Bioreactive Self-Assembled Monolayers on Hydrogen-Passivated Si(111) as a New Class of Atomically Flat Substrates for Biological Scanning Probe Microscopy

Bioreactive Self-Assembled Monolayers on Hydrogen-Passivated Si(111) as a New Class of Atomically Flat Substrates for Biological Scanning Probe Microscopy
of 13
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
  Bioreactive Self-Assembled Monolayers on Hydrogen-Passivated Si(111)as a New Class ofAtomically Flat Substratesfor Biological Scanning Probe Microscopy Peter Wagner, 1 Steffen Nock, and James A. Spudich  Department of Biochemistry, Stanford University Medical Center, Stanford, California 94305-5307  Wayne D. Volkmuth and Steve Chu  Department of Physics, Stanford University, Stanford, California 94305-4060 andRonald L. Cicero, Christopher P. Wade, Matthew R. Linford, and Christopher E. D. Chidsey  Department of Chemistry, Stanford University, Stanford, California 94305-5080 Received February 27, 1997, and in revised formApril 17, 1997 This is the first report of bioreactive self-as-sembled monolayers, covalently bound to atomi-cally flat silicon surfaces and capable of binding biomolecules for investigation by scanning probemicroscopy and other surface-related assays andsensing devices. These monolayers are stable undera wide range of conditions and allow tailor-madefunctionalization for many purposes. We describethe substrate preparation and present an STM andSFMcharacterization,partlyperformedwithmulti-walled carbon nanotubes as tapping-mode super-tips. Furthermore, we present two strategies of introducing   in situ  reactive headgroup functional-ities. One method entails a free radical chlorosulfo-nation process with subsequent sulfonamide forma-tion. A second method employs singlet carbene-mediated hydrogen–carbon insertion of aheterobifunctional,amino-reactivetrifluoromethyl-diazirinyl crosslinker. We believe that this new sub-strate is advantageous to others, because it (i) isatomicallyflatoverlargeareasandcanbepreparedin a few hours with standard equipment, (ii) isstableundermostconditions,(iii)canbemodifiedtoadjust a certain degree of reactivity and hydropho-bicity,whichallowsphysicaladsorptionorcovalentcrosslinking of the biological specimen, (iv) buildsthe bridge between semiconductor microfabrica-tion and organic/biological molecular systems, and(v) is accessible to nanopatterning and applicationsrequiringconductivesubstrates.  r  1997Academic Press INTRODUCTION The capability of the scanning force microscope(SFM) to image single biomolecules at (sub)nano-meter resolution in aqueous environments and withexcellent signal-to-noise ratio has redefined the con-cept of microscopy in biology. Since its invention adecade ago (Binnig  et al.,  1986), scanning forcemicroscopy has developed into a standard analysistechnique in materials and surface science (for re-cent reviews: Hamers, 1996; Bottomley  et al.,  1996).Due to better-defined sample preparation and fur-ther improvements in instrumentation, such assharper tips and the implementation of new modes(e.g., tapping and noncontact), SFM has been usedsuccessfully to investigate complex biological sys-tems (for reviews: Engel, 1991; Hansma and Hoh,1994; Bustamante and Rivetti, 1996; Shao  et al., 1996).Furthermore,theSFMhasbeenusedforvariousnonimagingapplicationssuchasnanostructurefabrica-tion (Park  et al.,  1995) and force interaction measure-mentsofdynamicevents(Florin  etal., 1995;Dammer  etal.,  1995). However, SFM is not yet a routine tool,especiallyforimagingofindividuallyimmobilizedglobu-larproteins,whichisstillamajorchallenge.Two factors are of special importance for high-resolution imaging of single biomolecules in their‘‘native’’stateandalsofornonimagingforcemeasure- 1 To whom correspondence should be addressed. Fax: (415)-7256044. E-mail: JOURNAL OF STRUCTURAL BIOLOGY   119,  189–201 (1997)  ARTICLE NO.  