electrowetting (EWOD) as an
Anti-fouling process for antibacterial
surfaces
STAVROS YIKA
(EWOD) as an Anti-fouling process for
antibacterial surfaces
Stavros Yika
Master’s Degree Project
Supervisor: Fredrik Carlborg
Examiner: Wouter Van der Wijngaart
June 2014
Microsystem Technology
Today the medical field is struggling to decrease bacteria biofilm formation which leads to infection. Also, biomedical devices sterilization has not changed over a long period of time which has resulted in high costs for hospitals healthcare managements. The objective of this project is to investigate electro-dynamic effects by surface energy manipulation as potential methods for preventing bacteria biofilm growing on medical devices.
Based on electrokinetic environments two different methods were tested: re-jection bacteria dielectrophoretic forces feasibility by numerical simulations; and electrowetting-on -dielectric by the fabrication of golden interdigitated electrodes on silicon glass substrates covered by a Teflon layer.
In the first experiment, numerical simulations of gold electrodes in buffer so-lution and frequencies were carried out to determine the forces required to reject bacteria. In the second experiment, interdigitated gold electrodes coated with a dielectric Teflon layer, were characterized in terms of breakdown voltage, dielectric adhesion and contact angle in terms of applied voltage. Finally the effect of EWOD on bacterial adhesion was tested.
Abbreviations
AC Alternate current CM Clausius-Mossotti DC Direct current DEP Dielectrophoresis EDL Electrical Double-Layer EP ElectroPhoresisEW ElectroWetting
EWOD ElectroWetting-on-Dielectric HCAI Healthcare Associated Infections
M9 M9 minimal medium - minimal microbial growth medium nDEP negative DEP
OSTE off-stoichiometry thiolene PDMS PolyDimethylSiloxane PBS Phosphate buffered saline pDEP positive DEP
UV Ultraviolet light
1 Introduction 8
1.1 Project Motivation . . . 8
1.2 Research Idea . . . 11
2 Background and State of Art 12 2.1 Biofilm formation and prevention . . . 12
2.2 Theory themes (DEP-EWOD) . . . 15
2.2.1 Dielectrophoresis Force . . . 15
2.2.2 Dielectrophoresis in DC and AC Fields . . . 16
2.2.3 EWOD in Direct current (DC) and Alternate current (AC) fields . . . 18
2.2.4 Theory limitations . . . 19
2.3 Modelling of biological cells . . . 19
2.3.1 Biological cells on DEP . . . 19
2.3.2 Hydrophobic effect on bacteria adhesion . . . 21
2.4 Modelling of electrodes . . . 21
3 Design and assembly of the experimental setup 24 3.1 Chip design and assembly . . . 24
3.1.1 Electrode design . . . 24 3.1.2 Chip fabrication . . . 27 3.2 Gasket fabrication . . . 36 3.2.1 OSTE Gasket . . . 37 3.2.2 PDMS Gasket . . . 41 3.2.3 Leakage tests . . . 43 3.3 Bacteria to be used . . . 45 3.3.1 Salmonella Typhimurium . . . 45
4 Measurements and results 47 4.1 Numerical simulations of DEP on bacteria model . . . 47
4.1.1 MATLAB calculations . . . 47
4.1.2 COMSOL analysis and results . . . 53
4.2 Experiments of EWOD surfaces on live bacteria . . . 64
4.2.1 Contact angle tests . . . 64
CONTENTS
4.2.3 Teflon adhesion to glass . . . 68
4.2.4 Teflon breakdown voltage tests . . . 71
4.3 Bacteria tests . . . 73
4.3.1 Teflon bacteria adhesion tests . . . 76
5 Discussion and outcome 78 5.1 Discussion . . . 78
5.2 Outcome . . . 82
5.2.1 Project method summary . . . 83
5.2.2 Outlook . . . 83
A Biofilm on hospital environments challenges 92 B MATLAB Program 94 B.1 Command set . . . 94
B.2 Flow Chart Diagram . . . 96
C COMSOL analysis and simulations 98 C.1 Drift velocity from Einstein– Smoluchowski relation . . . 98
C.2 DEP force measured at different heights . . . 99
D Bacteria formation 100 D.1 Bacteria . . . 100
D.1.1 ImageJ macro . . . 100
D.2 Medium . . . 101
E OSTE tests detailed information 102 E.1 Material specification . . . 102
F Chip fabrication detailed information 103 F.1 Design / L-EDIT . . . 103
F.2 Procedures . . . 104
F.3 Parameters . . . 106
2.1 Attachment of bacteria cells mechanism . . . 13
2.2 DEP variations by its applications . . . 15
2.3 Effect of ~E on particle . . . 17
2.4 Interdigitated electrode distance effect on Electric field[1] . . . 18
2.5 DEP cell modelling[2] . . . 20
2.6 Hydrophobicity and attach cell number (RP62A) relationship . . . . 21
2.7 Commercial Electrode-gasket sample for cells and/or proteins . . . . 22
2.8 Different existing Dielectrophoresis (DEP) devices classification . . . 23
2.9 ElectroWetting (EW) electric fields and equivalent Electric circuit . 23 3.1 Electrodes configuration . . . 25
3.2 Electrodes in square spaces . . . 26
3.3 Mask and expectable wafer design . . . 27
3.4 Wafer material layers . . . 28
3.5 Process of Photo-lithography . . . 29
3.6 Metal Mask Measurements . . . 31
3.7 Teflon technique to insulate on only electrodes . . . 33
3.8 Finished chip . . . 34
3.9 Etching time problem . . . 35
3.10 Electrode bad etching finish . . . 35
3.11 Commercial gasket and design . . . 36
3.12 Gaskets designs . . . 37
3.13 OSTE gasket on aluminium mould . . . 39
3.14 One hole gasket procedure . . . 40
3.15 Different made PolyDimethylSiloxane (PDMS) gaskets . . . 43
3.16 OSTE and PDMS gaskets on sealing tests . . . 44
3.17 OSTE gasket adhesion and leak . . . 44
3.18 Salmonella typhimurium . . . 46
4.1 Excel & Matlab CM plots . . . 49
4.2 CM on diff. mediums . . . 50
4.3 Different CM values for biological mediums for yeast and E.coli par-ticles within various frequency values . . . 52
4.4 Electric potential intensity . . . 54
LIST OF FIGURES
4.6 DEP Force . . . 58
4.7 COMSOL DEP simulation - Electric field analysis . . . 61
4.8 Maximum bacteria drift rejection velocity based on applied voltage . 62 4.9 EWOD setup for Contact angle measurements . . . 64
4.10 Contact angle by evaporation . . . 66
4.11 Contact angle within different voltages . . . 67
4.12 Waveform tests . . . 68
4.13 Teflon peel off from glass surfaces . . . 70
4.14 Teflon breakdown tests . . . 72
4.15 Setup used for bacteria tests . . . 73
4.16 Electrodes’ cells numbered and labelled . . . 74
4.17 Bacteria tests stages . . . 75
4.18 Bacteria . . . 77
5.1 DEP force (Newton)at 30 V at the edges . . . 79
5.2 Squared Electric field gradient affected by Teflon layer . . . 80
5.3 Droplet contact angle change & electrolysis . . . 81
5.4 Bacteria tests on with interdigitated electrodes . . . 82
B.1 Matlab flow chart . . . 97
C.1 DEP forces at different heights . . . 99
D.1 Bacteria biofilm matrix forces . . . 100
F.1 Different electrodes topology . . . 103
F.2 Electrode design attempt . . . 104
G.1 EWOD variety test . . . 108
1.1 Costs due to antibiotic resistant bacteria[3] . . . 9
1.2 Current Antibacterial surfaces methods . . . 10
2.1 Bacteria adhesion earlier experiences . . . 14
3.1 Different width of the electrode depending on the case used . . . 25
3.2 Photoresist curing time parameters . . . 28
3.3 Different used etching times on lithography . . . 30
3.4 Spindle and Feed speeds for the different materials . . . 30
3.5 Glass Blade parameters . . . 31
3.6 Silicon Oxide Blade parameters . . . 31
3.7 OSTE Formulation: Stoichiometry, monomers and curing agent . . . 38
3.8 Curing time for 2 mm OSTE . . . 41
3.9 Curing time for 1 mm OSTE . . . 42
3.10 PDMS Formulation: base and curing agent ratio . . . 42
4.1 Three-layer particle parameters . . . 48
4.2 DEP force calculated from COMSOL simulations . . . 63
4.3 Drift velocity for different bacteria . . . 63
4.4 Diffusion length of bacteria during one second . . . 63
4.5 Bacteria adhesion tests done in Karolinska laboratories . . . 76
D.1 M9 medium parameters . . . 101
E.1 Culture plate parameters . . . 102
F.1 UV exposure and development times - Second lithography attempt with standard softbake(110◦C) . . . 105
F.2 UV exposure and development times - Second lithography attempt with lower softbake(90◦C) . . . 105
F.3 Measurements of the electrodes and gaps in the second attempt from Tab.F.2 . . . 105
Chapter 1
Introduction
The aim of this chapter is to introduce the topic of the thesis and explain in further details the project definition, goals and the motivation behind it.
