<<

2228 copyrighto 1ee4by the A*""i"."%"fl;;li:"i ffJI"iSXTYJ*;13it;J0pu"-i",ionorthe copvright owner. Self-Assembled Monolayers Containing oMercaptoalkylboronic Acids Adsorbed onto Form a Highly Cross-Link€d, Thermally Stable Borate Glass Sur{acer

Robert I. Carey, John P. Folkers, and GeorgeM. Whitesides*

Department of Ch.emistry, Haruard (Jniuersity, Cambridge, Massachusetts 02138

ReceiuedAugust 5, 1993. In Final Form: April 8, 1994@

This paper describesthe preparation and propertiesof self-assembledmonolayers (SAMs) obtained by the ads'orptio.rof 11-mercaptoundecanyl-1-6oronicacid, HS(CHz)uB(OH)2,1, onto the surface-ofgoid substrates. In solutionin dry hydrocarbonsolvents, 11-mercaptoundecanyl-1- revelsibly for-rys a tri-ru-mercaptoalkylboroxinethat adsorbsonto gold and folms a hydrophobicboroxine bilayer: thi-s bilayer is maiginally stablg in air; it can be completelyhydrolyze!, to a monolayerin aqueousethanol, yieliing a hydiophiiic, tz-A-thick SAM that preJentsboronic acid grlups at its surface. Under - laturat"ed.y.loo.tu.te, SAMs exposingthe borbnicacid groupsdo not titrate appreciablyover a pH rang_e of 0 to t t-ihat is, the contact angle ii independentof the pH of the contactingaqueous . At pq 18 to 14, there is a sharp increas-ein the hydrophilicity ofthe SAM, indicating an-onsetof ionization of or under vacuum,the boronicacid tail groupsdehyllrate the boronicacid groups. In dry hydrocarbonsolvents 'C rapidly and revJrsibiy and form a cross-linkedborate glass at the surface. At 147 in hexadecane,SAMs travrni this cross-lini

Introduction sorption,l2 X-ray-induced damage to organic materials,l7 This paper describesstudies of the preparation and wetting,z's'7-10'12'18and 'ie characterization of self-assembledmonolayers (SAMs) Boronic acids are interesting groups to incorporate into formed by the adsorption of an @-mercaptoalkylboronic the surface of SAMs. In solution, alkylboronic acids react acid, 1l-mercaptoundecanyl-l-boronicacid, 1, onto a gold with simple ,diols, and polyols and form covalent substrate. This work is part of a program to study the reversibly under mild and easily controllable physicalorganic chemistry of organic interfaces.2 SAMs reaction conditions.20-22The ability of boronic acids to of alkanethiolates on gold are particularly useful sub- perform -specific molecular recognition in gels has strates for studying organic surface chemistry because lormed the basis of affrnity chromatography of carbohy- the high affinity of for gold3-6allows the incorpora- tion of a wide range of tail groups, X, into precursor alkanethiols ofthe general formula HS(CH2)"X.7-15SAMs of alkanethiolates on gold are useful model systems in fundamental studies of electrochemistry,lGprotein ad-

E Abstract published in Aduance ACS Abstracfs, May 15, 1994. (1) This research was supported in part by the Office of Naval Research and the Defense Advanced Research Projects Agency. (2) For general reviews, see: Bain, C. D.;Whitesides, G.M.Angew. Chem.,Int. Ed. Engl. L989,28, 506-512. Whitesides, G. M.; Laibinis, P. E. Langmuir L99O,6, 87-96. (3) Nuzzo, R. G.; Allara, D. L. J. Am. Chem. Soc. 1983, 105,4481- 4483. (4) Porter, M.D.; Bright, T'.B.;Allara, D. L.; Chidsey,C. E. D. J.Arn. Chem. Soc. 1987, 109, 3559-3568. (5) Bain, C. D.; Troughton, E. B.; Tao, Y.-T.; Evall, J.; Whitesides, G. M. J. Am. Chem. Soc. 1989. 111. 321-335. (6) Nuzzo, R. G.; Zegarski, B. R.; Dubois, L. H. J. Am. Chem. Soc. 1987. 109.733-740. (7) Bain, C. D.;Whitesides,G. M.J.Am.Chem. Soc.1988, 110,6560- 6561. Bain, C. D.; Evall, J.; Whitesides,G. M. J. Am. Chem. Soc. 1989, 111,7155-7164. Laibinis, P. E.;Whitesides, G. M. J. Am. Chem. Soc. 1992, 114,1990-1995. (8) Nuzzo, R. G.; Dubois, L. H.; Allara, D.L.,I. Am. Chem.Soc. 1990, 112,558-569. Dubois, L. H.; Zegarski, B. R.; Nuzzo, R. G. J. Am. Chem. Soc. 1990, 112,570-579. Chidsey, C. E. D.; l,oiacono, D. N. Whitesides, G. Langrnuir 1990, 6, 682-691. (9) X: alkyl : Bain, C. D.; Whitesides,G. M. J. Am. Chem' Soc' G. M. Science 1988, /t0, 5897-5898. (10) X : : Bain, C. D.; Whitesides, G. M. Langrnuir t993.114, 1989,5, 1370-1378. (11) X : electroactive group: Chidsey, C. E. D.; Bettozzl C. R.; Putvinski, T. M.; Mujsce, A. M. J. Am. Chem. Soc. 1990, 112,430L- 4306. Hickman, J. J.; Ofer, D.; Laibinis, P. E.; Whitesides, G. M.; Wrighton, M. S. Science(Washingtan D.C.) lggl' 252,688-691.

