ap-5-11.dvi Acta Polytechnica Vol. 51 No. 5/2011 Boron Doped Nanocrystalline Diamond Films for Biosensing Applications V. Petrák, J. Krucký, M. Vlčková Abstract With the rise of antibiotic resistance of pathogenic bacteria there is an increased demand formonitoring the functionality of bacteria membranes, the disruption of which can be induced by peptide-lipid interactions. In this work we attempt to construct and disrupt supported lipid membranes (SLB) on boron doped nanocrystalline diamond (B-NCD). Elec- trochemical Impedance Spectroscopy (EIS) was used to study in situ changes related to lipid membrane formation and disruption bypeptide-induced interactions. The observed impedance changeswereminimal for oxidizedB-NCD samples, butwere still detectable in the low frequencypart of the spectra. The sensitivity for the detection ofmembrane formation and disruption was significantly higher for hydrogenated B-NCD surfaces. Data modeling indicates large changes in the electrical charge when an electrical double layer is formed at the B-NCD/SLB interface, governed by ion absorption. By contrast, for oxidized B-NCD surfaces, these changes are negligible indicating little or no change in the surface band bending profile. Keywords: biosensor, nanocrystalline diamond, electrochemical impedance spectroscopy. 1 Introduction The increase in antibiotic resistance of pathogenic bacteria strains has spurred the development of novel antibiotics. One promising solution to these prob- lems is the group of antibiotics based on antimi- crobial peptides which are an abundant and diverse group of molecules that are produced by many tis- sues and cell types in a variety of plant and animal species. Their amino acid composition, amphipathic- ity, cationic charge and size allow them to attach to membrane bilayers and disrupt the membrane by the formation of pores [1]. They do not target spe- cificmolecular receptorsonthemicrobial surface, but rather interact directly with microbial membranes, which they can rapidly permeabilize. The monitor- ing of specific peptide-lipid interactions in antibiotic peptides, which affect the functionality of bacterial membranes, can play an important role in the re- search of new antibiotics [2]. Supported lipid bilayers (SLBs) are investigated as model systems of biologicalmembranes. They are composedof a lipidbilayer adsorbedon the surface of a solid substrate. In past decades, lipid membranes ona solid substratehaveattractedconsiderable inter- est, from the point of view of both fundamental and applied science. These structures have been exten- sively used to study the structure and properties of native biologicalmembranes and for investigating bi- ological processes such as molecular recognition, en- zymatic catalysis, membrane fusion and cell adhe- sion [3]. In addition, several applications based on lipid membranes have been developed, including the design of biosensors. A well-established technique for the formation of SLBs is the Langmuir–Blodgett technique, which is carried out by pulling a hydrophilic substrate through a lipid monolayer and sequentially pushing it horizontally through another lipid monolayer [4]. A second commonly employed technique for forming SLBs is vesicle fusion, in which a supported bilayer is formed by the adsorption and fusion of vesicles from an aqueous suspension to the solid substrate surface [5]. Commonly used substrates for SLBs are mica, fused silica and glass. Other substrates such as sil- icon, SiO2, platinum and gold have also been re- ported. In the case of diamond, SLBs can be formed on an oxidized hydrophilic surface and also on a hy- drogenated hydrophobic surface [6]. Diamond ex- hibits several special properties, such as good bio- compatibility and a large electrochemical potential window. These properties make diamond particu- larly suitable for biosensing [7]. In this application, aborondopednanocrystalline diamond (B-NCD) film serves as a solid support for SLBs and an active electrode for electrochemical impedance spectroscopy (EIS)measurementofmelit- tin induced membrane disruption. EIS was success- fully used for detection of disruption bymelittin on a free-standing lipid bilayer as well as on SLBs on gold surfaces. Ang et al. were able to detect membrane disruption of SLBs caused by Maigainin II on opti- cally transparent diamond [6]. The EIS detection of 84 Acta Polytechnica Vol. 51 No. 5/2011 membrane disruption by antimicrobial peptide LL- 37 has also been demonstrated. This work focuses on the effects of surface termination on the detection abilities of B-NCD film. In the presentworkwe have constructed a simple sensor for detecting thedisruption ofSLBs formedon a semi-metallically boron doped NCD electrode that serves as aworking electrode. SLBs are disrupted by membrane active peptide melittin. We report the re- sults ofEISofmembranedisruptiononhydrogenated and oxidized surfaces, and discuss the influence of B- NCD surface termination on the sensitivity of the sensor. 