1.1
Project Motivation
From ancient times until nowadays the absence of antibacterial surfaces has been present, its demand has been increasing and this is a current problem.
It is widely known microorganisms defence from other microorganisms’ has been to produce their own chemical substances, this is used today as antibiotics or antimi-crobial agents. Unfortunately our body is not as resistant as these microorganisms, making us more vulnerable. Indeed, this can be seen in infections caused by trans-mission of resistant microbes that lives in other’s patients systems like in hospital environments.
Antibiotics are becoming less effective while more resistant microorganisms are spreading all over the world. Antimicrobial resistance is mainly driven by healthcare practices, where antimicrobial are overdosed in patients. Indeed they do not need exactly that, as refereed in [4],[5]. For instance, a multi-drugresistant or-ganism (MDRO), also known as Superbug, is often resistant to fluoroquinolones[6]. Nowadays, places were healthcare is not well supported are still struggling with deaths caused by infections, and in more supported healthcare places antimicrobial resistance is the major cause of deaths, making antibiotics practically ineffective.
The main area of need is in hospital environments, where bacteria can grow almost on any medical instrument. Operating rooms (OR) must be a very ster-ile environment and antimicrobial resistance make it vulnerable to eventually be unsterilised surgeries; in fact, 80% of most common hospital acquired infections are associated to catheters, such as urinary ones[5]. Likewise, in the USA, 5-10% of hospital patients suffers healthcare-associated infections (HAIs), where 35% of them are urinary tract infections related to catheters.
incidences of infections.
From the public health point of view, due to a misused way of healthcare practices highly resistant bacteria have been created, which represent high costs in hospital-related infections solutions.
Hospital-acquired infections has become a major cost due to time and logistic invested in the use of disinfectants in hospital environments as can be seen in Table 1.1.
Hospital-acquired Infections Costs (US$)
USA $10 billion per year
Mexico $450 million per year
Thailand $40 million per year
Table 1.1: Costs due to antibiotic resistant bacteria[3]
The main problem of antimicrobial resistance lies in the bacteria infection treat-ment dependence. Consequently, new methods to avoid bacterial infections cannot prevent antimicrobial resistance from becoming a bigger problem, but help to step down as the most important public health concern[7].
On the other hand, what happens if this antibiotic dependence ends? There have been incredible reductions of use in antibiotics in the EU countries due to a ban on some of the antimicrobial growth promoters (AGP), which reduce antibacterial resistance on food germs[8].
Bacteria can survive several days on surfaces, and the only way to reduce biofilm formation is cleaning and disinfecting the used surfaces. Time to inactivate virus can range from five minutes to more than one day depending on the concentration of the inoculants and the disinfectant[9].
Reusable medical devices that are in continuous mucous contact should receive high-level disinfection (explained further in section A) between different patients, including endoscopes, Endotracheal Tube or Breathing Tube(ETT), anaesthesia circuits, and ventilators). In addition, some prosthetics, inner devices and operation theatres surfaces such as surgical tables need chemical sterilization instead of high temperature sterilization due to the material degradation. In this cases, better engineered materials are needed to suppress bacteria growth like covered fibers or wet fabrics with antibacterial substances.
Biofilms, mainly in endotracheal tubes is eliminated mechanically or by silver coating which is known to kill bacteria. Since the mechanical release of attached bacteria requires manual cleaning, trained personnel is required adding costs to hospitals bill. The metal coating can also wear off and become dangerous if swal-lowed.
1 Introduction
The majority of current solutions release an active ingredient that interacts with the microorganisms: either inhibiting from spreading or killing them.
CONTACT KILL RELEASE
Polymer grafts Antimicrobial
peptides (AMP) Silver or copper
Bactericides (i.e.chlorhexidine) Antibiotic coatings (i.e. gentamicin) Mechanism Covalent attach-ment of polymers with bacteria re-pellent or toxic groups. Disrupts the cell membrane, attaches by electrostatic attraction.
Silver ions pre-vent formation of enzymes and cause cell death. Copper mech-anism is not well under-stood. Prevents initial adhesion of bac-teria. Broad spectrum antibiotics in-corporated into surface coatings. Advantage Covalent bond, no leaching. Prevent surface adhesion or kill bacteria. Broad spectrum antibiotic effect. Can be covalently bonded to sur-face: Leach free. Unknown to cause antibiotic resistance
Can kill from a range.
Disadvantage Surface loading.
Difficult to stabilise in coatings. Leaches leads to detrimental effect on the environment. Difficult to control coating release at long term, leaching. Cause antibiotic resistance.
Table 1.2: Current Antibacterial surfaces methods
From Table 1.2, can be seen that releasing methods become a toxicity problem for living beings and the environment. The major problem with both techniques is that killed microorganisms remain in the surface making easy target for new bacteria to attach and blocking surface-active agents from diffusing. Consequently antibacterial surfaces become gradually inactive.
Previously all approaches to prevent health care related infections have centred around removing or killing the bacteria adhering on the surface of the devices. These have several drawbacks such as eventual build up of a biofilm consisting of dead bacteria and/or long term negative effects such antibiotic resistance caused by catheters coated with antibiotic substances. Instead the electrokinetic approach in-vestigated in this thesis focuses on creating a sufficiently unattractive environment on the surface as to completely prevent bacteria from attaching.
The solution would be to create a non-toxic surface that prevents bacteria to attach, reproduce and form biofilm. In fact, once bacteria biofilm is adhered into the surface, detaching it requires an enormous force.
gradient, and surface dielectric capacitance in an array of interdigitated electrodes.
The practical motivation behind the project is to investigate possible solutions for preventing bacteria biofilm formation in healthcare devices, such us on joint prosthesis (orthopaedic), on urinary catheters and on airways intubation tubes, where bacteria become the cause of Healthcare Associated Infections (HCAI) and Viral respiratory infections (VRI).
1.2
Research Idea
The research idea of this thesis is creating antibacterial surfaces by the use of interdigitated electrodes methods for preventing the bacteria to attach on medical devices or healthcare surfaces thus avoiding biofilm formation without the use of antibiotics or release of chemicals.
For that reason, different methods will be investigated to evaluate their effi-ciency to prevent bacteria adhesion: Electrophoresis (EP), Dielectrophoresis (DEP) and Electrowetting (EW). For instance, DEP can exert enough forces into particles depending on the dielectric values and medium. This rely on a very wide range of frequencies which define an attraction or repulsive force from the surface. Hence prevent bacteria to reach the surface and form biofilm.
On the other hand, ElectroWetting-on-Dielectric (EWOD) can create particle displacement as it can be done with liquid droplets depending on the applied voltage and the thickness of the dielectric layer on top of the surface. Therefore bacteria biofilm might be not only disrupted, from adhesion and further biofilm formation, but capable of be manipulated and transported along the surface.