SB973881189  1047-8477/97 $25.00Copyright  r  1997 byAcademic Press All rights of reproduction in any form reserved.  ments: (i) the chemistry, shape, and finite size of thetip and (ii) the topography and chemistry of thesubstrate used for immobilizing the target mol-ecules. The molecules must be tethered stronglyenough that they do not move under the action of thescanning tip. On the other hand this should notcause them to denature or hinder their interactionwith ligands or other biomacromolecules at thesolid–liquidinterface.Thenumberofavailableatomi-cally flat substrates fulfilling all these requirementsis very low to date and each has its own strengthsand weaknesses.Mica is unsurpassed regarding flatness and re-quired time for sample preparation (conveniencefactor). This surface has been successfully used tosimply adsorb DNA(Guthold  et al.,  1994; Hansma  etal.,  1995) and proteins (Yang  et al.,  1993, 1994;Schabert  et al.,  1995; Mueller  et al.,  1995, 1996).Unfortunately, it is very inert to specific chemicalmodification and not very suitable for higher sophis-ticated molecular architectures, e.g., to achieve acontrolled specific orientation of the biomolecule onthe surface.Instead of attaching the biomolecules directly toan inorganic substrate, ultrathin self-assembledmonolayers (SAMs) chemisorbed to inorganic sur-faces have been used to create organic interfaceswith tailor-made reactivity, structure, and proper-ties that can be controlled by choosing the appropri-ate functional headgroups in  v -position (Ulman,1996). Biomolecules can then be covalently or nonco-valently attached at specific sites to the reactivegroups of these monolayers and investigated bySPM. The two main groups of SAMs are alkylsilox-ane monolayers (‘‘silanes’’) on hydroxylated surfacesand alkylthiol/dialkyldisulfide monolayers on noblemetals, preferablyAu(111). Alkylsiloxane monolayers, obtained on glass or onthe oxide surface of silicon, exhibit higher stabilitythan alkanethiol monolayers. However, they have alimited range of functionalities in v -position that iscompatiblewiththereactivetrichloro-andtrialkoxy-silane moieties. In most cases they are more disor-dered and their formation is less reproducible. De-spite these obstacles, organosilane films have beenused successfully as substrates for biological SPM(Karrasch  etal., 1993,1994;Lyubchenko  etal., 1993; Vinckier  et al.,  1995).The power of mica to direct heteroepitaxial growthof thin metal films has led to polycrystalline goldsurfaces flat enough to be used as SPM substrates. Atomically flat terraces of these surfaces have beenpreparedusinganannealingprocessofthedepositedgold film as described, for example, by Chidsey  et al. (1988). The flatness of Au(111) surfaces could befurther improved using a lift-off technique with micaasreplicastructuretoobtainmeanroughnessvaluesof a few angstroms over very large areas (Wagner  etal.,  1995). Alkanethiol SAMs can be easily preparedon these gold surfaces from the liquid or vapor phasewith very high reproducibility and a wide variety of functional groups that can be displayed on thesurface(DuboisandNuzzo,1992;MrksichandWhite-sides, 1996). Such  v -functionalized SAMs made itpossible to image immobilized aminomodified DNA (Hegner  et al.,  1996), to determine the interactionforces of single proteoglycan molecules (Dammer  etal., 1995)andantigen/antibodybinding(Dammer  etal., 1996), and to obtain the first images of native clathrincagesandtheir insitu disassembly(Wagner  etal., 1994).Here we present a third possibility for preparingan atomically flat monolayer as a substrate forbiological SPM that combines, as we think, theadvantages of the SAMs mentioned above: Reactive,denselypackedmonolayerschemisorbedonhydroge-nated silicon [H-Si(111)] surfaces. They can be pre-pared by removing the oxide layer of Si(111) surfacesusing aqueous solutions of NH 4 F. This results inatomically flat, hydrophobic Si(111) (1 3 1) surfaces,where the dangling bonds are saturated by a mono-layer of hydrogen, i.e., each silicon surface atom isterminatedbyasinglesilicon–hydrogenbondperpen-dicular to the surface (Higashi  et al.,  1990, 1991).These hydrogen-terminated surfaces are importantfor semiconductor processing as passivation layersandduetotheirkineticstability,theycanbemanipu-lated in air. Their topography consists of atomically flatterraceswithexcellentflatnessoverlargeareas.Unfortunately, they have not yet been consideredassubstratesforbiologicalSPMtechniques,presum-ably because they are too hydrophobic and did notallow well-defined covalent crosslinking. The latterobstacle has recently been overcome with the prepa-ration of the first covalently bound methyl-termi-nated monolayers on H-Si(111) surfaces (Linford  etal.,  1995, 1997a,b; Chidsey and Linford, 1995). Thischemistry directs replacement of surface hydrogenswith alkyl chains in a radical process. The resultingmonolayers are similar to SAMs of alkanethiols on Au(111), but due to their silicon–carbon bond muchmore stable. Here, we suggest two different routes tomake these monolayers reactive toward biomol-eculesusing insitu activationchemistryandpresenta simple and reproducible method of preparing thesemonolayers for SPM purposes. MATERIALS AND METHODS  Materials  All chemicals were of the highest available purity. Ultrapurewater with a resistance of 18 M V · cm was generally used for allaqueous buffers (purified by passage through a Milli-Q purifica-tion system (Millipore)). 4 8 -[3-Trifluoromethyl-3H-diazirin-3-yl]-benzoic acid  N  -hydroxysuccinimide ester (TDBA-OSu) was from 190  WAGNER ET AL.  Photoprobes (Knonau, Switzerland). The SO 2  /Cl 2  gas mixtureused for the chlorosulfonation was prepared from two tanks withanhydrous SO 2  and Cl 2  (99.999%, Matheson) and gravimetricallyadjusted to 0.8% Cl 2  in SO 2  in a self-built stainless-steel tube.1-Octadecene (97%) and ethylenediamine were purchased fromFluka (Switzerland). N-type standard Si(111) wafers were fromSiltec (Salem, OR), with a resistance of 0.1–0.9  V · cm and amiscut of  DQ5 0.2°.  Monolayer Formation Octadecyl monolayers on silicon were prepared as follows: A Si(111) wafer was cut into 2  3  5 cm pieces and cleaned in asolution of H 2 O 2  (30%) and conc. H 2 SO 4  in a 1:3 ratio for 10 min at90°C.  Caution: Acidic solutions of concentrated hydrogen peroxideare very dangerous and can detonate in contact with organicmaterials.  After rinsing with ultrapure water, the wafer pieceswere etched in oxygen-free 40% NH 4 F for 10 min at roomtemperature, transferred without further water rinsing to aquartz cuvette, and attached to a double manifold system con-nected to a diffusion pump with a liquid nitrogen-cooled trap. Thesamples were repeatedly flushed with argon and ultimatelyevacuated to  # 6  3  10 2 5 Torr. Deoxygenated 1-octadecene wasfreshlyvacuum-distilledontotheH-Si(111)surfaceandirradiatedat 235.7 nm for 2 hr using a Hg pen lamp (9 mW/cm 2 at 2 cmdistance; Jelight, Model 823-3309-2). After rinsing with CH 2 Cl 2 and sonication in CH 2 Cl 2  for 5 min, the wafer pieces were washedwith nanopure water, blown dry under a stream of argon, andeither subjected to functionalization or stored under argon.  Functionalization of Monolayers Route A.  Silicon wafer pieces (5 3 5 mm size) with octadecylmonolayers were placed in a 10-mm quartz cuvette with thepolished side pointing upward. Then 200 µl of a 15 m  M   solution of TDBA-OSu in dry carbon tetrachloride was added and immedi-ately illuminated with a broad-band 350-nm lamp (Spectronics,MB-100)atadistanceof4cmfor15min.Afterrinsingwithcarbontetrachloride,CH 2 Cl 2 ,andwater,theaminoreactivesamplesweresubjected to bioconjugation.  