0743-7463194124L0-2228$04.5010O 1994 American Chemical Society SAMs Containing HS( CH) I B( OH)z Langmuir, Vol. 10, I{o. 7, 1994 2229

drates in a variety ofboronatematrices.23-27 and boronate Scheme l. Synthesis of roMercaptoalkylboronic esters have played a major role in enantioselective acids synthesis.2s-33In bulk solution in dry hydrocarbon a,b c,d Br-(CH2)e-CH=CH, --"-"+ HS-(CH2)e-CH=CHz + HS-(CH2)1r-B(OH)2 solvents,boronic acids dehydrate readily and form cyclic 75$ 80t trimers (boroxines),eq 1. Dehydration of boronic acid o Thiolacetic acid and NaOMe in methanol, 2 h.b Excess tail groups constrained in position to the surface of a quasi NaOMe in methanol, reflux, t h, N2." 2 equiv HBBr2.Me2S, two-dimensional, ordered SAM would form a surface CHzClz, Nz, 3 h.d Water, 10 min. comprising cross-linked boronic anhydrides (a "borate glass"). R = (CHr)r,-SH

Tl1Tz = Boroxine dryhydrocarbon solvents <.414 Bit"y", or (CH,;t'-59 R = (CH2\iS- vacuum I HS-(CH2)1J.B(OH)2 ^/\^ R\ /o\ /R II BB ,a-ra",(\oA (cH2)rr-sH o'"'o I B 1 I \ lsooctane

Hp I I HS{CH2h1-B(OH)2 We have investigated the interfacial properties of SAMs tl presenting boronic acid groups at the surface ofan ordered alkanethiolate monolayer on gold. The first goal in this = a B(oH), study was to developa convenient,gram-scale synthesis tl = (cHr)11 | | of the @-mercaptoalkylboronicacid, 1. The secondgoal II was to prepare and to characterize SAMs obtained from S =s- 1 and to measuretheirwettability, stability, and reactivity. Boronic Acid We particularly wished to determine how hydrophilic a surface of boronic acid groups would be and at what pH the boronic acid groups would begin to ionize. We also lsooctane Erhanol wished to establish conditions under which the boronic or i I vacuum HP acid tail groups would dehydrate and give a surface layer II ofcross-linked,boronic anhydride groups and to determine ifthis cross-linkedsurface would be more stable to thermal desorption than a structurally analogous SAM that was 7/////t ?v///////////////////////1////2 r-o.. -o.. -G. -O-r not cross-linked. The last goalwas to determine ifboronic nrBBB,n acid groups exposedon the surface of a SAM would react Borate Glass with cis-diols,catechols, simple alcohols,and alkyltrichlo- rosilanes in a manner similar to that of boronic acids in bulk solution. Figure l. Stylizedillustrations ofmonolayer structures. (top) Proposedstructure of the boroxinebilayer that is obtainedby Results and Discussion the adsorptionof2 fromisooctane solution onto gold substrates. This bilayeris hydrolyticallyunstable, exhibits high hysteresis Synthesis of HS(CHz)rrB(OH)2. We required a me- in its contactangles, and is prolablynot well ordered.(middle) dium chain-length substituted with a boronic acid Proposedstructure of the 17 A monolayerof the boronicacid at one end and a thiol at the other. The chemistry formed by exposureof the boroxinebilayer or the monolayer necessaryfor introducing the thiol group has been well of the borateglass to washesryith alcoholand water. (bottom) studied7 and many reagents for hydroboration are com- Proposedstructure of the 17-Aborate glass surface formed by merically available.s4-40The s5mthesisin Scheme1 yields dehydrationof the monolayerof the boronicacid undervacuum or in dry hydrocarbonsolvents like hexadecane.

(23) Poole, C. F.; Zlatkis, A. J. Chromatogr. 1980, 184,99-183. (24)Mazzeo, J. R.; Krull, L S. BioChromatography 1989, 4,I24- multigram quantities of c{r-mercaptoalkylboronicacids 130. from ar-bromoalkenes. (25) Igloi, G. L.; Koessel,H. Methods EnzymoL 1987, 155,433-448. Preparation and Characterization of SAIVIs: Au/ (26) Soundararajan,S.; Badawi, M.; Kohlrust, M.; Hageman,J. Anal. Biochem. 1989, 178, 125-734. {[S(CHz)rr]e(BO)s], Au/S(CH2)uB(OH)2, and Au/tS- (27) Hageman, H. H.; Kehn, G.D. Methods in Molecular , uol (CH2)rlBOl,. Figure 1 is a schematicillustration of the 11, Practical Chromatography; Kenney, A., Fowell, S., Eds.; monolayer structures that are formed from 1. Compound The Humana Press Inc.: Totowa, NJ, 1992; pp 45-7I. I exists in dry hydrocarbon solvents predominantly as (28) Matteson, D. S.; Jesthi, P. K.; Sadhu, K. M. Organometallics 1984.3. 1284-1288. the boroxine 2. From solution in isooctane,the boroxine (29) Matteson, D. S.;Ray, R.;Rocks,R. R.;Tsai, D. J. Organometallics 2 adsorbed onto gold substrates and formed a boroxine 1983,2,1536-1543. bilayer that was marginally stable to and that (30)Tsai, D. J. S.; Jesthi, P. K.; Matteson, D. S. Organometallics 1983.2. 1543-1545. had a thickness (as determined by both X-ray photoelec- (31)Matteson, D. S.Acc, Chem.Ees. 1988.21.294-300. tron spectroscopy(XPS) and ellipsometry) that varied with (32) Matteson, D. S. Pure Appl. Chem. 1991, 63, B3g-S44. the conditions used in its preparation. XPS analysis (33)Matteson, D. S. Chem..Reu.1989, 89, 1535-1551. showed a strong thiol signal, in addition to the thiolate (34)Brown, H. C.; Ravindran, N.; Kulkarni, S. U. J. Org. Chem. 1980.45. 384-389. signal that is normally observed in SAMs. The surface (35) Brown, H. C.; Bhat, N. G.; Somayagi, Y. Organometallics lg83., of this boroxine bilayer was hydrophobic, and high 2,1311-1316. hysteresis in contact angles measured w"ith drops of (36) Racherla, U. S.; Khanna, V. V.;Brown, H. C. Tetrahedron Lett. 1992.33. 1037-1040. buffered water suggestedthat the bilayer was disordered (37) Brown, H. C.; Campbell, J. B. J. Org. Chem. L980,45,g8g-ggb. or reactive with water. Treatment ofthis boroxine bilayer (38)Lane, C. F.; Kabalka, G. W. Tetrahedron 1976,32. 981-990. with ethanol and water yielded a hydrophilic monolayer (39)Brown, H. C.; Gupta, S. K. J. Am. Chem.Soc. 1gZ1,gJ, 1816- 1818. composedofboronic acid tail groups thatwas consistently (40) Brown, H. C.; Gupta, S. K. J. Am. Chem. Soc. lg7b, 97, S24g- 17 A thick as measured by ellipsometry. Dehydration of 5255. this monolayer under vacuum or by exposure to a dry 2230 Langmuir, Vol. 10, No. 7, 1994 Carey et al.

