copyright o reeb by the A-""i"""TflJili'""i ff#tfl1TYJh133t;j1p""-i..ion or the copyright owner. 813

Self-Assembled Monolayers of Long-Chain HydroxamicAcids on the Native Oxides of Metalsl

John P. Folkers, Christopher B. Gorman, Paul E. Laibinis, Stefan Buchholz, andGeorge M. Whitesides*

Department of Chemistry, Haruard Uniuersity, Cambridge, Massachusetts 02138

Ralph G. Nuzzo*

Departments of Materials Science and Engineering and Chemistry, Uniuersity of lllinois,Urbana-Champaign, (Irbana, Illinois 61801

Receiued September 79, 1994e

Long-chain alkanehydroxamic acids adsorb on the native oxides of metals and formed oriented self-assembled monolayers (SAMs). This study examined SAMs of hydroxamic acids on the native oxides ofcopper, silver, titanium, aluminum, zirconium, and iron. These SAMs were characterized usingwettability,X-ray photoelectron spectroscopy (XPS), and polarized infrared external reflectance spectroscopy (PIERS).Alkanehydroxamic acids give better monolayers than the corresponding alkanecarboxylic acids on certainbasic metal oxides (especially copper(Il) oxide). On the native oxide of copper (which has an isoelectricpoint greater than the pK" of the hydroxamic acid), the ligand is bound to the surface predominantly asthe hydroxamate. The strength of the interaction between copper oxide and the hydroxamate allowsincorporation of polar tai l groups into the monolayer. On acidic or neutral metal oxides (e.g., TiO2), thepredominant species bound to the surface is the hydroxamic acid. Alkanehvdroxamic acids on titaniumdioxide bind relatively weakly but. nonetheless. form SAlts that are more stable than those from carboxylicacids (although not as stable as those from alkanephosphonic acids r.

Introduction

We have examined the properties of self-assembledmonolayers (SAMs) obtained by the adsorption of long-chain alkanehydroxamic acids (general formula CHs-(CHz),CONHOH) onto the native oxides of freshly evapor-ated frlms of copper, silver, titanium, aluminum, zir-conium, and iron and have compared these SAMs to otherscommonly used with these substrates. We have usedwettability, X-ray photoelectron spectroscopy (XPS ), andpolarized infrared externai reflectance spectroscopy(PIERS) in characterizingthese monolayers. We concludethat monolayers of alkanehydroxamic acids (or theiranions) are more stable than monolayers obtained by theadsorption of carboxylic acids on all of the surfaces(especially copper oxide and silver oxide), but they areless stable than monolayers obtained by the adsorptionof alkanethiols to copper or silver.

Of the many systems available for the formation of self-assembled monolayers, monolayers obtained by the chemi-sorption of alkanethiols onto gold and silver,2 a monolayersobtained by the hydrolysis and polymerization of alkyltrichlorosilanes onto hydrated surfaces,5'6 and monolayersobtained by the adsorption of carboxylic acidsT-ls orphosphonic acidslG onto metal oxides have been the most

8 Abstract published in Aduance ACS Abstrocls, January 15,1995.

(1) This research was supported in part by the Office of NavalResearch and the Advanced Research Projects Agency (ARPA). XPSspectra were obtained using instrument facilities purchased under theARPAruRI program and maintained by the Harvard UniversityMaterials Research Laboratory. S.B. acknowledges a grant from theStudienstiftung des deutschen Volkes (BASF-Forschungsstipendium ).R.G.N. acknowledges support from the National Science Foundation(cHE-930095).

(2) Nuzzo, R. G.; Al lara, D. L. J. Am. Chem. Soc. 1983, 105,4481-4483. Porter , M. D. ; Br ight , T. B. ;Al lara, D. L. ; Chidsey, C. E. D. J.Am. Chem. Soc. 1987, 109, 3559-3568. Bain, C. D. ;Troughton, E. B. ;Tao, Y.-T.; Evall, J.; Whitesides, G. M.; Nuzzo, R. G. J. Am. Chem. Soc.1 9 8 9 , 1 1 1 , 3 2 1 - 3 3 5 .