Chapter 2
Background and State of Art
In this chapter current anti-biofouling technologies history is reported. Further-more, the theory themes behind the experiments are presented. Additionally, this chapter provides a basic theory about EW and DEP.
2.1
Biofilm formation and prevention
Surfaces have an important role in the micro-organism activity; biofilm formation is formed as a defence from adverse conditions (low nutrients, extreme temperatures). Adherence to the surface happens by the secretion of extracellular polysaccha-ride substance from cells (known as slime).
The biofilm formation sequence starts with the conditioning of the surface, followed by the colonisation of the bacteria and finally its multiplications. This microenvironment can increase bacteria growing if two factors are ruled: either by the removal of microcolonies parts and future growth in farther places, or increased by fluids flows that carries nutrients.
Biofilm’s structure is very similar to a pile of Lego pieces where bacteria canR
overlapped each other in different spots and orientations, making easy to intercon-nect themselves by its flagella and cellulose (like bricks and cement) in a common secretion medium. By this, bacteria is allowed to breath and be protected from outer biocides. Factors that increased the formation come from nutrients presence, to warm temperature, and flows stagnation on fluids channel dead-ends[12]. Microbial adhesion mechanism, DLVO theory In a dielectric medium, ions are dispersed all over it. But there will be a gradient of ions density where a in-terface is contained. For example, when a charged particle is introduced in this medium counter ions begin to screen all over its surface by self-organization man-ner however not in an uniform way. In fact, counter ions structured in layers with different characteristics which behaviour is explained through Debye-H¨uckel theory[13].
so called Stern layer[15] which is located in the Inner Helmholtz Plane (IHP). The second layer is composed of a diffuse ion layer, which is attached by Coulomb forces. This two equalled opposite charged layers are defined as an Electrical Double-Layer (EDL). It is important to consider the distance that ions affect beyond the EDL, which is called Debye Length, for phosphate-buffered saline (PBS) is 0.7 nm.
The kinetic stability between energy interactions of particles is explained by Derjaguin-Landau-Verwey-Overbeek (DLVO) theory[13]. This theory combined the van der Waals attraction forces and the electrostatic repulsion forces at low sur-face potentials (potential energy <<thermal energy) considering EDL limits[16]. First, bacteria cell is transported to the surface by sedimentation forces and hydro-dynamic forces (drag force), until reaching the Diffusive boundary layer illustrated in Fig 2.1. At this point the only main force is the diffusion, where its diffusion transportation behaviour is purely by Brownian motion; here DLVO forces are present between surface and the bacteria cell. However, the cell attachment is re-versible due to still week surface interaction. This interaction range is relative small (< 1 µm), where both surfaces present negative charges until EDL is overcome. Finally stronger irreversible forces arrived at attachment where its range begins from 5 to several hundreds of nanometers[17].
Figure 2.1: Attachment of bacteria cells mechanism by Dickinson[18]
Previous bacteria adhesion experience in different surfaces There are several surfaces that have been tested for bacteria adhesion measurements, as well as new antibacterial surfaces that prevent chemical leaching. As it can be seen in table 2.1.
2 Bac kground and State of Art
Test target Tested surface Bacteria Results Solution
catheter-associated urinary tract infection: CAUTI. 2 cm-long uri-nary catheters segments. Pseudomonas aeroginosa, Stapholoc-cus aureus, CNS,Enterococcus spp., Streptococ-cus spp. and Salmonella ty-phirium strains MAE52 and MAE50.
•179 bacterial isolates from catheters (96) and urine (83) samples.
•4 from 5 catheters surfaces detected coexistence of gram-negative and gram-positive bacteria.
•E. coli was the most infection agent in catheters.
The prevention of biofilm, hence preven-tion of infecpreven-tion.[19]
bloodstream infec-tions, fungal infection risks in intensive care unit, gastrointesti-nal tract surgery and power antibacterial agent treatment conse-quences. SETM: Silatos sil-icone sheet Candida albicans and non-albicans spp.
•More biofilm formation by non-albicans Candida species because its higher capacity •Although C. Albicans Can-dida, a more pathogenic specie, are the main cause of bloodstream infections.
Fluconazole is the main counter agent for both species but resistance to it has been displayed by both too.[20]
biofilm formation in catheters and bio-prosthetic devices. SE, Thermanox cell culture coverslip and polystyrene 96-well plate. Candida species. Antifungal: fluconazole, am-photericin B and caspofungin (Sigma-Aldrich).
•10 from 11 samples were seen to have highly formed biofilms.
•Amphotericin B is effective against biofilms.
•The choice of the ma-terial is critical for re-ducing biofilms •Amphotericin B toxic-ity against mammalian cells and kidney damage is a drawback.[21] characterize the Hfq (RNA) dependency on biofilm formation. polystyrene, glass and cholesterol gallstones. Salmonella en-terica serovar typhimurium.
Hfq dependency for regulating biofilm development.[22] —
antibacterial:”New ma-terial”, leakage free.
polyelectrolyte multilayer (PEM): PVAm and Anionic PAA (polymers); different pulps (fibers). gram-positive Bacillus subtilis, gram-negative Escherichia coli ATCC11775.
Satisfactory in terms of bacte-rial growth inhibition.
•Varied because of the dependence on polymer content and cationic charge of the fibers. •pH used was raised to physiologic values where low values im-prove antibacterial effects.[23]
2.2
Theory themes (DEP-EWOD)
In this part all physical theory is presented and explained. DEP is explained in two parts, what DEP force is and how it can be an important part in the manipulation of dielectric particles. EWOD is explained by its basic mechanism and applications. Both mechanisms were selected for their ability to repel particles from the surface. EP Electrophoresis is the motion of particles by the appliance of an uniform
Elec-tric field.
DEP Dielectrophoresis is the manipulation force of polarisable (uncharged) dielec-tric particles by the use of non-uniform elecdielec-tric field.
eDEP electrode-less DEP due to a dielectric constriction of the electric field on a narrow space.Fig 2.2a
TW DEP Travelling wave DEP produces forces that pushes particles in/off the direction of wave propagation.(depending on the polarity of imagi-nary factor of CM)Fig 2.2b
(a) eDEP (b) TWDEP
Figure 2.2: DEP variations
This last two are DEP variations that are considered in selected methods as a possible result in the tests purpose.
EW Electrowetting is the surface properties modification due to applied electric field.
EWOD electrowetting-on-dielectric is EW on top of dielectric-coated elec-trodes.
2.2.1 Dielectrophoresis Force
In this part Dielectric Force is explained, as well the Clausius-Mossotti (CM) factor. Before explain of Dielectrophoresis, electrophoresis must be first defined. Electrophoresis is the coulomb force created due to a electrostatic charged par-ticle that moves (if it is DC potential) or stay still (if it is AC potential).
Electrophoretic mobility is described by Eq.2.1 and considering a “Thick double layer” is described by Eq.2.2
2 Background and State of Art
where εm, εo, are medium and vacuum permittivity, ζ is the zeta potential and
η is the dynamic viscosity.
µe=
2 εrεoζ
3 η (2.2)
Considering this and restricting it to dielectric particles, the study is focused on how electric field gradient affect the particle charge.
In order to define Dielectrophoresis (DEP), DEP force must be detailed as it is the main physical concept that influence this phenomenon.
The main force exerted by a gradient electric field in a general manner is de-scribed by Eq.2.3 where ε∗ stands for complex permittivity.
The time-average force is:
h FDEPi = 2πr3εoεm<{ ε∗p− ε∗ m ε∗ p+ 2ε∗m }∇| ~Erms|2 (2.3)
where ε∗ is the complex permittivity(vacuum, medium or particle), r is radius of the particle, and ~Erms is the electric field.
But considering spherical or cylindrical particles the equation can be reduced to Eq.2.4. FDEP = πr2l 3 εoεm<{ ε∗p− ε∗ m ε∗ m }∇| ~E|2 (2.4) ε∗= ε + σ ω (2.5)
where l is the cylinder length, ε is the dielectric constant σ stands for conduc-tivity and ω is the frequency domain. Last but not least <{f } is the real part of Clausius-Mossotti(CM) factor {f}.