We used the relative intensities ofthe two sets ofAu(4fl peaks to determine the difference in thickness between the monolayer and the partial bilayer. The attenuation ofthe signal was modeledas an exponential decaythrough the hydrocarbon overlayer using an attenuation length of 40 A for the Au(4f) electrons.4r-43For the layers whose data are presented in Figure 2, the partial bilayer was approximately I A thicker than the monolayer. This 1000 800 600 tmo 2oo o difference in thickness correspondedapproximately to an additional 50Voof a monolayer. For both of the layers, the binding energy of the B(1s) electrons was 192 €V, in good agreement with 8{no the a+ published value forp-chlorobenzeneboronicacid of 191.b I 6000 t :I eV.44 The intensity of the B(ls) peak was compared to that of the corresponding O(1s) peak to obtain a ratio of "t.4ooo + d" {ooo+ T the two elements in the samples. The signal could ,tot ,*al be attributed solelyto the boronic acid groups ofthe layers, or .l because gold does not form a native and because spectra were taken on freshly prepared layers, thus 2q) 195 190 185 173 168 163 158 limiting potential oxidation of the sulfur Binding Enorgy In eV Bindlng Energy in eV atoms.as The raw intensities for both samples were scaledby the Scofield crosssections ofionization (0.486for B(1s);2.93 for O(1s))6 2(X)00 r l2ooo and by the efliciency (or f +t of the detector "6tendue"). The t*T lsooo+ ? magnitude of the latter correction is dependent on the 8OO0 f + f ct!1()ooo+ kinetic energies of the electrons;we used the reciprocal "t" eooo t { t of the kinetic *T t energy as an approximation of this depen- sooo+ 2000+ t dence.aT f + OI sL In the case of the monolayer, we did not include the ffi effect of attenuation of the intensities since we assumed 541 536 531 526 93 88 8il 78 Binding Energy in eV BindlngEnergy in eV that the boronic acid groups were at the monolayer- Figure ?. XPS spectrafor a monolayerand a partial bilayer vacuum interface. Correcting the raw intensities for the derivedfrom the adsorptionof HS(CHz)rrB(OH)zonto gold. fhe crosssections ofionization and the 6tendue yielded a ratio top is a representativesurvey spectrum of the monolayer, of to oxygen of 1.1 in the monolayer. This result insl-udilglabels tor the individual peaks. Notethat the intensity suggests that the surface of the monolayer is predomi- ofthe B(ls) peak at L92 eV is too low to be seenin a survey nantly composedof a cross-linked borate glass, rather spectrum. Belowthe surveyspectrum are individual element scans for boron, sulfur, gold, and oxygen. For the sulfur than offree boronic acids groups, forwhich we would have spectrumof the bilayer, two spectraare shown. SpectrumA expecteda ratio ofboron to oxygen closer to 0.5. is the actual data. SpectrumB is a manipulationof the data To obtain a ratio of boron to oxygen for the partial in spectrumA to show peaks the dueto thiol anddisulfide. This bilayer, we needed to correct the data for attenuation of manipulationwas carried out by scalingthe s^pectrum for the these signals through the monolayerfor attenuationthrough an extra 9 A of hydrocarbon 9 A of additional hydrocarbon (the thicknessdetermined from the ratio of the Au(40 peaks) in the partial bilayer, in addition to the cross sectionsand and subsequentlysubtracting these data from spectrumA. the 6tendue. For this additional correction, we used attenuation lengths of 37.5 and 30 A for the B(1s) and hydrocarbon solvent left a boronic anhydride or "borare O(1s)electrons, respectively, which were estimated from glass" monolayer; this SAM was 1Z A tni.t (by XpS) and an empirical equation that describesattenuation lengths had an atomic ratio of oxygen to boron that was 0.9 as through hydrocarbon as a function of the kinetic enerry determined by XPS. The XPS spectra for four separate of the electrons.a2 Scaling the raw intensities by these samples of the monolayer had an average value of 1.0 * factors yielded a boron to oxygen ratio of 1.1, suggesting 0.15 for the ratio of oxygen to boron. This value suggests formation of boroxine groups between the layers. that the speciespresent on the surface under the ultiahigh Figure 2 also summarizes XPS data for the S(2p)regions vacuum (UHV) conditions of XPS was indeed the ar*ry- for both layers. The spectrum glass dride and notthe boronic acid, forwhich one would expect of the borate mono- a ratio of oxygen to boron of 2.0. layer exhibited the expected two peaks in a 2:1 ratio due We note that all alkylboronic acids readilyinterconvert to spin-orbit coupling of the 2p electrons.a8The binding between the acid (in air) and the boroxine when exposed to heat or vacuum.20 SAMs of 1, which concentrate and (41) Bain, C. D.; Whitesides, G. M. J. Phys. Chem.1989, 93, 1670- present boronic acid tail groups at a monolayer surface, 1673. (42) Laibinis, P. E.; Bain, C. D.; Whitesides, G. M. J. Phys. Chem. can be expected to behave in a similar manner, inter- 1991,95,7017-702r. converting between the anhydride form under vacuum (43) Feldman,L.C.;Mayer, J. W.Fundamentals of Surfare andThin g. _andthe hydrated boronic acid form upon exposure to FiIm Analysis; Elsevier Science: New York, 1986; Chapter humid air. (44) Hendrickson, D.N.;Hollander, J. M.;Jolly, W.L.Inorg. Chem. 1970, 9, 612-615. X-ray Photoelectron Spectroscopy CXPS). XpS was (45) Tarlov, M. J.; Newman, J. G. Langmuir 1992,8, 1398-1405. useful in the structural characterization of a boronic Huang, J. ; Hemmin ger, J. C. J . Arn. Chem. Soc. 1993, 11 5, 3342- 3343. (46) anhydride surface and a partial boroxine bilayer. XpS Scofield, J. H. J. Electron Spectrosc. Relat. Phenom. 1976,9, spectra were (using t24-r37. collected for boron the B(ls) peak), (47) Seah, M. P. In Practbal Surface Analysis by Auger and X-ray sulfur (s(2p) peaks),gold (Au(4f) peaks),and oxygenio(rr) Photoelectron Spectroscopy;Briggs, D., Seah, M. P., Eds.;John Wiley peak). Spectra for a monolayer and a partial bilayer are and Sons: Chichester, 1983; Chapter 5. Using l/E may be a slight presented in Figure 2, along with underestimation of the 6tendue, but it does approximate the effect. a representative survey Large exponents in the denominator would decrease the ratio of oxygen spectrum of the monolayer. to boron.