(3) Walczak, M.M.;Chung, C. ;Stole, S. M.;Widr ig, C.A. ;Porter , M.D. J. Am. Chem. Soc. 1991. 113. 2370-2378.

( 4) Laibinis, P. E. ; Whitesides, G. M.; Allara, D. L. ; Tao, Y.-T. ; Parikh,A.N.; Nuzzo, R. G. J. Am. Chem. Soc. 1991, 113,7152-7767.

thoroughly studied.l; SAMs of alkanethiolates on goldand silver are generally the most useful in scientificapplications,l8 because of their stability and because thespecificity of coordination of the sulfur atom to the metalallows polar groups containing oxygen and nitrogen to beincorporated into the SAM.2'|1're

Alkanethiols do not adsorb, however, to the surfaces ofmany metal oxides.20 Carboxylic acids and phosphonicacids adsorb, but the adsorption can be weak.7-16 These

. ( 'hen. Srrr ' . 1980. 102.92*98. Netzer L. ; Sagiv,1983. 105. 671- 676. Moaz, R. :Sae-tv, J.J.Col lo idt 0 0 . 1 6 5 4 9 6 .

S, R : Tao. \ ' . -T. : Whi tesrdes, G. M. Langmuir 1989,

D. Chem. Phys. Lett. 1990. 167. 198-

07 43-74631951241I-0813$09.00/0 @ 1995 American Chemical Society

814 Langmuir, Vol. 11, No. 3, 1995

systems also cannot be used to incorporate polar groupsinto a monolayer.ll Phosphonic acids have large headgroups and can be inconvenient to synthesize.

We have examined hydroxamic acids (1) as an alterna-tive to carboxylic acids (3) for self-assembled monolayerson the surfaces of metal oxides because their complexationconstants with metal ions in aqueous solution are severalorders of magnitude larger than those ofcarboxylic acids2land because they have similar sizes. Synthesis of long-chain hydroxamic acids is also straightforward.22'23

When complexed to a single metal ion in solution, thegreater stability of complexQs with hydroxamate relativeto those with carboxylate reflects the larger bite size ofthe hydroxamate, which forms five-member chelate ringswith metal ions (2t. Incontrast, carboxyiate forms eitherstrained. four-member chelate rings (4t or monodentatespecies.rl r; We had hoped that this difference in bindingconstants in solution would also be reflected in the strengthof binding of hydroxamic acids to the ions on the surfaceof a metal oxide. In fact, although hydroxamic acids doform more stable monolayers than carboxylic acids, theyseem to use only one oxygen atom for coordination.

Results

Synthesis of Hydroxamic Acids. The long-chain,methyl-terminated hydroxamic acids were synthesizedby reacting the corresponding acid chloride and O-benzylhydroxylamine, followed by removal of the benzylgroup by catalytic hydrogenation over palladium on carbon(Scheme 1).22 The hvdroxvl-terminated hvdroxamic acid

r 18 r Appl icat ions of th io ls on gold and s i lver : Chidsey, C. E. D. Sciencer \ \ ' osh rng ton . D .C l . r 1991 . 251 .9 I9 922 . H ickman, J . J . ; O fe r , D . ;Laib in is. P. E. : \ \ 'h i tes ides. G. ! I . :Wrighton, M. S. Science(Washington.D.( ' . , 1991. 252. 688 691. Pr ime, K. L. ; Whi tesides, G. M. Sciencer l | 'os ht r tgtLtr t . D.C . , 1991. 252. 1164- 1 167. Joyce, S. A. ; Houston, J. E. ;) I i cha lske . T . A . App / . Ph t ' s . Le t t . 1992 ,60 , l I 75 -7177 . Kumar , A . ;Biebuvck. H. A. : Abbott . N. L. ; Whi tesides, G. M. J. Am. Chem. Soc.1992. 11J. 9188 9189. Abbott , N. L. ; Folkers, J. P. ;Whitesides, G. M.Sc ience rV 'ash ing ton , D .C . ) 1992 ,258 ,1380-1382 . Ross , C . B . ; Sun ,L. : Clrooks. R. *-1. Langmuir 1993,9, 632-636. L6pez, G. L. ; Biebuyck,H. A. : Frisbie. C. D. ; Whitesides, G. M. Science ( Washington, D.C.) 1993,260.647 619. Lopez, G. L.; Albers, M. W.; Schreiber, S. L.; Carrol, R.W. :Pera l ta . E . : Wh i tes ides , G . M. J .Am.Chem. Soc . 1993 , 115 ,5877-5878. Kumar. A.; Whitesides, G. M. Appl. Phys. Lett. 1993, 63, 20022004 .