It can be concluded from above that the force magnitude depends strongly on the polarizability of the particle in respect of the medium. This motion of particles resulting in polarization forces by an gradient electric field was named dielectrophoresis by Herbert A. Pohl in 1950[24].
2.2.2 Dielectrophoresis in DC and AC Fields
This part shows how DEP works in DC and AC fields. Finally the two cases are compared for our purpose.
Figure 2.3: Movement of the particle under ~E
Non-uniform AC electric fields Aside from the charged particles (they presents little movements due to the exerted opposite forces produced by the AC field),depending on dielectric particles properties, it will follow the same be-haviour of non-uniform DC. As a consequence, positive DEP(<{CM }> 0) and negative DEP (<{CM }< 0) will work as trapping and rejecting particle method respectively.
From above cases, it can be seen that DEP force depends strongly on ~E gradient. Based on it, this information can be deduced:
• FDEP is proportional to particle volume and εm. Its direction is within ∇| ~E|
rather than ~E.
2 Background and State of Art
(a) Electric field on 5, 10 and 15 µm gaps
(b) Electric field gradient vs. gap distance
Figure 2.4: Interdigitated electrode distance effect on Electric field[1]
2.2.3 EWOD in DC and AC fields
In this part EWOD is similarly presented and compared not only by DC and AC fields but the difference in frequency values.
To this extent, DC voltage is applied to see how the contact angle change as part of the force created by the electric field and how that force overcome the surface tension (Gibbs free energy) while the potential increases between the electrodes and the medium (capacitance).
By applying AC voltage the result is the same but depending strongly on the frequency applied the contact angle changes and return to its original position (zero potential)in a dynamical speed.
The electrowetting principle reckon on the described Lippmann-Young equation Eq.2.6 cos θ = cos θo+ 1 2 C γLG V2 (2.6)
where θo is original contact angle(without electric field), θ is the one with the
electric field, C is capacitance per unit area(between electrodes and liquid), γLG
2.2.4 Theory limitations
In this section limitations are displayed and justified, including: the Debye layer, the ion diffusion on Electrical Double-Layer (EDL) as seen on Section2.1, REDOX on electrolysis and joule-heating as part of the inherently happening in the proposed microenvironment.
It is important to consider also the limitations presented in every experiment in regard of the theory presented.
• FDEP is proportional to square of voltage. (reversing BIAS does not reverse
force)
• Joule heating effect can be neglected as the resistance and low voltage can reduce the electrical current considerable in this stationary test.
• As long as the Debye layer is able to be present in ionic medium, it must be taken on count but can be neglected if the force is high enough compared to EDL.
• Other limitations not accounted for like taking as constant values for di-electric, permittivity, conductivity, viscosity, resistivity, and more inherent properties of the particle, medium and surface materials.
2.3
Modelling of biological cells
In this part Biological particles are modelled in a layered approximation, also is presented the effect of a hydrophobic dielectric for EWOD applications.
The bacteria that was used is Salmonella typhimurium, which is characterized for having cylindrical surface and a tail.
2.3.1 Biological cells on DEP
In this part it is illustrated the approximation modelling of bacteria cell as a three concentric layer particle and details of calculation of the complex permit-tivity needed to calculate the CM factor.
2 Background and State of Art
(a) DEP multi-shelled cell modelling
(b) Spheric approximation of multi-shelled particles by smeared-out sphere approach
Figure 2.5: DEP cell modelling[2]
From Fig 2.5a can be seen that having the conductivity and permittivity data the right frequency can be calculated.
2.3.2 Hydrophobic effect on bacteria adhesion
In this part is presented a study of how surface energy is directly translated into hydrophobicity, and how adhesion works on that.
Bacteria adhesion is measured in free energy of interaction (∆Giwi). And so
it is the hydrophobicity measurement, where if ∆Giwi> 0 it is hydrophilic, and if
∆Giwi< 0 it is hydrophobic.
Figure 2.6: Relation between the number of attached cells of Staphylococcus epider-midis ATCC 35984 (RP62A) and the degree of hydrophobicity (∆Giwi) of various
types of substrata,ref [25]
From the Figure2.6 it can be seen how linear is the relationship between both measurements, confirming by this way how hydrophobic bacteria interact easier with hydrophobic surfaces and same with the hydrophilic interactions.
2.4
Modelling of electrodes
In this part electrodes configuration is selected by its purpose and an electrical modelling is approximated according to the material electrical values.
About electrode-gasket methods, there are already some samples of Disposable Electrode Arrays built and commercialized like in Fig 2.7, which are able to interact between cell proteins, and make cell proliferation measurements[26].
Moreover, it can be found also in the market antimicrobial-impregnated central venous catheter which use two antiseptics(chlorhexidine and silver sulfadiazine) to avoid bacteria colonization, which has been used in transplants, septic patients (included burned patients)[27].
After seen the some cell manipulation electrodes in the market, electrodes the-ory and limitations presented for DEP-EWOD phenomenon, the next step is to select the best electrodes array for the Project goal.
2 Background and State of Art
Figure 2.7: Electrode-gasket sample for Cell-extracellular matrix (ECM) protein interactions, signal transduction assays,detection of invasion of endothelial cell layers by metastatic cells, barrier function measurements and cell proliferation applications[26]
The selection of the electrodes array, in the electrode pattern, is interdigitated beneficial to levitating the particles in DEP and making the particles flow over them without reaching the bottom.
By having an open EWOD design, with interdigitated electrodes, rejecting bac-teria can be possible.Fig 2.9
Figure 2.8: Classification of DEP devices according to the configuration of mi-croelectrodes: (A) parallel or interdigitated, (B) castellated, (C) oblique, (D) curved, (E) quadrupole, (F) microwell, (G) matrix, (H) extruded, (I, J) top-bottom patterned, (K) side-wall patterned, (L) insulator-based or electrodeless, and (M) contactless,[28]
(a) EW electric fields affecting bacteria
(b) Electric Modelling of the setup
Chapter 3
Design and assembly of the
experimental setup
This part describes the wafer fabrication which consist of the assembly of the electrodes array onto a glass surface using standard MEMS technology processes.
Several aspects of the design had to be considered in order to adapt the ex-perimental device to the measurement setup available; transparent substrate and thin device layer for observation of the bacteria through an inverted fluorescent microscope, oil based lenses, culture time, and Teflon layer deposition.
The design and assembly can be summarized to 3 stages: electrodes design, chip fabrication and gasket moulding. The electrodes design was started on L-Edit program, where the interdigitated electrodes arrangement resulted to beR
the most suitable for the DEP rejecting bacteria application as it is explained later on. The chip fabrication was done in the clean room in KTH Kista facilities because of the micro scale proportions of the electrodes design; metal sputtering, lithography, etching, dicing and finally Teflon spinning were done to build the final desired chip.
Finally, it is shown how gasket design and fabrication were made for containing bacteria on top of the produced electrodes.
3.1
Chip design and assembly
This part describes the chip manufacturing which includes electrodes design, metal sputtering, photo-lithography, etching and dicing needed to obtain the wafer with gold electrodes on top of it. Finally a Teflon layer is spun which makes possible to act as an capacitor for EWOD tests purposes.
3.1.1 Electrode design
The investigated electrode design of interdigitated electrodes was chosen be-cause it was the most suitable and flexible design for DEP and EWOD applications according to the Fig 2.8 [28] as it was shown in Section 2.4, because this kind of configuration are often found in Travelling wave DEP and DEP applications, where electric fields induce movement on floating particles. But also this configuration is used in pulsed DEP, which has a similarity with the pulsed behaviour of EWOD actuated with an on-off signal.
Besides the common EWOD configuration, a wire-free configuration[29] is more useful because a tip on top of the droplet is not longer needed, and match up with the selected DEP configuration mentioned in the paragraph before.