2 Experimental Planar sensor electrodeswere preparedbymicrowave plasma-enhanced chemical vapor deposition (MW PE-CVD) from methane/hydrogen mixtures in an ASTeX reactor, as described in [8]. The substrates were 2 inch silicon wafers (thickness 550 μm, crys- talline orientation (100), p-type doped with boron and resistivity from 1 to 20 Ω·cm), which were diced into samples 10 mm by 10 mm after deposition. The diamond layers had a typical thickness of 150 nm with an average grain size of 50 nm, as determined byX-raydiffraction and atomic forcemicroscopy. To ensure good electrical conductivity of the diamond layer, the CVD deposition was performed with an admixture of trimethylboron to the CH4 gas with a concentration ratio of 200 ppm B/C. The B NCD samples served as theworking electrode in our home- made set-up that allows impedance read-out. Prior to the measurements, the diamond sam- pleswere either hydrogenated inH2 plasma (50Torr, 800◦C, power 4000 W, duration 15 min) or oxidized by UV-ozone for 30 minutes. For the evaluation, the resulting contact wetting-angles were 95◦ ± 2◦ for the hydrogenated diamond and 14◦ ± 3◦ for the oxidizeddiamond. TheB-NCDsampleswere cleaned in a mixture of H2SO4/KNO3 (2 : 1 wt%) heated to 250◦C for 10minutes. The samples were then rinsed in deionizedwater and dried under a streamof nitro- gen. Fig. 1: A schematic cross-section of the setup for EIS measurements The B-NCD samples were mounted on a copper backing contact, using an electrically conductive eu- tectic transfer tape. Rubber O-rings (Viton) with an inner diameter of 6 mm were pressed between the active electrode and the body of the sensor, form- ing a cell with a total inner volume of 160 μl. The cell was filled with 140 μl of 10 mM Hepes buffer solution. A gold counter electrode 500 μm in di- ameter was immersed in the solution. The working and counter electrode were connected to the 4194A Impedance/Gain-Phase Analyser (Hewlett Packard, USA) with shielded cables. After a stable signal was obtained at 25◦C, a 100 μM solution of DOPC:DOPS (1 : 4) liposomes with negative chargewas added. Themembranewas formed by the vesicle fusion method. Lipid mem- brane formation was completed within 30 minutes. Formembranedisruption, a2 μMactiveamphipathic α-helical peptide, melittin, was added. The impedance measurement, 10 mV AC poten- tial signal (U) was applied and the resultingAC cur- rent was measured (I). Each 15 seconds, a sweep of 50 frequencies ranging from 100 Hz to 1 MHz was done. 2.1 Equivalent circuit Theequivalent circuit used formodeling theEISdata was used for diamond to model the processes in the sensor in several other applications [9]. The equiva- lent circuit, shown in Figure 2, can be divided into three components. (1) The first component is the se- ries resistance RS. This comprises the solution and the electrode resistance between the gold and the B- NCD working electrode. (2) The second component is a parallel combination of resistance R1 and a con- stant phase element Q1, and corresponds to the dou- ble layer on the surface of the electrode. (3) The third component, which corresponds to the space- chargeregion in theB-NCD,also consistsof aparallel combination of resistance R2 and a constant phase element Q2. The data was fitted to the model in ZSimpWin (PrincetonApplied Research, USA). The quality of the fit is determined by the χ2 test. If the result is below 1 · 10−3 it is a good assumption that the model that was used is correct. Fig. 2: The equivalent circuit used for modeling divided into components. R represents resistance; Q is a constant phase element 85 Acta Polytechnica Vol. 51 No. 5/2011 3 Results The data series were fitted over the total measured frequency range from 100 Hz to 1 MHz. The resis- tance of the solutionwas calculated from 8 measure- ments, and was found to be 99 Ω ± 12 Ω. The con- stant phase elements in the equivalent circuit showed a value of n close to 1, suggesting the capacitance character of the circuit element. 