I 223L SAMs Containing HS(CH) rB(OH)2 Langmuir, VoL 10, No. 7, 1994 energy of the peak at lower enerry, the S(2pBl-2) peak, 0.5 was 161.6 eV, the expectedvalue for thiolates.s'6'15 o oao o'a aa For the biiayer, there are two sets of peaks in the S(2p) a^ . o 0.4 Oe-OOa aOOOOO- region, whose overlap gives rise to the appearance of a oo-o triplet. Superimposed on the two peaks for the thiolate groups are two peaks at higher binding energy. Peak 0.3 fitting the spectrum with four peaks yielded a binding acos0 o Au/S(CHJfiB(OH)2 energy of 163.2 eV for the S(2p312)of the secondset of o.2 peaks.ae The binding energJ of this peak corresponds to o Au/S(CH2)11OH groups in the secondlayer of thiol and possibly disulfide 0.1 the partial bilayer.6 The XPS data present a picture of the layers consistent with the stylized illustrations drawn in Figure 1. For the 0.0 monolayer, the sulfur atoms are adsorbed to the gold -0.0 surface as thiolates, and the boronic acid groups are . = 0" r = 0, AulS(cll2)irB(oHh forming a cross-linked borate-glass-like net- dehydrated, -o.2 o = 0a o =0, aurs{cH2}rloH work. The partial bilayer has all of characteristics of the monolayer plus the presenceof sulfur atoms not bound to (undercyclooctane) the gold surface. The partial secondlayer is stable in the -0.4 ^aaa atat' of the XPS (10-e Torr) au ^^o^o'oooooo ultrahigh vacuum conditions -cos 0 o - "" a suggesting binding of the secondlayer to the monolayer -0.6 presumably through boroxine groups as observed for boronic acids in solution. shows -0.8 o Interfacial Properties: Wetting. Figure 3 30 that the fZ-A-thict- SAU bearing the boronic acid tail sepEcg!goPlioo: groups had a hydrophilicity and hysteresis between the -1.0 'l-T lr lr ll ll l'l advancing and receding contact angles that is similar to 024 6 8 101214 the structurally analogous 11-hydroxyundecane-l-thiol (HUT). Contact angle under cyclooctanes0sug- PrY gested that the boronic acid groups of the SAM obtained Figure 3. (top) Hysteresisin the contactangle of buffered from 1 began to ionize at pH 12; this ionization was aqueoussolutions measured with dropsunder cyclooctane as reflected by a decrease in both advancing and receding a iunction of pH. Values of hysteresiscannot be determined t\^'hich,in practical contact angles. The SAM composed of HUT remained when the recedingangle of water is zero ter;s, iJ any value of 9,Hzoless than or equalto un-ionized over the titration range of pH 0 to pH 14 (as experimental and thereforethere are no valuesof hysteresisfor SAMs of both advancing and 10;) reflected in the independence of L abovepH : 11. (bottom)Advancing and recedingcontact receding contact angles to pH). anglesofbuffered aqueous solutjons measured urth drops-under In air, the contact angles of buffered aqueous drops for cyc"looctaneon monolayersof HS-(CH2)11-B(OHlzand HS- the SAM having terminal boronic acid groups were (eHz)rr-OH. The sizeof the symbolsgives our bestestimate constant over a range of pH 2 to pH 10 (0uH'o: 16-24"). of the error of the measurement. At pH 11 or above, the SAM of the boronic acid was wet completely by buffered aqueous drops. The receding pK^z:12.5 in water). The origin of the apparent shift in contact angles in air for the SAM of the boronic acid were ihe pK" to higher values has not been completely-estab- zero at all values of pH. lished; the environment of the interface might influence These titration results in air and under cyclooctane both the properties of the functional groups and the establish that SAMs of alkylboronic acids do not begin to partitioni.tg olttydtoxide into the interfacial region.2'lo ionize significantly until the aqueous drop is at least pH Desorptlon of Au/tS(CH2)uBOl" in Contact with 11. Thus, in analogy to our previous frndings for SAMs Hexadecane. Manytechnologicalapplications ofSAMs, bearing carboxylic acid and phosphonic acid tail groups,lo'5r particularly in microelectronic device fabncation, will alkylboronic acids at the monolayer-water interface iequire greater thermal stability than that displayed by require contact with stronger for ionization than SAMs of simple alkanethiolates on gold. Our prior studies alkaneboronic acids in bulk solution (ca. pK"t : 9.2 and of the thermal desorption of alkanethiolates from gold have used methyl-terminated SAMs thathave low surface (48) Briggs, D.; Rivi6re, J. C.InPractical Surface Analysis b1 Auger free energies;the SAMs of alkaneboronic acids have high and X-ray-Fhotoelectron Spectroscopy; Briggs, D., Seah, M. P., Eds.; surface energies and contaminate more easily than ratio John Wiley and Sons: Chichester, 1983; Chapter 3. Althoughthe methyl-terminated SlUVIs.2'10'51Typically, thermal de- of the peaks is expected to be 2:1 due to spin-orbit coupling, the slight apparent increase in the intensity of the S(2pvz) peak and the small sorption is measured by exposing the SAM to a reservoir shoulder on the high energy side of these peaks is due predominantly of hexadecane held at constant temperature and measur- to sulfur- bond cleavage during irradiation: the intensity of the ingthe change in ellipsometric constants ofthe SAM (after peaks increased with the amount of irradiation. higher energ'y and washing) as a function of time. We l+g) We frt the peaks by constraining the binding energy of S(2prn) removing peak for the thiolate to be 1.2 eV gxeater (or 162.8 eV) than the S(2prn) determined the thermal stability of the SAIVIsin this study of the thiolate and also constraining its intensity to be half that of the inhexadecane at 147 "C: the choiceoftemperature allowed S(2p) S(2pgp)peak. For information on the spin-orbit coupling of the reproducibly to observe complete desorption of the peak, see: Salaneck,W.R.;Lipari, N.O.; Paston,A.;Zallen, R.; Liang, us K. S. Phys. Reu.B 1975, 12,1493-1500. monolayer in a short time.52 (50) The contact angle were performed under cyclooctane, Figuie 4 is a plot of the rate of desorption vs time for Ca, in order to increase the contact angles and to reduce the rate of SAMs of I and SAMs of HUT (used as a reference). HUT contamination of the surface ofboronic acids. Under air as the ambient medium, a pure, clean surface of the boronic acid is wet (or very nearly pH, but the high interfacial free energJ (52) Although we observed thermal desorptior of the monolayers at wet) by water at all values of .c, causes the surface composed of boronic acids to become contaminated g0 to 90 quantitative analysis of the data was hinderqlby-atendency reason, at a rate that is fast enough to interfere with the measurements. Cg ofthe monol-ayers to contaminate as a function oftime. For this purity at low cost. we determined the thermal stability of the SAMs in this study at a is relatively nonvolatile and can be obtained in high oC. (51) Lee, T. R.; Carey, R. I.;Whitesides, G. M. Unpublished results. higher temperature, 147 2232 Langmuir, Vol. 70, No. 7, 1994 Carey et al,