In this case a wire-free configuration[29] on top of a glass substrate of golden interdigitated electrodes is presented, in contrast to the conventional EWOD con-figurations as it can be seen in Figure.3.1a.
The L-Edit software was used to design the electrodes layout, which have been planned to have three different widths (a, b in Table 3.1)keeping the length of each electrode constant (l=6.450 mm) as seen in Fig 3.1b. The a and b variables were chosen within the bacteria scale size(<10 µm).
a µm b µm
Case 1 5 10
Case 2 10 20
Case 3 20 50
Table 3.1: Different width of the electrode depending on the case used
(a) EWOD conventional configuration vs wire-free configuration[29]
b a
6.450 mm
(b) Interdigitated electrodes design where the width ’a’ and ’b’ are variable and the length is kept constant.(see table3.1)
Figure 3.1: Electrodes configuration
3 Design and assembly of the experimental setup
plates. For that reason, the gasket should stand on top of each square regions (as seen on Fig 3.2b) where each square is an array of electrodes as seen in Fig 3.2a. From the 9 electrodes squares, 6 are active and the last 3 on the bottom are going to be used as negative control (no power supply is applied by not connecting to the power grid).
(a) Gasket on chip to see the electrodes length reason
6.450 mm
(b) Interdigitated electrodes spaced on a quarter of a 10 cm diameter wafer.
Figure 3.2: Electrodes in square spaces
One test area is composed of 9 electrodes squares. For the sake of fabricating as many tests areas (chips) as possible, 4 chips are inserted in one wafer are as it can be seen in Fig 3.3b. In addition, each chip has been numbered with case 1, 2 or 3 as used in Table 3.1 to differentiate each chip from each other, this is shown in Fig 3.3a.
(a) Metal mask design for wafer lithography
(b) mask designing on diced wafer(what is expected)
Figure 3.3: Mask and expectable wafer design
3.1.2 Chip fabrication
The following subsection describes the process and details involved in the manu-facturing of the wafer. From the layer structures to the Teflon spinning of the final finished wafer. All procedures of this fabrication where done on Kista Campus at KTH facilities.
The procedure is outlined below: 1. Glass wafer: 500 µm of thickness
2. Sputtering Metal: TiW (200 ˚A)under Au (1500 ˚A) layer 3. PhotoResist(PR): Type nLOF2070 is used
4. Development: three minutes of duration 5. Etching: First layer (Au)
6. Etching: Second layer (TiW) 7. Striping: Last layer (PR) Process
3 Design and assembly of the experimental setup
500 um 20 nm 150 nm
10 um
Figure 3.4: The material layers of the final wafer
Metal Deposition
The metal deposition was made in a metal sputtering machine (KDF 844NT,R
Model 844GT), using DC sputtering system. In order to improve the 150 nm gold layer adhesion on the glass wafer, 20 nm of titanium tungsten alloy (TiW) was sputtered.
Structuring of gold
Lithography In this part is detailed the photo-lithography method, where used photoresist and development times are included. Different development times under Ultraviolet light (UV) exposure were tried in order to get the best result.
To define an etch mask for the gold, a photoresist layer (AZnLOF2035 RER 600 [2:1]) is spun on the wafer, , in a OPTI spin machine model SST20(SSE) atR
3000 RPM for 30 s with a final soft bake at 90◦ C for 60 s.
Next, the UV exposure dose was optimized through some tests to reach the ideal exposure parameters (Table 3.2) on a Karl Suss , MA6/BA6KSM modelR
machine (Fig 3.3a).
Lamp intensity Lamp power Wavelength Mode CP
with 10 s of Lo Vac Contact 18,3 mWcm2 350 W 450 nm
and Alignment Gap of 40 µm
Lastly, to finish the lithography process the pattern was developed. After bak-ing on a hot plate to cross-link the photoresist, the wafer is developed for 4:05 min to remove the unexposed areas of photoresist film. Subsequently 1 minof post expo-sure bake is performed. This process is automatically executed by MAXIMUS ma-R
chine.
The photo-lithographic process is shown in Fig 3.5
(a) Wafer with sputtered metals (b) Wafer preparation for lithography
(c) Wafer exposure to UV
(d) Development for removing unexposed PR
(e) Etching of metal (f) Wafer finished after PR remove
Figure 3.5: Process of Photo-lithography
Etching Etching consists in the removal of unwanted metal layers from the glass surface, such as sputtered metal that does not belong to the electrode pattern.
In this part, each metal etching is detailed with their respective recipe. Having many different electrodes widths made this a complex task resulting in some over etching or poor etch finishing.
As the layer of gold is the first one from the top, it is a good idea to start with it.
Au ETCH FORMULA: • 60 g of I2
• 240 g KI • 2400 g H2O
3 Design and assembly of the experimental setup
Because our electrodes have different width, different etching times were used with the aim of etching more metal without affecting the electrode arranged width .
Au ETCHING METHOD:
For every glass wafer, the etch was performed for 30 s at 10 RPM. Additionally, extra time was tried according to the following Table 3.3 in manual mode(static).
#1 #2 & #3
1st wafer – –
2nd wafer 5 s –
3rd wafer 5 s 3 s
Table 3.3: Different used etching times on lithography In the same way the TiW layer is etched.
TiW ETCH FORMULA:
• 31% 50 H2O2 @50◦ C: etch rate 1.25 nm/s.
TiW ETCHING METHOD:
The etching liquid is poured manually onto the glass wafer during 20 s. This process is refined by feedback from observation under the laboratory microscope.
In this way, time can be measured and see how far the etching can go without over etching the electrodes. Both methods were used in the actual devices to etch different material layers, in Au etch method the best time was 5 seconds, and for the TiW etch method was 20 sand then steps of 5 to 8 s.
Dicing In this part it is explained how the glass wafers are diced and the param-eters used for the saw machine setup DISCOTM DAD 320.
The dicing of the wafer is more delicate in glass wafers. A different speed from the silicon wafer dicing should be used to avoid break as on Table 3.4).
Spindle speed Feed Speed
Glass 14’000 RPM 1
Silicon oxide 30’000 RPM 5-10
Table 3.4: Spindle and Feed speeds for the different materials
The parameters used in this process are shown in the tables Table 3.5 and Table 3.6.
P1A851 Work thickness (wafer thickness) Blade height Tape thickness (blue tape) Rnd(round work size) z-axis down speed
Machine Parameters 0,8 mm 0,0650 mm 0,05 mm 110 mm 10 mm/s
Table 3.5: Glass Blade parameters
ZH05-SD2000-N1-70 EE Work thickness (wafer thickness) Blade height Tape thickness (blue tape) Rnd(round work size) z-axis down speed
Machine Parameters 0,55 mm 0,1 mm 0,08 mm 120 mm 10 mm/s
Table 3.6: Silicon Oxide Blade parameters
(a) Electrode bus connection
(b) Contact pads
(c) Electrode # 1 (narrowest)
(d) Electrode # 2
(e) Electrode # 3 (widest)
Figure 3.6: Metal Mask Measurements
3 Design and assembly of the experimental setup
behaving as an insulator makes possible to create potential difference between the substrate and the used medium. Additionally, it makes easy to reuse the surface after bacteria attachment. In an ideal case, Teflon was thought to be covering only the top of the electrodes because the gasket could adhere easier to the spaces where the square electrodes were not located.
Teflon (C2F4)n is formed with the mix of 0.6 % Teflon AF-1600 (DuPontR TM)
and FC–40 in a dilution ratio of1/7.
Depending on the thickness of the layer, the spinning speed and acceleration are modified. The spin coater used was SPIN 150– NPP. In order to have 480 nm of thickness in our samples, a 1200 RPM speed is selected within 60 s at 100 RPM/s of acceleration.
Subsequently, the chip is heated up in a two-steps procedure on a hot plate at 165◦C for 10 min, and then 200◦C for 25 min. The hot plate used was IKA C– MAGR
H57. Teflon fumes may be dangerous and care must be taken to avoid inhalation. Thereafter, the samples were left to cool down for 20 min.