3.1 SLB on oxidized B-NCD The lipidmembranewasmeasureddirectly during its formation and subsequently themelittin induced dis- ruption was measured by EIS. The modeling showed significant changes in the equivalent circuit related to the formation of the lipid membrane. The Nyquist plot in Figure 3 consists of a semicircle in the higher frequency part of the spectra, which represents the space charge region, and the second semicircle in the lower frequency corresponds to the interface capaci- tance. The change in the absolute value of the impedance upon formation of the membrane on the surfacewas low. However, the changewas detectable at low frequencies, as can be seen on the Nyquist plot in Figure 3. The absolute impedance value at a frequency of 255 Hz had risen from 1453 to 1540 Ω. Fig. 3: Nyquist plot showing the initial state prior to ad- dition of liposomes (�), state after membrane formation (�) and addition of melittin (�). Fits to the equivalent circuit are indicated with solid lines After membrane formation, the free liposomes in the solution were flushed with 1 mM Hepes buffer. The changes in impedancewereminimal in the entire frequency range and remained at a value of 1540 Ω for a frequency of 255 Hz. The same is true for the equivalent circuit values, which remained almost un- changed. After the addition of melittin, the difference in the high frequency part was minimal. However, a small changewas observed in the impedance spectra. The absolute value of the impedance at a frequency of 255Hz decreased from 1540 to 1461Ω. Themax- imumchange of the absolute impedance value during themeasurementwas only 5%during thedisruption. 3.2 SLB on an hydrogenated B-NCD Thefirst curve (�) inFigure 4 shows the initial state, when only Hepes buffer was present, and the second curve (◦) shows the result state after the addition of liposomes. An increase in the size of the semi- circle corresponding to the molecular bilayer on the B-NCDsurface canbe seen on theNyquist plot. The modeling showed that the main detectable change in the equivalent circuit was a decrease in resistance R1 from 55 to 28 kΩ, together with a change in the capacitance of the constant phase element Q1 from 279 nF to 141 nF. By contrast, the resistance value represented by the R2 capacitance of Q2 changed from 0.45 nF to 1.38 nF. The absolute value of the impedance at a frequency of 4.9 kHz rose from 4.4 to 7.6 kΩ, so the impedance rose by 175% of the value prior to addition. Fig. 4: Nyquist plot representing changes in impedance after the addition of liposomes (�), state aftermembrane formation (◦) and the addition of melittin (�). Fits to the equivalent circuit are indicated with solid lines. The change after the addition of liposomes into the solution and after membrane disruption is clearly visible Redundant free liposomes in the solution were subsequently flushed with 1 mM Hepes buffer. The flush of the liposomes did not result in any signifi- cant change in the impedance characteristic, and the absolute value of the impedance changed only from 7.6 to 7.5 kΩ. This represents a 2% change in the impedance at 4.9 kHz frequency. The membrane active peptide, melittin, was added after the system had stabilized. The absolute value of the impedance at a frequency 5 KHz de- creased from 7.5 to 5.2 kΩ. This represents 68% of 86 Acta Polytechnica Vol. 51 No. 5/2011 the impedance value before the addition of melittin. Data modeling using the equivalent circuit showed a change in the values of the R1 ‖ Q1 elements of the circuit. The values of elements R1, Q1 increased. The resistance of R1 increased from 28 to 41 kΩ and the capacitance of Q1 increased from 141 nF to 356nF.However, the resistance of R2 decreased from 7.3 to5.1kΩ, andthecapacitanceof Q2 changedfrom 1.35 nF to 0.83 nF. 3.3 Comparison of hydrogenated and oxidized surfaces The main difference in sensitivity can be attributed to the difference in hydrogenated and oxidized sur- faces. In the case of a hydrogenated surface, the band bending is upwards, and the addition of neg- ative charge SLB at the surface leads to increased band bending. By contrast, in the case of an oxi- dized surface the band bending is downwards, and the addition of SLB should reduce the surface band bending. However, the important fact is that we are working with B-NCD, in which free holes are present. When we add the negatively charged SLB on a hydrogenated surface and increase the negative charges at the surface, the holes fromB-NCD diffuse to the B-NCD surface, leading to a large change in the impedance of the system. On the other hand, when we work with an oxidized surface there are no free electrons in B-NCD and therefore the change in the surface band bending is limited. This is themain reason why the B-NCD sensor will work much more effectively with hydrogenated surfaces. 