20,0 1. Wetting Properties and Thicknesses of SAIVIs Acido ^ 15.0 Derived from I l-Mercaptoundecanylboronic

o r 100 monolayer system 9,Hro o,H"o thickness, A o c Au/S(CHz)rrB(OH)z 150 t7b 6o 3t.. Au/t tS(CHz)r r le(BO)e) 97 50 2gb An/S(CHz)rrB(OH)z* ZtOClz * 150 17b C+HgB(OH)2" Au/S(CH2)rrB(OH)z* ZrOClz * 15 17b HzNCHzCHzPOsHz" * CreHszSiClsd t02 65 37" F Au/S(CH2)118(OH)z Au/S(CHz)rrB(OH)z* 110 74 25" CFs(CFz )s( CHz )zSiClsd Au/S(CHz)rrB(OH)z* 58 30 CFs(CFz)z(CHz)zOH/ Au/S(CHz)uB(OH)z* catechof 62 35 2V Au/S(CHz)118(OH)z* Pinanediol 95 78 2lbe 93 74 17bs n,, /-: Au/s(cHz)rr-B<:'"' - -o"/-,vt'Yl Tlme (sec) Figure 4. Thermal desorptionof monolayersof thiols on gold 'C: a in iontact with hexadecaneat L47 (O) 1l-mercaptoundec- Q^H2oand 9rH'orefer to advancingand recedingcontact angles acid;(O) 1l-hydroxyundecane-1-thiol.In the of water, respectively, using drops of water buffered at pH 7. anyl-1-boronic b firit-order plot of the correctedthicknesses against time, ?r - Thickness determined by XPS relative to a monolayer of undec- ?- representsthe differencein ellipsometricthickness between anethiol at 15A. " Theseexperiments attempted to build multilayers long times. The inset showsa plot of ellipsometric onto the surface of sAMs of I by the sequential deposition of zroclz time 7 and (See thicknessvs time. and either an alkylboronic acid or an alkylphosphonicacid. refs 52-54). By XPS, no zirconium was observed,and there was no increasein the thickness of the films. d SAMS of l were treated and t have the same C11alkyl chain length and similar with an alkyl- or perfluoroalkyltrichlorosilane in dry CHzClzor in solubility in hexadecane. The primary difference in the the vapor phase.-"Thickness determined by ellipsometry. f 11tit chemical properties of the two SAMs is that the SAM of boronate sAM and analogoussAMs derivatized with ethylene the boronic acid readily dehydrates to a cross-linked, glycol, octanol, and dodecanolwere not stable to high vacuum' prolongedexposure to borate glass surface that we expected to be more stable excessivewashes with aqueousethanol, or e A thick SAM of a pinanediol boronate ester was desorption than the similar, but non-cross- air. The 21 toward thermal prepared by the esterification of a preexistinq SAM o-f 1 with linked surface of a SAM of HUT. The data in Figure 4 'C, pinanediol in solution from isooctane. The 17 A thick SAM of a show that in hexadecaneat 147 a SAM of the boronic pinanediol boronateester was preparedby the adsorptionof3 onto acid was approximately 5 times more stable to thermal i gold substrate. Both ofthese SAMs were stable to repeatedwashes desorption than a SAM ofHUT. These data are consistent with water and to vacuum. with the hypothesis that cross-linking the tail groups in SAMs enhancestheir stability to thermal desorption, but SAM obtained from 3 (see Table l-).53 The conclusion from it may also reflect differences in solubilities. the experiments with pinanediol is that SAMs presenting (the Reactions: Formation and Stability of Boronate boronic acid tail groups react cleanly and rapidly Esters on SAMs. Alkylboronic acids in solution can form derivatization requires only 5 min in a 5 mM solution of proper cis- strong covalent bonds reversibly under mild conditions pinanediol in isooctane) if presented with the with a-hydroxyacids, catechols,y-hydroxyaldehydes, and diol. cis-diols. In this sectionwe describethe reactions of SAMs In solution, boronate esters of catechols are much less pinanediol. of 1 with pinanediol, catechol,simple alcoholsand glycols, stable toward hydrolysis than esters of and alkyltrichlorosilanes (Table 1). Treatment ofa boronic acid-terminated SAM with catechol from dilute solution in isooctane or from concentrated We compared the properties of SAMs prepared by the solution in yielded a derivatized monolayer surface adsorption of 3 from solution in isooctane to SAMs lHF that was 3 A thicker than the underivatized boronic acid prepared by the esterification in isooctane solution of a SAM and that had an advancing contact angle with water preexisting monolayer of I with pinanediol. We chose that varied from 62' to 65" depending on the sample. pinanediol as the diol constituent because the resulting Although several washes with ethanol and water did not boronate esters are very stable and easily manipulated seem to change the wettability or thickness, a prolonged without complicationsdue to hydrolytic instability.20'26'27'ea exposure (overnight) to aqueous ethanol reversed the Compound 3 was prepared by refluxing equimolar esterification and yielded the surface exposing boronic amounts of 1 and (*)-pinanediol inheptane for 10 min (eq acid groups again. When an analogous esterification was 3 was stable to an aqueous extractive 2). Compound attempted with 3-fluorocatechol, very little change in to elution through a of silica gel. workup and thickness or wettability was observed, and after one wash with ethanol and water, no fluorine could be detected by heptane'reflux> -1--/ "l>t-(cH2)!l-sH^ XPS. We conclude that SAMs of 1 form only moderately HS-(.H2)1r-B(oH)2 X I stable boronate esters with catechols and that the greater v,,'"v (2\ the electron drawing constituent on the catechol, the lower \-9tot - .r 3 the stability of the derivatized monolayer.sa f7l \l/ (53) The reason for the observed difference in thickness and wet- tability of the two SAMs is probably that the tail group of 3 is so bulky, that it prevents the formation of a complete monolayer. In_th_e_caseof SAMs formed from 3 and SAMs formed by reaction of the SAM formed by the surface reaction, the underlying SAM of the pinanediol with a preexisting SAM of 1 are both hydro- boronic acid has already formed from I and is densely packed and highly react with these ordered boronic acids at phobic, is stable toward water (pH : 7) for at ordered. Pinanediol can then and each the surface, filling up all of the sites and making a more ordered, more Ieast 24h. The SAM formed in the surface reaction was densely packed layei ofpinanediol boronate terminal groups than the actually slightly more hydrophobic and thicker than the monolayer formed from 3. SAMs Containing HS(CH, r B( OH)z Langmuir, Vol. 10, No. 7, 1994 2233