Cut pattern in film
1 2 Cutted by hand and plotter
Fit the sticker film to the wafer 3
Heat it up for Teflon curing 5
Press all over to reduce bubbles 4
Result in the wafer 6
Figure 3.7: Teflon technique to insulate on only electrodes
3 Design and assembly of the experimental setup
(a) Diced finished wafer and soldering macro electrodes
(b) Final built test device
Figure 3.8: Finished chip
Comments Some problems arose during the wafer fabrication process from de-sign to the etching part that are worth noticing.
The electrode array design was supposed to mimic the 96– wells plate that are used in biological laboratories (Fig 3.11a as it is going to be explained in Section3.2). In this case, we used only nine wells design because those were enough for our anti-fouling experiments.
During the etching, the major challenge was adapt the etching times defined for the smallest electrodes and the bigger ones.
For example, it was needed more time for the smallest ones (Case1) so over-etching were seen (red shrinking arrows in Fig 3.9); while for the bigger electrodes (Case 2 and 3), the required time was less and it resulted in under-etching of the smallest electrodes (green increasing arrows in Fig 3.9).
Figure 3.9: Etching time problem between Case 1 electrodes and Case 2 electrodes
(a) Broken electrode bus (b) Photoresist residue
(c) Photoresist and non-etched spots (d) Irregular patterns over electrodes
3 Design and assembly of the experimental setup
This problem was solved opening the short-circuited electrodes with a sharp diamond needle (same used for cutting the glass slides for the gasket moulding.
Next, the main concern in Teflon spinning was to achieve a thicker layer enough to act as a capacitor but thin enough to allow electric field to act on particles by EWOD. Several iterations of spinning time and curing temperature were done to achieve the final height (∼500 nm).
Some of the key problems were the way Teflon was spun on the spinning ma-chine. If there were any dust particle or bulk on top of the surface, it got affected making some sections covered by Teflon in a thinner level.
It has been found that using 2.25% of dielectric layer (Teflon) is to thin that breakdown becomes easy to happen. In fact, pin holes in the Teflon layer facilitate electrolysis to occur.
Finally the wafer is covered with Teflon and soldered to copper electrodes as it can be seen on Fig 3.8b
3.2
Gasket fabrication
In this section a gasket design is developed in order to contain the bacteria medium on top of our chip (manufactured diced wafer glass). Two materials are analysed to be used as a gasket for the bacterial medium container.
Gaskets were done to mimic the well (Micro-titre) plates that are used by the bacterial tests.
The chip was made based on the size of the wells from the Cell Culture Plate (96 Well) Sterilin , which well diameter φ is constant (φ= 6,45 mm), as it can beR
seen in Fig 3.11a.
(a) Commercial Cell Culture Plate (96 Well) Sterilin R
(b) Design of the gasket to be done
Figure 3.11: Commercial gasket and design
the sake of practicality which in fact had good adhesion and made-up with one big single well.
The main design limitations presented were the thickness (calculated for the medium quantity), size (covering all the electrodes, giving space for the macro elecrodes connections and enough for improve adhesion to prevent leaking), easy detachment(for fast measure of bacteria after incubation)and evaporation of the medium during the time without affecting the observation procedures.
3.2.1 OSTE Gasket
In this part, off-stoichiometry thiolene (OSTE) was selected mainly for the tuning abilities of the material and its versatility for complex moulding.
The first gasket was a squared shape with nine wells in it, where each well correspond to each square electrode in the chip. Because adhesion of the OSTE was not uniform all over glass surfaces and some wells leaked due to the used thickness, a second gasket was made with a single well for better tuning of the properties. In this case, two thicknesses were tested.
With the intention of calculate the exact amount of UV-curing time, gaskets were made by reducing the layer thickness and also making only one hole for prac-tical measurements: 5 mm, 2 mm and 1 mm respectively(Fig 3.12). The sealing tests were shown in Subsection 3.2.3.
Figure 3.12: OSTE nine hole and one hole gaskets designs
Formulation
In this section the design is sought to fit into the well plate structure but is modified for better clamping and sealing.
The only elements used for the formulation were two types of monomers, a thiol functional group and an allyl functional group. Combining the two components, an off-stoichiometric thiol-ene network polymer is fabricated.
The OSTE formulation is based on 100 % thiol excess as can be seen in Table 3.7.
Mixing procedure
3 Design and assembly of the experimental setup
Off stoichiometry Thiol Allyl TPO-L(initiator)
50% PETMA 1.125x TATATO 1x 0.5 wt%
PETMP 1.125x TATATO 1x 0.5 wt%
80% PETMA 1.35x TATATO 1x 0.5 wt%
PETMP 1.35x TATATO 1x 0.5 wt%
Table 3.7: OSTE Formulation: Stoichiometry, monomers and curing agent
end of the mixing, which can be solved by introducing the sample to vacuum for at least 30 min at a pressure of approximate -72 kPa.
Preparation – moulds
In this part aluminium and glass moulds were used for the gasket shaping as il-lustrated in Fig 3.13 and Fig 3.14 taking in consideration the size of the wells.
The aluminium mold consisted in two squared frames, one base and a glass for lithography. For assembling it only was required to screw the parts together.
In case single hole wells, aluminium mould were not needed as a result of not having big samples and trying only a little piece of polymer. Thus suitable glass slides were used for cleanliness and practicality.
In order to made a single hole, a very simple and quick mould should be used. By this, some glass slides were cut with a diamond burr type Flamme, from the Hobby Drill 2000 brand,as it can be seen in section 1 of Fig 3.14.
Once many 2 cm x 2 cm are cut, are glued with double side Scotch 3M tapeR
to a single glass slide which was previously covered with a Xerox film.(to detachR
easily after curing).
Clean mold and equip film mask
1 2 Assemble the mold, films, mask
Adjust the screws in it 3
UV curing 5
Input prepolymer in mold 4
Cured Polymer - Gasket 6
3 Design and assembly of the experimental setup
Cut and glue the glass slides
1 2 Glass slide with film covers
Prepare the cover slip mask 3
Clamping and UV curing 5
Input prepolymer and top mask 4
Cured Polymer - One Hole 6
Figure 3.14: OSTE one hole gasket on glass slide mould
Lithography
In order to cure the pre-polymer, UV curing using an OAI UV curing lamp,R
was used. Obviously, single hole wells used much less quantity of pre-polymer, consequently UV curing time decreased significantly (less than 10 seconds). In bigger gasket case, undoubtedly a nine patterned film mask was used as it can be seen in the bottom corner of the first sequence of the Fig 3.13.
Table 3.8 and Table 3.9 shows the different times used and the polymer finish characteristics.
One hole gasket test of 2 mm
Thiol monomer Off stoichiometry Time (s) Results
PETMA
50 %
26 too much curing time X
20 still over-cured, no hole X
15 half well X
10 defined well √
8 more sticky, neat √
6 very clear, uncured, affected by cover film X
80 %
27 too much curing time X
26 still overcooked X
15 almost defined hole X
10 hole, and more clear √
8 uncured, bottom too raw & soft, clearest X
PETMP
50 %
27 too much curing time X
20 over-cured, no hole X
15 over-cured X
10 still no hole, bottom uncured X
8 same as 10 & no adherence X
5 half well, blurred X
3 almost defined hole, blurred, detachable √
2 almost defined hole, clearer, softer √
1.5 uncured, very stiff, detachable X
80 %
27 too much curing time X
5 almost defined hole, bottom raw, not clear X
3 defined hole but stretched, clearer X
(very well defined wells) 2 defined hole & softer, breaks easily √
softer than 50 % 1.5 defined hole, easier to detach √
Table 3.8: Curing time variations for best OSTE 2 mm sample
Development
Finally, the development of the gaskets was possible by the use of solvents such as acetone or butyl acetate. Importantly, de-moulding from the aluminium and glass moulds were made by detaching the polymer from the Xerox films that lied onR
top and on bottom of the moulds and helped by the solvents actuation.