Route B.  Silicon wafer pieces with octadecyl monolayers wereplaced in a custom quartz cuvette and connected to the vacuumline described above. After repeatedly flushing the system withdry argon and evacuating to a pressure of less than 6 3 10 2 5 Torr,a gas mixture of 0.8% Cl 2  in SO 2  was then dosed onto the samplefrom a self-built tank (also connected to the vacuum line). Thefinalpressurewasadjustedto2Torr. Caution:Chlorineandsulfurdioxide are toxic and extreme care must be taken when handlingthese gases.  The samples were then immediately illuminated at351 nm (5 mW/cm 2 intensity at 2 cm distance; Jelight, Model84-2011-2, 41 nm bandwidth) for 10 sec. The cuvette was evacu-ated, flushed three times with argon, and left under argon. A 100 m  M   solution of ethylenediamine in anhydrous dimethylsulfoxide (distilled from CaH 2  and stored over molecular sieves)was transferred into the cuvette using dry transfer needles.After10 min the sample was rinsed first with DMSO and then withultrapure water and finally blown dry under a stream of argon toget a well-defined and stable amino-terminated surface.After thewafer was cut into smaller pieces the samples were stored underargon or subjected to further modification.  Immobilization of a dsDNA Fragment Using Route A  Amino-substituted oligodeoxyribonucleotides (Operon, Ala-meda) were used as primers in PCR. The vector pET28a( 1 )(Novagen, Madison) including a gene (712 nucleotides) coding forthegreenfluorescentproteinfrom  Aequoriavictoria wasusedasatemplate. For the PCR reaction, the plasmid was linearized by Sma I and amplified in 30 cycles. The resulting 1752-bp DNA fragment was extracted with phenol according to Sambrook  et al. (1989) and precipitated with ethanol.After separation of the DNA on a 0.8% agarose gel, the band corresponding to the 1752-bpfragment was eluted using the Qiaex II kit from Qiagen (Hilden,Germany). To remove traces of agarose, the DNA was extractedagain with phenol and precipitated. Finally, the DNA was dis-solved in water to a concentration of 0.1 mg/ml and stored at 4°C.Coupling of the amino-terminated DNA fragment to the  N  -hydroxysuccinimidyl-activated monolayer was routinely per-formed in water at a concentration of 1 µg/ml for 1 hr at roomtemperature. After washing with water, the sample was driedwith filtered (0.2 µm) compressed air of high purity and immedi-ately used for SFM imaging.  Ellipsometry Measurements Monolayer thickness measurement was done on a Gaertner Variable Angle Ellipsometer L116B using a helium–neon laserand an incident angle of 70°. An index of refraction of 1.46 wasused for the octadecyl film. Scanning Force Microscopy SFM was carried out on a Multimode NanoScope III fromDigital Instruments Inc. (Santa Barbara, CA) equipped with anE-scanner with a maximum scan range of 15 µm. Microfabricatedmonocrystalline silicon tips (Nanoprobes, Santa Barbara, CA)were used with nominal force constants of 75 N/m for tappingmode using conventional silicon tips. Multiwalled carbon nano-tubes (MWNTs) were mounted on silicon cantilevers with forceconstants of 10 N/m as described elsewhere (Dai  et al.,  1996). Thetapping amplitude was chosen to be 25 nm and the setpoint was24.5 nm. Typical scan rates were 1–3 Hz.All images are based onunfiltered data except image leveling. Scanning Tunneling Microscopy Scanning tunneling microscopy was carried out on a home-builtinstrument with an inverted sample holder equipped with aBurleigh Inchworm (trademark) piezoelectric motor, a Teflon cellmounted in a 6-in., six-way stainless steel cube that can beevacuated or filled with inert gas and a controller from DigitalInstrumentsInc.FordetailsseeWade  etal. (1997).STMtipswereelectrochemically etched from 0.25-mm Pt/Ir wire.  