Although multilayers can be formed from SAMs of stable to hydrolysis. Derivatization of boronic acid or phosphonic acids by the sequential deposition of ZrOClz boronic anhydride monolayers with alkyltrichlorosilanes and o,ar-alkylphosphonic acids,ss-s?attempts to build yields hydrophobic surfaces that are held together by a analogous multilayer systems from ZrOClz and alkylbo- borosilicate network. ronic acids were unsuccessful due to the incomplete formation or the hydrolytic instability of the zirconium Experimental Section boronate compounds. XPS spectra contained no peaks for zirconium and showed no increase in the thickness of Materials. Absolute ethanol (Quantum Chemical Corp.) was prior (Aldrich, a SAM of 1 after it had been treated with ZrOClz and deoxygenated with Nz or Ar to use. Isooctane 99Vo) and hexadecane (Aldrich, 99Vo) were percolated twice (seeTable 1). C4He-B(OH)z through activated, neutral alumina (EM Science). Water was Formation of a Borosilicate Bilayer on SAIVIs: deionized and distilled in a glass and Teflon apparatus. Un- Reaction of Alkyltrichlorosilanes with Boronic Acid decanethiol (Pfalz & Bauer) was distilled prior to use. Un- Tail Groups. Partial bilayers attached by borosilicate decylenic bromide (Pfalz & Bauer) and dibromoborane dimethyl bonds have been formed by treatment of SAMs of 1 with complex (Aldrich) were purchased and used without either oftwo different alkyltricNorosilanes : C 16H3?- SiCl3 further purifi cation. Hydroxyundecanethiol was available from and CFg(CF2)dCH2)z-SiCls.These reactions were carried previous studies.T out both with the alkyltrichlorosilane in the vapor phase Preparation of Substrates. Gold substrates were prepared -200 (Materials and in solution in dry CHzClz. The advancing contact by electron-beam evaporation of A of gold angles of water and hexadecane on these bilayers were Research Corp., Orangeburg, NY 99.999Vo)onto single crystal silicon(100ttest wafers (Silicon Sense: Nashua, NH; 100 mm closetothose oftypical long-chain alkyl and perfluorinated -500 precoated g,HD: diameter, rm thick) that had been with 50-100 alkyl-terminated SAMs (0^H"o: 115 and 44). The A of titanium to improve adhesion. The substrates were stored properties ofthese surfaces also compare well with bilayers in Fluoroware wafer holders until used in experiments, generally formed in a similar fashion from the reaction of - as soon as possibleafber evaporation. terminated SAMs with alkyltrichlorosilanes,ss although Formation of Monolayers. Adsorptions were carried out the hysteresis in the contact angles of drops of buffered in 10-mL glass weighing bottles that had been cleaned with water is slightly higher for the borosilicate bilayer than "piranha solution" (7:3 concentrated HzSOq/3IVoHzOz) at 90 "C that observedfor analogously derived siloxane bilayer from for t h, and rinsed first with distilled water and then with copious HUT.59 amounts of deionized water. WARNING: "Piranha solution" should be handled with caution. It should not be allowed to contact significant quantities of oxidizable organic materials. In Conclusions somecircumstances (Most probably when mised with significant The ar-mercaptoalkylboronic acid, 1, adsorbs onto gold quantities of an oxidizable organic material), it has d,etonated from isooctanesolution and forms a boroxine bilayer than unexpectedly.6OThe weighing bottles were stored in an oven at 100 'C use, Adsorptions were carried out in 10 mL can be hydrolytically cleaved by treatment with aqueous above until of deoxygenated solvent (ethanol or isooctane) at a total thiol yield ethanol and a 17 A thick monolayer bearing boronic concentration of 1 mM. acid tail groups. Under high vacuum or in dry organic Thermal Desorption Experiments. Desorptions were solvents, the boronic acid monolayer reversibly dehydrates carried out in an unstirred solution of hexadecane in a 25-mL and forms a cross-linked boronic anhydride or "borate weighing bottle that was partially immersed in an oil bath glass" surface. This cross-linking to a 5-fold decrease thermostated at 153 + 1 "C. The hexadecane was percolated in the rate of thermal desorption in hexadecane atl47 "C twice through a column of activated alumina and was purged relative to the rate of thermal desorption observed for a with Nz immediately prior to use. The temperature of the structurally analogous monolayer of hydroxyundecane- hexadecanein the glass bottle was I47 + I "C. Slides (-1 cm thiol. Boronic acid or boronic anhydride groups on the x 3 cm) with preformed monolayers were immersed in 20 mL of surface of a SAM can be derivatized reversibly with cis- hot hexadecane, removed at the proper time intervals, rinsed with heptane, and blown dry with a stream of . diols and formhydrophobic surfacescomposed ofboronate Ellipsometry was used to follow the desorption ofthe monolayers; esters;boronate ester groups formed from pinanediol are ellipsometric constants were measured on the complete mono- layer and then after each removal from solution. The thickness (54) Attempts to derivatize SAMs of the boronic acid, 1, with simple of the adsorbed organic layer was calculated using the optical alcohols or glycols such as , octanol, or perfluorododecanol constants for clean gold. Corrected thicknesses (?, - ?