3.2.2 PDMS Gasket
3 Design and assembly of the experimental setup
One hole gasket test of 1 mm
Thiol monomer Off stoichiometry Time (s) Results
PETMA
50 %
15 too much curing time X
10 over-cured X
6 defined hole √
4 defined hole √
80 % 8 almost defined hole X
6 a bit uncured X
PETMP
50 %
6 too much curing time, no hole X
4 over-cured, no hole X
2 defined hole √
1 defined hole, bit uncured √
80 %
4 almost defined hole X
2 almost defined hole X
1 top face stuck to cover film X
0.5 too uncured, melted √
Table 3.9: Curing time variations for best OSTE 1 mm sample, curing agent for PETMP 80% was 0.1 g for the others 0.05 g
PDMS known as well as Poly(dimethylsiloxane), is an organic polymer widely used in microfabrication. It was chosen due to its simple formulation; of course the cost of it was to make a simpler moulding than the desired one, but for practical instances.
Contrary to OSTE, PDMS is not cured using UV, and the patterning must be defined in the mould. The manufacturing is even more simple because the lack of specific curing times.
Formulation
In this part is described the parameters used for the manufacturing of the PDMS gasket. The formulation used is shown in the Table 3.10.
base curing agent
10x 1x %
Table 3.10: PDMS Formulation: base and curing agent ratio
Mixing procedure
Elements mentioned in Table 3.10 are mixed. Both elements are stirred for at least 10 min until is fully mixed.
Cast of PDMS films
In this part the glass moulds are specified as a part of the curing step. After the mixing procedure the uncured PDMS is poured very carefully (to avoid any air bubble) into a glass mould which consist in a glass surface delimited with glass slides depending on the desired height. Carefully a Xerox film is extended all overR
the liquid PDMS. The prepared filled mould is then introduced in a 75◦C heated oven for 3 h. Once the PDMS is cured, the well could be cut out using a razor blade to give the adequate shape. Different samples can be seen in Fig 3.15
(a) Chip with gasket
(b) PDMS gasket in use with bacteria
(c) PDMS gaskets within different heights
Figure 3.15: Different made PDMS gaskets
3.2.3 Leakage tests
In this part, the properties of the gaskets are discussed how gaskets were resulted and the problems or successful outcomes from the manufacturing process.
This parameter can be reflected on the gasket pictures in Fig 3.16
3 Design and assembly of the experimental setup
(a) OSTE gaskets with different sealing tests
(b) OSTE one hole gaskets within dif-ferent curing times
(c) 2 cm PDMS gasket with sealing property on Teflon layer
(d) 1 mm PDMS gasket with hydropho-bic property without leaking
Figure 3.16: OSTE and PDMS gaskets on sealing tests
Comments In this stage, two gasket materials were tested. In the nine-hole OSTE gasket, the lithography problem was due to very thick layers, the base remained uncured creating a non-uniform surface due to the unknown exact UV curing time, while the top face was satisfactory. Even though, sealing tests were made on nine-holed gaskets which the best sealing was with PETMP 80% but not last for more than 1 h as it is shown in Fig 3.17
In the single hole OSTE gasket lithography parameters considered mainly: hole circular shape, polymer transparency and sealing, where the best ones were: 2 mm sealing test: PETMA 50% 10 s, and PETMP 80% 1.5 s.
1 mm sealing test: PETMP 50% 1 s (PETMA had not good hole shapes). There was not enough time to return to the nine-hole test because bacterial tests came out and a gasket was required for it. Nevertheless nine-hole tests shown not having the right curing time because on sealing tests liquid leaked from under the gasket.This is displayed in Fig 3.16a
In the PDMS gasket, sealing was quite good, and even with different gasket heights (5 mm, 1 mm) it was capable to retain the bacteria medium on top of the manufactured wafer, taking in consideration that the wafer’s surface was a thin layer of Teflon. This because of its rubber-texture of the PDMS material that stick easily to the surface, but with a little help of pressure to keep air bubbles out of the bottom surface of the gasket. It is important to mention that for very thin gasket films overflow is difficult to occur due to the hydrophobicity of the surface as is shown in Fig 3.16d.
3.3
Bacteria to be used
In this part, a brief description of the bacteria used and its morphology and topol-ogy.
Regarding the bacteria maturity, they cause infections by sporadic or epidemic mediums[5], these can be divided in several typologies where common ones are Gram-negative bacteria and gram-positive bacteria. Generally, the more resistant bacteria against antibacterial agents are the gram-negative because of their several cellular layers.
3.3.1 Salmonella Typhimurium
Salmonella species are reported as an often bacteria acquired in the community, as in Table 5 of the Infection Control Guide in Hospital Personnel[31].
3 Design and assembly of the experimental setup
(a) Salmonella close up
(b) Salmonella en-tanglements
(c) Biofilm forma-tion
Figure 3.18: Salmonella typhimurium
Measurements and results
This section consists of three parts. The first is a numerical simulation of the DEP force experienced on particles under the influence of spaced planar electrodes. Dif-ferent models for the bacteria are compared and optimal frequencies are determined for repulsion from the surface. Pro and cons are discussed for DEP techniques. The second part is an experimental study of EWOD to influence the bacteria adhesion of Salmonella Typhimurium. This study is carried out in cooperation with the group of Prof. Ute R¨omling from the Karolinska Institute. The third part is an experimental setup and several experiments for evaluating the bacteria attachment.
4.1
Numerical simulations of DEP on bacteria model
In this part, Clausius-Mossotti(CM) factor plots parameters are calculated for the later experimental bacteria. Similarly, simulations are done on how the gradient of electric field from two gold electrodes in a liquid solution (water) affects the modelled bacteria.
DEP simulations were studied using COMSOL simulation and complemented by MATLAB for calculations of effective CM values. The system was modelled as: two co-planar gold electrodes under a layer of Teflon with a water medium. MATLAB curves, illustrating the difference of the CM factor by changing the particles’ medium can be seen in Fig 4.2 (different medium conductivities displayed) as explained in the following subsection.
4.1.1 MATLAB calculations
4 Measurements and results
Three layer particle information
Particle Layers (outer to inner) Radius [µm] σ [ µS/cm] ε [F/m]
Yeast 1 (cell wall) 2.5 140 60 2 (cytoplasm) 2.36 0.0025 6 3 (nucleus) 2.35 2000 50 E. Coli 1 (cell wall) 1 500 60 2 (cytoplasm) 0.98 0.0005 10 3 (nucleus) 0.975 1000 60
Table 4.1: Different parameters to consider on three-layer particle(E. coli density:ρ = 1100 kg/m3 ± 3% )
By the smeared-out sphere approximation considered in Fig 2.5b formulas in Eq 4.1 and Eq 4.2 are solved for the bacteria analysis by MATLAB programming.
ε∗BACT = ε∗3 r3 r2 3 + 2ε∗21−ε∗3 ε∗21+2ε∗3 r 3 r2 3 −ε∗21−ε ∗ 3 ε∗ 21+2ε ∗ 3 (4.1) ε∗= ε + σ ω (4.2)
From above, εx is the permittivity of the layer x, thus x can be one layer(the third
one) or a combination of layers(the first two ones); then σ stands for conductivity of the medium and ω is the frequency domain of the applied ~E. Last but not least <{f } is the real part of Clausius-Mossotti(CM) factor {f}.
It can be compared the results of the calculated values plotted on excel in Fig 4.1b and the plot from Matlab in Fig 4.1c to see how the few values obtained in Excel was a good approximation but not as detailed as the ones obtained by Matlab complex calculations.
From the MATLAB calculations, using five different medium conductivities (2, 10, 17.78, 31.62, 38, 50.62, 100 and 380 µS/cm) where DI water has 2 µS/cm, it is shown how the CM-frequency curves were affected by the medium and how Yeast and E. Coli bacteria change its pDEP and nDEP areas (explained in subsection 2.2.2) as a consequence of it.