X-Ray Photoelectron Spectroscopy (XPS)  X-ray photoelectron spectroscopy spectra were obtained on aSurface Science Model 150 XPS spectrometer with an Al K  a source (1486 eV), quartz monochromator, hemispherical analyzer,and multichannel detector. The spectra were accumulated at atake-off angle of 35° and an angular acceptance of 30°, with a 250 3  1000 µm 2 spot size at a pressure of less than 1  3  10 2 8 Torr.Peaks were analyzed after a Shirley background subtraction andthe correction for the number of scans, the atomic sensitivityfactors, and the differential photoelectron attenuations. RESULTS Octadecyl Monolayer on Silicon  Asummary of the processing steps involved in theformation of atomically flat monolayers on Si(111) isshown in Fig. 1. It is well established that wetchemical treatment of oxidized Si(111) surfaces byaqueous NH 4 F removes the oxide layer and resultsin the termination of all surface silicon danglingbonds by hydrogen (Higashi  et al.,  1990). TheH-Si(111) surface shows a single sharp peak in thep-polarized infrared spectrum, but not in the191 BIOREACTIVE MONOLAYERS ON HYDROGENATED Si(111)  s-polarized spectrum, which is direct evidence thatthe surface consists of silicon–hydrogen bonds ori-ented perpendicular to the surface (data not shown).UV-catalyzed free radical addition of neat 1-octa-decene forms a densely packed monolayer covalentlybound to the substrate via stable Si-C bonds. Theresulting monolayer has a thickness of 19 Å, asdetermined by optical ellipsometry. This is consis-tent with the theoretical length of a trans-extendedC18-alkyl chain with a tilt angle of 35° to the surfacenormal. The surface is very hydrophobic, showingthe typical advancing contact angle of 113° in waterfor methyl-terminated monolayers, and the expectedsymmetric and asymmetric methylene stretches inthe infrared spectrum, as well as the asymmetricin-plane methyl mode (data not shown). X-ray photoelectron (XPS) survey spectra of thesestepsareshowninFigs.2a–2ctodisplaytheelemen-tal composition of the surfaces. The survey scan of aclean superficially oxidized silicon surface in Fig. 2ashows the oxygen 1s (O1s) and silicon 2s (Si2s) and2p (Si2p) signals, but also a small carbon 1s (C1s)signal due to hydrocarbon contaminations duringsample transfer in air. The survey scan of theH-Si(111) surface (shown in Fig. 2b) is completelyfree of C1s and O1s peaks as expected. The completedisappearance of the O1s peak was only achieved inthose cases where the surface was not washed withwater after the NH 4 F etching step (see also SFMimages in Fig. 3 and Wade  et al.,  1997). Si2p narrowscans of the fluoride-ion-etched silicon hydride sur-face also indicate no oxide present (not shown).Exposure to degassed 1-octadecene results in a highC1s peak due to the formation of the monolayer (Fig.2c). Detailed XPS analyses of the unmodified andmodified octadecyl monolayers will be publishedelsewhere (Wagner  et al.,  1997).Figures 3a–3d show STM (a, b) and SFM (c, d)imagesoftheH-Si(111)surface.InFig.3aaconstant-currentSTMimage(  z range10nm)isshownwithan  x ,  y  range of 4 3 4 µm 2 . The surface is extremely flat,exhibiting a mean roughness value of 0.194 nm.Characteristic structural features are the atomicallyflat terraces, thousands of angstroms across andseparated with step heights of 0.31 nm (one bilayerof silicon atoms). The surface appears defect-free incontrast to the oxide, which often exhibits an amor-phous rolling-hill structure, or the conventional,annealed gold surfaces having atomically flat ter-races but separated by deep groves and holes.Theshapeoftheterracesisrelatedtothemiscutof the single crystal wafer and to the NH 4 F etchanisotropy that causes surface removal by a flow of monohydride steps. The width of the terraces de-pends partly on this miscut (here  DQ5 