4 Conclusions An impedimetric characterization of membrane for- mation and disruption on a hydrogenated and oxi- dized B-NCD surface has been carried out. For a hydrogenated surface, significant changes have been observed in the properties of the B-NCD/SLB inter- face on interacting with the membrane active pep- tides. Data modeling indicated large changes in the electrical charge occurring at the diamond surface, and also the creation of an electric double layer at the B-NCD/SLB interface, which is governedby ion- absorption. By contrast, for an oxidizedB-NCD sur- face, these changes are negligible. This indicates that there are fewor no changes to the surfacebandbend- ing profile. Acknowledgement The research described in this paper was supervised byProf. PatrickWagner fromHasseltUniversity and Prof. Milos Nesladek from the Faculty of Biomedical Engineering, Czech Technical University in Prague. Financial support from the Academy of Sci- ences of the Czech Republic (grants KAN200100801 & KAN400480701), COST MP0901 — NanoTP, MSM6840770012 “Transdisciplinary Research in the Field ofBiomedicalEngineering II”. andCTU(grant No. CTU 10/811700) are gratefully acknowledged. The Erasmus student exchange programme is also gratefully acknowledged. References [1] Izadpanah, A., Gallo, R. L.: Antimicrobial Pep- tides, Journal of the American Academy of Der- matology, vol. 52, no. 3, 52, p. 381–390. [2] Castellana, E. T., Cremer, P. S.: Surface Science Reports, Solid supported lipid bilayers, vol. 61, no. 10, p. 429–444. [3] Dufrene, Y. F., Lee, G. U.: Advances in the characterization of supported lipid films with the atomic force microscope, Biochimica et Biophys- ica Acta – Biomembranes, vol. 1509, no. 1–2, p. 14–41. [4] Solletti, J. M., Botreau, M., Sommer, F., Brunat, W. L., Kasas, S., Duc, T. M., Ce- lio, M. R.: Elaboration and Characterization of Phospholipid Langmuir-Blodgett Films, Lang- muir, vol. 12, no. 22, p. 5379–5386. [5] Kalb, E., Frey, S., Tamm, L. K.: Formation of supported planar bilayers by fusion of vesicles to supported phospholipid monolayers, Biochimica et Biophysica Acta, vol. 1130, no. 2, p. 307–316. [6] Ang, P. K., Loh, K. P., Wohland, T., Nes- ladek, M., Van Hove, E.: Supported Lipid Bi- layer on Nanocrystalline Diamond Dual Opti- cal and Field-Effect Sensor for Membrane Dis- ruption, Advanced Functional Materials, vol. 19, p. 109–116. [7] Nebel, C. E., Rezek, B., Shin, D., Uetsua, H., Yang, N.: Diamond and biology, Journal of the Royal Society Interface, vol. 4, p. 439–446. [8] Williams, O., Nesladek, M. et al.: Growth, elec- tronic properties and applications of nanodia- mond, Diamond and Related Materials, vol. 17, no. 7–10, p. 1080–1088. [9] Vermeeren, V., Daenen, M., Grieten, L., et al.: Diamond-based DNA sensors: surface function- alization and read-out strategies, Physica Status Solidi A, vol. 260, no. 3, 520(2009). 87 Acta Polytechnica Vol. 51 No. 5/2011 About the authors Václav Petrák was born in Prague, Czech Repub- lic. He graduatedwith amaster degree from theFac- ulty of Biomedical Engineering of the Czech Techni- cal University in Prague in 2010. In 2008 he joined department of Functional Materials at the Institute of Physics of the Academy of Sciences of the Czech Republic. In 2010 he worked for 5 months at the Institute of Material Research in Hasselt (Belgium) duringhisErasmusstudent exchange. He is currently working onhis doctoral thesis. His professional inter- ests are nanocrystalline diamond, nanodiamond par- ticles, biosensors, and also neural circuits and data analysis. His personal interests are family, traveling and photography. Jaroslav Krucký was born in Prague, Czech Re- public. He graduated with a bachelor degree from the Faculty of Biomedical Engineering of the Czech Technical University in Prague in 2009. In 2009 he joined the department of the FunctionalMaterials at the Institute ofPhysics ofAcademyofSciences. He is currentlyworking onmethods for preparingND thin films with selective growth. In future, he is planning to work on diamond micro-electron array electrodes for in vivo and in vitro neural measurements. Marie Vlčková graduated with a bachelor degree from the Faculty of Biomedical Engineering of the Czech Technical University in Prague in 2010. She is continuing her studies at FBMI, and is currently working on optimization of the surface pretreatment of substrates for diamond CVD deposition. Her pro- fessional interests are biomedical applications of dia- mond films. Václav Petrák E-mail: vaclav.petrak@fbmi.cvut.cz JaroslavKrucký Marie Vlčková Department of Biomedical Technology Faculty of Biomedical Engineering Czech technical University, Kladno, Czech Republic 88