-) were yielded derivatived monolayers that were at best transiently stable as calculated by subtracting from the ellipsometric thickness at judged by XPS, ellipsometry, and contact angles. Afber treatment of readings obtained long SAMs of l with simple alcohols and glycols in a dryhydrocarbon solvent, time / the mean of several at times after the contact angles with water and hexadecane increased and there desorption was essentially complete. were small changes in the ellipsometric constants. In all such cases, Instrumentation. Ellipsometric measurements were per- however, the derivatized SAMs were highly unstabie in the presence formed on a Rudolf Research Type 43E03-2008 ellipsometer of moisture, and extensive washing with ethanol and water always equipped with a He-Ne laser (,1: 6328 A) at an incident angle restored the underivatized, surface acids. The reasons for ofboronic were rinsed in this observation are that simple alcohols and glycols do not form boronate of 70'. Samples and blown dry a stream ofnitrogen esters that are sufficiently stable to make a surface that is resistant to prior to characterization. Values of thickness were calculated hydrolysis. Partial derivatization is typically not detected by XPS using a program written by Wasserman,6l following an algorithm because if the alcohol is volatile, the small amount the alcohol adsorbed by McCrackin and co-workers;62 in the calculation, we used a to the surface prior to an XPS experiment will be removed under the refractive index of 1.45 for the SAMs.63 high vacuum conditions in the XPS chamber. Contact angles were measured on a Ram6-Hart Model 100 (55) Lee,H.; Kepley,L. J.;Hong, H.-G.;Mallouk,T.E.J.Am.Chem. Soc. 1988, 110,6L8-620. Lee, H.;Kepley, L. J.; Hong, H.-G.;Akhter, goniometer at room temperature and ambient humidity for both S.; Mallouk, T. E. J. Phys. Chem.1988,92,2597-260L. water and hexadecane. Advancing and receding contact angles (56) Putvinski, T. M.; Schilling, M.L.; Katz,H. E.;Chidsey, C. E. D.; were measured on both sides each of at least 3 drops of each Mujsce, A. M.; Emerson, A. B. Langrnuir 199{J,6,1567-157t. liquid per slide; data in the frgures represent the average ofthese (57) Vermeulen, L.; Thompson, M. E. Nature 1992,358, 656-658. (58) Ulman, A.; fillman, N. Zongrnuir 1989, 5, l4L8-I420. (b) Netzger, L.;Sagiv, J.J.Am.Chem. Soc. 1983, 105,674. (c) Pomeranz, M.; Segmuller, A.; Netzer, L.; Sagiv, J.Thin SolidFilms 1985,132,153. (59) In our hands, treatment of monolayers of HLIT with octadecyl- trichlorosilane in dry CH2CI2yielded bilayers that had hysteresis in the contact angles with water consistently l}Volower than bilayers prepared in an analogous fashion from monolayers of I and octadecyltrichlo- rosilane. 2234 Langmuir, Vol. 10, No. 7, 1994 Carey et al. measurements. A Micro-Electrapette syringe (Matrix Technolo- half-saturated ammonium solution. The aqueouslayer gres. Lowell, MA) was used for dispensinfremoving the liquids was extracted with CHCIg (2x). The CHCI3 extracts were onto/from the SAMs (ca. 1 ml/s). The method usedfor measuring combined, washed with HzO (2x), dried over MgSOa, and the advancing and receding angles has been described previ- concentrated in vacuo to yield a crude oil (1.5 g, 94Vo)that was ouslY.2'10 purified by flash chromatography using hexanes/ethyl XPS was carried out using a SSX-100 spectrometer (Surface (95:5)as eluent to yield a clear colorlessoil (1.2g,75%o). lH NMR ScienceInstruments) using monochromaticAl Ka X-rays (photon (400 MHz, CDCh) d 5.8 (m, 1 H), 4.95 (m, 2 H), 2.52 (q, J = 8 enerry :1487 eV; the peaks are referenced to Au(4fid at 84.0 Hz,2H),2.04 (q,2 H), 1.6(m, 2H),t.2-1.4 (-, 13 H). MS EI eV). For quantitation of the individual elements in the SAMs, (pos): 185 (M+). we used the 4f doublet for gold ("Au(40" at 84 and 87 eV for l-Mercapto-ll-undecanylboronic Acid. To a solution of Au(4fzn')and Au(4fsrz),respectively), peaks the ls for carbon, l0-undecene-1-thiol(1.1 g, 5.9 mmol) at 0 "C in freshly distilled oxygen,and boron at284,532, and 192eV, respectively,and the CHzClz was added via syringe dibromoborane dimethyl sulfide 2p peaks for sulfur between 161 and 164 eV. Data for