(a) <{CM } value from ap-plied frequency
(b) Excel CM plot from frequency values for general idea
(c) Matlab CM plot vs. frequency for detailed data
4 Measurements and results 102 103 104 105 106 107 108 -0.2 0 0.2 0.4 0.6 0.8 1 1.2 frequency (Hz) R e [K (w )]
Plot of Clausius-Mossoti factor(CM), sm=3.162278e-03 YEAST E. COLI 102 103 104 105 106 107 108 -0.1 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 frequency (Hz) R e [K (w )]
Plot of Clausius-Mossoti factor(CM), sm=1.778279e-03 YEAST E. COLI 102 103 104 105 106 107 108 -0.1 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 frequency (Hz) R e [K (w )]
Plot of Clausius-Mossoti factor(CM), sm=1.000000e-03 YEAST E. COLI 102 103 104 105 106 107 108 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1 frequency (Hz) R e [K (w )]
Plot of Clausius-Mossoti factor(CM), sm=5.623413e-03 YEAST E. COLI 102 103 104 105 106 107 108 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1 frequency (Hz) R e [K (w )]
Plot of Clausius-Mossoti factor(CM), sm=1.000000e-02 YEAST E. COLI
These values were selected from different mediums used in different scientific publications about DEP. From the variety of ranges obtained, yeast and bacteria can be observed how they are affected by the medium conductivity.
It can be seen from Fig 4.2 that the higher the conductivity of the medium, the smaller the area under the curve that is in the real part of the CM factor. Regarding the negative area under the curve, this means that if you want more repulsive forces it is convenient to have higher conductive values of the medium and also shows lower frequency values to work with.
In addition to the selected mediums, additional common mediums related to the healthcare environment were investigated.
Blood conductivity the conductivity of the blood varies depending on different blood composites: Hematocrit that reflects the red blood cells/blood volume ratio (the higher of it, the lower the conductivity); plasma is the major com-ponent which contains suspended ions (the higher of it, the higher the con-ductivity); electrolytes change the blood resistivity (the higher of them, the higher the conductivity); erythrocyte helps the oxygenation of the blood by diffusion, which orientation define the conductivity (at higher flow rates, the higher of it)[33]. The average value, by [34] experiments, is 6666.67 µS/cm. Urine conductivity the conductivity of the Urine is normally around 280 µS/cm
[35]
4 Measurements and results
(a) Medium DI water and NaCl
solu-tion (b) Ethyl alcohol and human blood
(c) Skin and Urine (d) Medium saliva and gastric juices
4.1.2 COMSOL analysis and results
In this section, a complete simulation about how electric field acts on two coplanar wave-guide geometry electrodes is described. Here DEP force is the main active force and the direct relation with the frequency is portrayed.
From the COMSOL simulations, there were two measurements of electric po-tential intensity, electric field intensity and DEP force on the medium, as presented on Fig 4.4, Fig 4.5 and Fig 4.6 respectively.
4 Measurements and results
(a) Frequency at 1 kHz
(b) Frequency at 1 MHz
4 Measurements and results
(a) Vo=1 V (b) Vo=2 V (c) Vo=5 V
(d) Vo=10 V (e) Vo=20 V
4 Measurements and results
(a) Yeast at 1 V (b) E. Coli at 1 V (c) E. Coli at 2 V
(d) E. Coli at 5 V (e) E. Coli at 10 V (f) E. Coli at 20 V
Bacteria rejection possibility Based on the presented electrical field simula-tions, it can be calculated if bacteria rejection is viable.
The bacteria simulated in the COMSOL simulation was Yeast at a frequency of 100 Hz,in order to start with the least visible value from the Fig 4.1c. The real part of Clausius-Mossotti factor used for this particle was -0.4687 considering the negative sign as the zone of the attraction for DEP as it can be seen in the same figure.
The CM factor was calculated using MATLAB, where medium was DI wa-ter (conductivity from COMSOL 38 mS/m) and the yeast was considered as a 3-layer permittivity particle, which resulted in complex permittivities 1.1746 × 10−8 − i2.5778 × 10−6 and 7.0834 × 10−10 − i6.0479 × 10−5 (F/m) for yeast and water respectively.
The possibility of bacteria rejection can be measured by the exerted DEP force from the electrodes, which should be larger enough to stop (and maybe reverse) the speed of the bacteria due to the usual Brownian motion. The start point to do this, is the Einstein– Smoluchowski relation from the Kinetic theory [37] which describes the Brownian motion on a medium.
The general equation is the Eq.4.3.
D = µ κB T (4.3)
where D is the diffusion constant, µ is the mobility, κB is the Boltzmann’s
constant and T is the absolute temperature.
It can be found diffusion bacteria values like: Pseudomonas aeruginosa (2.1 × 10−9 m2/s), Klebsiella pneumoniae (0.9×10−9m2/s)[38] and Escherichia coli (0.8× 10−12 m2/s)[39]. Considering room temperature (T = 300 K)and the Boltzmann’s constant (κB = 1.3807 × 10−23 m2. kg/s2. K) and replacing the diffusion values
on Eq.4.3 mobilities are obtained: Pseudomonas aeruginosa (5.1 × 1011 m/Ns), Klebsiella pneumoniae (2.2 × 1011 m/Ns) and Escherichia coli(1.9 × 108 m/Ns).
From the general equation is derived also mobility as particle’s drift velocity(vd)
to an applied force (F )[40], where it is demonstrated from Eq.C.4. µ = vd
F (4.4)
where the acting force is evaluated through the gradient of the squared electric field on the vertical axis(∇| ~Ey|2), due to see the effective force that is going to act
on the bacteria.
Velocity is going to be calculated by the multiplication of mobility and force deduced from Eq.4.4.
As mobility values are already obtained, it remains only FDEP values to be
calculated. The force FDEPy is obtained from ∇| ~Ey|2 values obtained from the
simulation.
4 Measurements and results
voltage is increased. The highest values of ∇| ~Ey|2 are on top of the electrodes going
into the diagonal direction while the lowest are very near the electrodes’ corner edges. This electric field intensity analysis can complement also the distance effect seen in Fig 2.4.
In the simulation is illustrated three different distances (as seen in Fig C.1) to see how FDEP changes with the variation of height. From the three distances, 1
and 3 µm were no enough force to cover as much vertical area as it happens with the last distance. For that reason, from the resulted mesh it can be considered 7 µm of distance (height) from the electrodes as an acting DEP force effect for rejecting purposes.
(a) DEP force in Cartesian plane, with selected distance as space of rejection, 7 µm
(b) DEP forces in distance x-axis (abscissa) at 7 µm above the electrodes
(c) DEP force (blue dotted line)and drift velocity (yellow Pseudomona, ma-genta Klebsiella and green Escherichia) in distance x-axis
4 Measurements and results
Using Eq.2.4 and COMSOL ∇| ~E|2 values, DEP force is calculated in Table 4.2 considering radius of the particle (bacteria) r = 0.5 µm, vacuum permittivity εo = 8.854 × 10−12, water medium εm = 80.1, and<{CM } = −0.4687 extracted
from MATLAB calculations.
Replacing values from Table 4.2 into Eq 4.4, and using the mobilities of each bacteria, the Drift velocity of the particle (bacteria) is illustrated in Table 4.3
It can be seen in Fig 4.8 that the Drift velocity increases exponentially with the applied voltage.
0.00E+00 2.00E-02 4.00E-02 6.00E-02 8.00E-02 1.00E-01 1.20E-01 1.40E-01 1.60E-01 0 5 10 15 20 25 30 35 (m /s) (Volts)
Bacteria average Drift speed
Pseudomonas aeruginosa Klebsiella pneumoniae Escherichia coli
Figure 4.8: Maximum bacteria drift rejection velocity based on applied voltage Comparing the velocity obtained with the Diffusion length, which defines how far the concentration is dispersed in a medium during a period of time. It is defined by Eq.4.5[41].
L = 2 √
D t (4.5)
Replacing the Diffusion values of the bacteria can be seen on Table 4.4 the common length that moves during one second.