0.2°) andcould be increased up to micrometer size by justchoosing special prime quality wafer material with DQ, 0.05° (see also Discussion). Figure 3b shows a500 3 500 nm 2 scan of the same sample with 2 nm  z range. Tapping-mode SFM images (taken in air) of adifferent H-Si(111) surface, previously rinsed withwater after the fluoride etch, are shown in Figs. 3cand 3d. In general, the shape and size of the terracesand the resolution of the images are similar to theSTM images in Figs. 3a and 3b. The 1000  3  1000nm 2 scan in Fig. 3c shows two additional surfacefeatures compared to Figs. 3a and 3b: Small islands,a few nanometers in diameter, randomly distributedover the whole surface and larger islands preferablylocated at steps. The first are attributed to siliconoxide and most likely stem from the water rinse,since they are not seen in the STM image. This canbe better seen in the 500 3 500 nm 2 scan in Fig. 3d.Surface oxidation related to the water contamina-tion has been previously investigated by high-resolution electron energy-loss spectroscopy(HREELS, Graef   et al.,  1989) and STM (Wade  et al., 1997). This defect structure can be completely re- F IG . 1.  Monolayer formation on silicon. Removal of the oxidelayer from Si(111) by aqueous NH 4 F etch solution results inatomically flat hydrogen-terminated Si(111) surfaces. Furthertreatment with freshly distilled oxygen-free 1-octadecene for 2 hrunder illumination with ultraviolet light leads to a closely packedmonolayer of 19 Å thickness. 192  WAGNER ET AL.  moved by omitting the final wash step, as shown in thesurfacesofFigs.3aand3b.Thus,itisimportanttonotethat the differences in surface topography in Figs. 3aand 3b and 3c and 3d are not related to the probingtechnique.Theresolutionofbothsetsissimilar,whereasthe SFM in tapping mode was able to image theoxide spots with less noise than the STM (not shown).The second important structural feature in Figs. 3cand 3d are bigger islands approximately 10–15 nm indiameter. We assume that these spots are trace metalcontaminations (not detectable by XPS) being adsorbedduring the etch process and resulting in triangulatedshaped steps due to a pinning effect. These defects arethe subject of future investigations and can probablybeeliminatedbyafurtherimprovementintheprotocol.Figures 4a–4c show SFM tapping-mode images of the methyl-terminated octadecyl monolayer usingMWNT. The structural topography of the monolayerrepeats that of the underlying silicon surface in formof shape and size of terraces which are separated by0.31-nm-highsteps.Themonolayerisextremelyflat,with a mean roughness of   R a  5  0.159 nm for the2000  3  2000 nm 2 scan in Fig. 4a, i.e., the UV light-catalyzed radical addition of alkenes does not deterio-rate the flatness of the H-Si(111) surface (see alsothecrosssectioninFig.4e).Thesurfacefeaturesseeninthe 500  3  500 nm 2 scan in Fig. 4b are smaller andhigher in density than the oxide islands describedabove.They,aswellasafewdepressions,arethesubjectoffurtherinvestigations,butareprobablyjustreflectingthe high resolution of the MWNT tapping tip onthe organic interface. This can also be seen in Fig. F IG . 2.  XPS analysis (survey scans) of different surfaces used in this study. (a) Silicon oxide (starting material) after treatment withH 2 O 2  /H 2 SO 4 . (b) H-Si(111) surface after removal of the oxide film. (c) Octadecyl monolayer. (d) Octadecyl monolayer activated by SO 2  /Cl 2 treatment with subsequent reaction with ethylenediamine (e) Octadecyl monolayer after photoreaction with TDBA-OSu. Note: Because of the low cross section of nitrogen and sulfur, these peaks are weak and should only be quantified using narrow scans (not shown). 193 BIOREACTIVE MONOLAYERS ON HYDROGENATED Si(111)
Similar documents
View more...
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks