Second year Report
AN ATOMIC FORCE MICROSCOPY INVESTIGATION OF THE INTERFACIAL PROPERTIES OF BIOCOMPATIBLE NANOSTRUCTURED TITANIUM OXIDE FILMS
Introduction
Metal oxides, like Titanium Oxide (TiOx) are known for their biocompatibility with various cell lines or as a standard support for microarrays. Metal oxides used here are primarily nanostructured and are deposited by pulsed microplasma cluster source (PMCS) [1, 27]. Clusters produced from PMCS are deposited over a substrate, producing nanaostructured thinfilms thickness having thickness of tens of nanometer. In general, nanostructured materials may be defined as those materials whose structural elements - clusters, crystallites or molecules—have dimensions in the 1 to 100 nm range [2]. Biocompatibility is mainly investigated by two well established methodologies. First method is to expose a surface to a particular cell line and observe the growth of cells. Primarily, lysine treated coverslips are used as standard support for growing various cell lines or chemically functionalized surfaces for developing protein or DNA based microarrays. Carbone et alshowed the biocompatibility of nanostructured titanium dioxide (ns-TiOx) film. They spotted an array of streptavidin molecules for a retroviral microarray on to a TiO2 nanostructured film. Previously work shown compatibility of TiO2 with a various human cell lines [3,4]. Second method is to measure protein adsorption followed by verification of their bioactivity. The question to be addressed here is what makes a surface biocompatible? Is it possible to describe interfacial properties of a surface which are essential for its biocompatibility? At molecular level a protein surface exhibits polar, non-polar or charged functional groups. Of this oxygen, nitrogen and other electronegative elements in various functional groups actively take part in hydrogen bonding, whereas non-polar and aromatic groups are hidden inside, away from polar water molecules. Measurement of such single molecule interactions can be carried out by using Atomic Force Microscopy (AFM) and with the development of force spectroscopy it is possible to study strength of bonding between different functional groups. As an example, Lieber et al performed Chemical Force Microscopy (CFM) by chemically modifying tips and substrates with –CH3 and –COOH groups. Their experiment involved both force spectroscopy and friction measurements. Their results showed that magnitude of the adhesive interactions between tip/sample functional groups decreases in the following order: COOH/COOH > CH3/CH3 > CH3/COOH [5].
Likewise many techniques are available to differentially functionalize a surface for attaching directly or by means of linkers various functional groups and/or proteins. When CFM is used under force volume mode for mapping target molecules or functional groups on a surface then this technique is know as Single Molecular Force Recognition Microscopy (SMFRM) [6].Force Volume Mode (or Imaging) involves taking Force vs. Distance (FD) curves over every point corresponding to a pixel of the AFM image. Both approaching and retracting part of the curve can provide crucial information about the interplay of molecular interactions between two different functional groups or surfaces, in our case Si3N4 tip against glass and ns-TiOx.
In force volume mode, one can map physio-chemical heterogeneties in the substrate. Many research groups have reported using SMFRM for mapping molecular recognition sites over cellular surfaces [7] or for exploring strength of molecular interactions and visualizing single molecules [8] both in air and in solution.
The aim of this project is to uncover interaction between metal oxide nanostructured surface with proteins and various functional groups like -COOH, -OH, etc. This study would help us to understand underlying mechanism of protein adsorption and in developing superior biocompatible materials having myriad of biomedical applications. The results presented here demonstrate the technique of force spectroscopy for investigating distribution of possible protein binding sites over a biocompatible ns-TiOx surface. Direct measurement of adhesion between AFM tip (Si3N4) and ns-TiOx and between AFM tip and a reference surface like piranha cleaned glass surface provide us good control against corrugated TiOx surface. The adhesion of Silicon Nitride (Si3N4) tip over reference glass can help us interpret the role of morphology and chemistry of our target material. Issue of having a reference material against the target TiOx surface is solved by coupling Nanosphere lithography (NSL) with Supersonic Cluster Beam Deposition (SCBD). In NSL, masks are made byself-assembly of polymeric nanospheres. A two-dimensional (2D) arrays produced after material deposition are used to study magnetic dots arrays based on magnetic metals or alloys, localized Surface Plasmon Device (LSPR) devices for biosensing, etc [9,10]. Here we have tried to pair the technique of NSL with SCBD. By producing 2D array of ns-materials we can have many different reference substrates for investigating nanoscale heterogeneities.
Materials and Methods
Sample preparation
Glass coverslips were cleaned by sonicating with piranha (1:3 30% H2O2:H2SO4) solution for 10 minutes. Original nanosphere (3 micron diameter) solution obtained from duke scientific is diluted twice with 1:400 Triton-x: Methanol solution. 10 microlitre of the diluted solution is used for spin coating spheres over piranhna cleaned glass coverslips. The nanospheres were allowed to dry in ambient conditions to form a 2D hexagonally close packed array [9].TiOx is then deposited over the sphere mask.
Deposition
An ns-TiOx film is deposited (Fig. 2 & 3) by a SCBD apparatus equipped with a PMCS. The PMCS operation principle is based on the ablation of a titanium rod by a helium plasma jet, ignited by a pulsed electric discharge. After the ablation, TiOxions thermalize with helium and condense to form clusters. The mixture of clusters and inert gas is then extracted in vacuum through a nozzle to form a seeded supersonic beam, which is collected on a set of standard glass microscope slides located in the beam trajectory. The clusters kinetic energy is low enough to avoid fragmentation and hence a nanostructured film is grown. Rms roughness of ns- TiOx films can be typically controlled during deposition in the range 2-40 nm, with corresponding specific areas (the ratio of the surface to the projected area) in the range 1-2 [11,12]
AFM-Force Spectroscopy
The sphere mask is removed form the substrate by sonication in milli-q water. After the removal of the spheres, sample is imaged by AFM in tapping mode in order to visualize nanoscale patterns. This is followed Force Volume (FV) imaging of the as-deposited and annealed samples both in air and in milli-q water. Milli-Q water refers to ultrapure laboratory grade water that has been filtered and purified by reverse osmosis. It’s free of any dissolved ions which may passivate the TiOx surface during longer measurements. Force spectroscopyinvolves taking FD curves over every point corresponding to a pixel of the AFM image. FD curves were acquired on a three micron size area containing two triangular islands of ns-TiOx along with reference glass substrate. Resolution is kept to 128 points per line and measurements were done at 4 Hz scan rate.
Fig. 1 – A typical Force vs. Distance (FD) curves. Data is initially obtained in form of Deflection Error (DE) Vs. Z (nm). Later on, DE is converted to Force of adhesion.
For calculating, Force of adhesion (Fad) force constant (k) of the cantilever is multiplied with Z-sensitivity (Zsens) and Deflection Error (DE) signal -
Fad = k (nN/nm) x Zsens (nm/V) x D.E. (V) (1)
The above equation is derived a Hook’s law-
F = - k.δ x (2)
Where δ x = Zsens. x D.E., k is the force constant and F is the force experienced by the cantilever.
Data obtained from FV maps was analyzed using Matlab routines and force measurements on ns-TiOx and glass surface are segregated for acquiring average value of force of adhesion.
AFM cantilevers
Gold coated Contact Si3N4 cantilevers (DNP-20) were purchased from Digital Instruments-Veeco Metrology (Santa Barbara, CA). They were used as received from the company. Spring constant of cantilevers used were determined using thermal noise method [26]. Before start of each experiment, Z-sens is calculated on hard substrate like glass coverslip both in air and water.
Result and Discussion
Nanopatterned ns- TiOx Thin Films
On removal of nansospheres by sonication, patterning was studied by imaging the substrate in tapping mode. Size of the patterns obtained by NSL extends to few hundred microns when seen under an optical microscope.
Fig.2 – Deposition of nanopatterned ns-TiOx by coupling SCBD with NSL
5 um x 5 um; Z – 100 nm
Fig.3 – A high resolution image of Titanium dioxide pattern.
Length of single triangular island is approximately around 700 nm.The main objective of developing such patterns is to have reference and target material simultaneously in an experiment. The presence of reference during an investigation of the target material can help us in resolving any ambiguity which may not be apparent when reference and target material are investigated independently.
Mapping Molecular Forces
FV mapping was carried out with Si3N4 in air and in milli-q water on annealed and as-deposited sample. Simultaneous topography and adhesion force mapping gave us 128 x 128 force curves in a 3 micron size area. Concurrent study of height and adhesion maps revealed distribution of Fad between TiOx and glass surface and independent quantification of molecular forces between the tip and the substrate. An example is shown in the fig. 3.
Fig.4 - A typical Topo-adhesion map in aqueous medium. Here we can realize contrast between the two types of measurements. TiOx islands have shown different Fad against glass surface. At some points (black arrows) we cannot identify any feature on topography but disparitiesin chemistry of the substrate are evident in adhesion map.
Fad on glass and TiOx is segregated from each other and average values of adhesion are obtained for interpreting molecular forces in-correlation to surface morphology. Thereafter, ratio of TiOx against glass (TiOx/Glass Table 1 & 2) is calculated for identifying any ambiguities during the measurements. Differences in ratios can help us to pinpoint faulty measurements which could be due to dirty AFM tip or substrate, large changes in humidity during air measurements, incorrectly functionalized substrate or AFM tip, high background noise which can overshadow large number of force curves in an experiment, etc.
Probing Measurements in Air
Measurements in air showedhigher forces of adhesion on glass. Whereas this effect was reversed in milli-q water, titanium showed higher force of adhesion in aqueous medium. Data gathered from various histograms in air is shown in the Table 1.
Fig.5 – Topography and adhesion maps of annealed and as-deposited substrate.
Table - 1
Air / TiOx (nN) / Glass (nN) / TiOx/GlassAsdep / 66 ± 2.64 / 74.66 ± 3.05 / 0.88 ± 0.05
Annealed / 68.33 ± 0.57 / 74.33 ± 2.51 / 0.91 ± 0.03
When AFM tip is interacting with a surface, it experiences a force of adhesion which is a combination of electrostatic force Fel, the van der Waals force Fvdw, the capillary force Fcap and forces due to chemical bonds or acid–base interactions Fchem [13]:
Fad =Fel +Fvdw + Fcap +Fchem:(3)
While imaging in air, capillary force is the major force contributing to adhesion onto a flat substrate. Meniscus formation is the consequence of reduced vapour pressure at the curved surface. This leads to capillary condensation and plays a major role in creating a force of attraction between tip and the substrate [13]. Fig.6- Schematic of capillary column
The resulting capillary force Fcap between a plate and a sphere with radius R has been calculated by O’Brien and Hermann [14] to be
Fcap = 2πRγ (cosθ1 + cos θ2),(4)
where R is the tip radius, γ is the surface tension of the liquid, θ1 and θ2 are the contact angles between the two surfaces and the liquid.
We can see that, through cosθ1 the surface energies of biocompatible oxides have entered the equation. Cosθ1 is measure of the wetability of the surface.
Wetability of a flat homogeneous surface like glass is explained by Young equation [15] -
Cosθ = γsv – γsL(5)
γ
whereθ is the contact angle between liquid and surface and γ, γsv and γsL are the surface tensions at the liquid/ vapor, solid/vapour and solid/liquid interfaces, respectively.
For describing Fad of a rough surface like ns-TiOx, we move into Wenzel regime [17] where the Cosθ1 (θ1 - apparent contact angle on a rough surface) is more accurately explained by roughness factor (r) (r = A/A0; where A is the ‘true’ surface area, A0 the nominal area).
Cosθ1 = r Cosθ(6)
Our collegues in CIMAINA have collected data on wettability of ns-TiOx and crystalline TiOx surface [16]. Crystalline TiOx surface is intrinsically hydrophobic in nature and hydrophobicity increases as we increase roughness of ns-TiOx. Thus, we can state that larger the roughness, larger is the specific area, larger is the contact angle and more hydrophobic is the surface.
On ns-TiOx we experience lower adhesion than on glass, which is more hydrophobic as compared to piranha cleaned coverslip. On one side piranha treated glass is very hydrophilic where TiOx is intrinsically mildly hydrophobic, on the other there is a prominent role played by morphology. A rough morphology can hamper a complete meniscus formation and apparent angle appears to be larger, thereby reducing overall force of adhesion. For estimating changes in the wettability of annealed ns-TiOx, one needs to collect more data to get better statistics for estimating smaller difference in Cos θ1.
Probing Measurements inMilli-q Water
Fig..7 – Topography and adhesion maps of annealed and as-deposited substrate in Milli-q Water
Table - 2
Milli- Q Water / TiOx(nN) / Glass(nN) / TiOx/GlassAsdep / 0.29 ± 0.05 / 0.08 ± 0.009 / 3.625 ± 0.74
Annealed / 0.45 ± 0.32 / 0.13 ± 0.007 / 3.461± 2.46
Experiments performed in milli-q water (pH-4.9) showed higher adhesion on ns-TiOx as compared to glass coverslips. This effect can be explained by examining chemical nature of the substrate and AFM tip. Our treatment of coverslips with piranha introducedmany silanol groups whichcan interact with silicon nitride (Si3N4) surface. A virgin silicon nitride surface primarily consists of nitrogen bonded to three silicon atom. In aqueous medium, Si3N4 surface undergo hydrolysis to produce silanol and basic silylamine (secondary and/or primary) surface groups [23, 24].
Hoh et al measured strong force of adhesion on glass surface [18]. Depending on the annealing or chemical treatment, silica surfaces can have minimal 4.6 ± 0.2 hydroxyl groups per nm2[25]. But still we have observed higher adhesion over titanium in comparison with our reference glass. The ratio of adhesion on TiOx is 3-4 times (Table -2) higher as compared to glass.
Fig.8 -Schematic of the twopredominant typesof hydrogen bonds likely to form between an oxidized silicon nitride surface and glass. (c) At low pH most hydroxyl groups are protonated resulting in little electrostatic repulsion, and a large number of bonds. (d) At pHs over 9most of the hydroxyl groups are not protonated. Even allowing for a shift upward of the hydroxyl group pKa as the surfaces are brought together, resulting in electrostatic repulsion between the surfaces and very few bonds [18].
The possible explanation could be ability of titanium to form co-ordinate bonds with nucleophiles like oxygen or nitrogen and coordinate bonds are much stronger then Hydrogen-bonds. As-deposited and annealed titanium surface can co-ordinately bound with water molecules (physisorbed) and hydroxyl (chemisorbed) groups present in an aqueous medium. On annealed titania, there can be more number of hydroxyl groups as water under goes dissociative chemisorption at “empty” (five-coordinated) Ti sites [19].
A review of electrokinetic and titration experiments, for titanium dioxide gives the following average values [20]
RutilePZC 5.3 IEP 4.8 or 5.6
Anatase PZC6.2 IEP 6.1
PZC is "point of zero charge" more correctly defines the situation when there is net zero charge on the surface, and is readily determined by potentiometric titrations. The pHPZC is a very important value for adsorption measurements and surface characterization. It plays a crucial role in understanding sorption of protons and hydroxyl groups which depends on acid- base properties of the surface. Whereas Isoelectricpoint (IEP) characterizes a state of a surface such that electrical potential at the slip plane (the thickness of the solvent/ion film which moves with the particle) inside the double electric layer is equal to zero. Hence, due to the presence of slip plane, surface IEP and PZC change in the opposite direction on sorption of ions over the surface [20, 21].
In our case, both as-deposited and annealed titanium is amorphous in nature and in milli-q dissolved ionic species are almost absent. Under such conditions we may consider pHPZC = pHIEP and investigate chemical properties of nascent ns-TiOx surface. IEP of amorphous TiOx has range of 3.5-6.7 and we are working with milli-q water at pH 4.9. Therefore, the surface or TiOx will be positively charged or neutral. At low pH, all the functional groups on the tip will be protonated and tip will experience lower level of repulsive interactions against the TiOx surface.
Fig.9 - Formation of chelation and H-bonding of incoming neucleophilic groups with cationic Ti sites on TiOx surface.
Hence, incoming nucleophiles like nitrogen or oxygen on Si3N4can displace adsorbed molecules and form coordination complex at Ti+ sites (Fig. a). In additon, hydrogen atoms linked to oxygen and nitrogen form hydrogen bonding with oxygen atom on the substrate (ns-TiOx). Therefore, on adding strength of coordinate bonds with H-bonds, we observe overall higher adhesion over titania as compared to glass. This effect is more enhanced over corrugated surfaces where tip can interact with more number of sites because of the higher effective surface area, thus forming a stronger bond in comparison to a silanol surface.
The same concept can be extended to protein molecules. Surface of a protein molecule contains many groups like – OH, -COOH,-NH2, etc. They can effortlessly take part in co-ordinate and hydrogen bonding over metal oxide surface. Coordinate bonds are as strong as covalent bond plus formation of hydrogen bond can produce irreversible binding of proteins with ns-TiOx.Therefore, for having deeper understanding on single molecule - single group interactions, AFM tips will be functionalized (This is briefly explained in the following sections) for elucidation of mechanism involved in protein adsorption.