gold, complex(12.0 mL, 1.0M in CHzClz,12 mmol). The reactionwas carbon, and sulfur were taken using a spot size of 600 4m and allowed to warm to room temperature over 15 min and was a pass energy of 50 eV. Data for oxygen and were boron taken refluxed for 3 h. Most of the CHzClzwas removed in vacuo, and using a spot sizeof 1000prm and a passenerg'y of 100eV because the reaction was then transferred to a stirred mixture of water of the low intensity ofthe boron peak; both regions were checked (10 mL) and ether (50 mL). The suspensionwas stirred for 10 at the lower pass enerry to show that there was only one peak min and transferred to a separatory funnel. The aqueous layer in each region. T\vo scans were taken for gold, three scans for was extracted with ether (4 x ), and the ether layers were combined carbon, eight scans for oxygen, thirty scans for boron, and frfty and extracted with brine (2x), dried over MgSOa, arrd concen- scans for sulfur. Each scan at the pass energy of 50 eV took trated in vacuo to yield a white solid (1.31 g, 96Vo)that was approximately 1.5 min; each scan at the pass energJ of 100 eV purified by flash chromatography using hexanes/ethyl acetate took approximately 2 min. (3:1)as eluent to yield a white solid(1.1 g,80Vc). lH NMR (400 Peaks in the spectra were fitted using a 80VoGaussian/207o MHz, CDCIg)6 2.48(q, 8 Hz, 2 H), 1.55(m, 8 Hz, 2H),I.2-L4 Lorentzian peak shapes, using a Shirley background subtrac- (m, 17 H), 0.82 ft, 8 Hz,2 H). FABMS (glycerolmatrix): 289 tion.6a For each present, the sulfur peaks were f(NVglycerolester)H*1, 576 f2(Wglycerol ester)H+]. FABMS fit using two peaks,corresponding to the S(2p3,2)and S(2p12) (nitrobenzyl alcoholmatrix): 350 l(ltl/mononitrobenzylester - peaks due to spin-orbit coupling.a3'asThe ratio of these peaks HzO )H*1,503 [(tr&dinitrobenzyl ester)H+], 1006 is 2:1, and the splitting betweenthem is 1.2 eV.as In the fitting t2(lwdinitroben- zyl ester)H'1. program, these values were used as constraints to obtain realistic fits to the data. For both the monolayer and the bilayer, two sets (*)-Pinanediol-l-mercapto-11-undecanylboronate. A of peaks were used in the fit. In the former, the set of peaks at solution of 1-mercapto-11-undecanylboronicacid (100 mg, 0.43 higher energT was shown to be due to X-ray-induced damagelT mmol) and (*)-pinanediol (73 mg, 0.43 mmol) in heptane was at the sulfur-carbon bond: the relative intensity of these peaks refluxed for 10 min and then left at ambient temperature for 1 increased with time of irradiation from 25Voto 43Voaft,er 4 h of h. The heptane was removed in vacuo, and the resulting oil was irradiation. For each sample, spectra in the S(2p) region were passedthrough a column of silica gel using hexanes/EtOAc (97: taken first so as to minimize effects due to damage. 3) as eluent. The appropriate fractions were collected to yield lH l0-Undecene-l-thiol. To a clear solution of NaOMe (465 a clear oil (130 mg,807c). TLC: fu: 0.8 (hexanes). NMR mg, 8.6 mmol) in degassedmethanol (10 mL) was addedthioacetic (300 MHz) 6 4.26-4.24 (1 H, dd, CHOB pinanyl);2.54-2.47 (2 acid (0.70 mL, 1.15 equiv) and undecylenicbromide (2 g,8.6 H, q, CH2-Sl 2.37-2.29(1H, D, pinanyl);2.24-2.18 (1 H, m, mmol). The reaction was refluxed under Nz for 3 h. Additional pinanylt;2.05-2.03 (1 h, t, pinanyl); 1.92-1.88(1 H, septet, NaOMe (930 mg, 17.2 mmol) was added, and the reaction was pinanyl);1.85 and 1.81(1 H, m, pinanyl);I.64-1.52(2H,m, refluxed for another t h. Most of the CH3OH was removed in Cl-Hz); 1.41 I.39 (2 H, m, C,t-Hz);1.38(3 H, s, CHe);1.33 (1 vacuo, and the reaction was quenched by addition ofdegassed, H, t, SH); I.29 I H, s, CH3);1.25 (14 H, s);0.84 (B H, s, CHs); 0.82-0.78 (2H,t, CH2-B). High resolutionFABMS: calcdfor (63) In the calculation ofellipsometric thickness, we assume an index CzrHssBOzS,366.2764; found, 366.2760. ofrefraction of 1.45; the indices ofrefraction for various similar to those used in this study vary from 1.42 to 1.46: undecane (liquid), Acknowledgment. 1.42; undecanol (liquid), 1.44; undecanethiol (liquid), 1.46; octa- XPS spectra were obtained by decanethiol (solid), 1.46; docosane(solid), 1.45. CRC Handbook of using instrumental facilities purchased under the DARPA/ Chemistry and Physics,65th ed.; Weast, R. C., Ed.; CRC Press, Inc.: URI program and maintained by the Harvard University Boca Raton, FL, 1984. This pqtential variation in the refractive index Research laboratory. NMR spectra were obtained in the affects the thickness <*1.5 A, which is less than the variation we normally observe for the SAMs of alkanethiolates on gold. Bruker NMR Facility at Harvard. R.I.C. was an NIH (64) Shirley, D. A. Phys. Reu.B 1972,5,4709-4714. Postdoctoral Fellow (GM-13678).