111. Ribeiro M, Monteiro FJ, Ferraz MP. Infection of orthopedic implants with emphasis on bacterial adhesion process and techniques used in studying bac-terial-material interactions. Biomatter 2012; 2:176–94.

Acta Universitatis UpsaliensisDigital Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Science and Technology 1250

Editor: The Dean of the Faculty of Science and Technology

A doctoral dissertation from the Faculty of Science andTechnology, Uppsala University, is usually a summary of anumber of papers. A few copies of the complete dissertationare kept at major Swedish research libraries, while thesummary alone is distributed internationally throughthe series Digital Comprehensive Summaries of UppsalaDissertations from the Faculty of Science and Technology.(Prior to January, 2005, the series was published under thetitle “Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Science and Technology”.)

Distribution: publications.uu.seurn:nbn:se:uu:diva-249181

ACTAUNIVERSITATIS

UPSALIENSISUPPSALA

2015