t 19 t Nuzzo, R. G.; Zegarski , B. R. ; Dubois, L. H. J. Am. Chem. Soc.1987. 109, 733-740.

r20r The exceptions being copper oxide, silver oxide, and possiblyother soft metal oxides such as those on platinum, zinc, and lead. Forstudies on the removal the native oxides from copper and silver usingalkanethio ls, see ref 4 and Laib in is, P. E. ; Whi tesides, G.M. J. Am.Chem. Soc. 1992, 114.9022-9028.

t21 i Martell, A. E.; Smith, R.M. Critical Stability Constants; PlenumPress: New York, 1977; Vol. 3. For comparison, the binding constantsfor acetohydroxamate with Fe3* and Al3* in aqueous solution are 2.6x 101r and 8.9 x 107 M r, respectively, and the binding constants foracetate with Fe3* andAl3* are2.4 x 103 and 3.2 x 10t M 1, respectively.

t22 t Hearns , M. T . W. ; Ward , A . D . Aus l . J . Chem.1969 ,22 ,16 l *173. Rowe, J. E. : Ward, A. D. Aust . J . Chem. 1968,21,2761- 2767.Fujii, T.; Hatanaka, Y. Tetrahedron 1973,29,3825-383L. Sun, Y.;Martell, A. E.; Motekaitis, R. J. Inorg. Chem. 1985,24, 4343-4350.

( 23 ) Karunatne, V. ; Hoveyda, H. R. ; Orvig, C. Tetrahedron Lett. L992,,33. 1827- 1830.

(24) Cotton, F. A.; Wilkinson, G. Adua nced Inorganic Chemistry, SthEd.; Wiley-Interscience: New York, 1988.

{25) Entropy of formation (due to chelation) is the dominant factorrn the strong binding between hydroxamic acids and discrete metalions in aqueous solution. The enthalpies of formation for thesecomplexes in water tend to be small and positive. For examples, seeref 2 1 and Monzyk, B.; Crumbliss, A. L. J . Am. Chem. Soc . 1979, 101 ,6203-62L3.

Folkers et al.

Scheme 1. Synthesis of Hydroxamic Acids

cH3(cH2)^coc, - "r*oP

----_ cH'{cH2)^coN ^.p

ooH o { c H 2 } l s c o o H * H 2 N o J

E D C > H o { C H 2 t , " C o N r o J

l-\Y H2,Pd

R(CH2).CONHO -- l ---+ R(CH2).CONHOH

R = C H s , O H

was slmthesized using 1-ethyl-3-(3-(dimethylamino)pro-pyl)carbodiimide (EDC) to couple the benzyl-protectedhydroxylamine to 16-hydroxyhexadecanoic acid.23 Thesame deprotection scheme was used for this compound.

Preparation ofSubstrates andMonolayers. MetaVmetal oxide substrates were prepared by electron-beamevaporation of 600-1000 A of the pure metal onto siliconwafers that had been precoated with 50 A of titanium asan adhesion promoter. The samples were exposed to airupon removal from the chamber. because we wereinterested in studf ing the adsorption of monolayers tothe native oxides of the metals. In this paper, instead ofwriting "metaVmetal oxide" for each reference to asubstrate with a native oxide, we will refer to them simplyas the metal oxide. The oxidation states of the metal inthe native oxides are given in Table 1 and the Experi-mental Section.

Solutions of alkanehydroxamic acids and other adsor-bates were prepared in either isooctane or ethanol andwere not degassed. The nominal concentration in each ofthe solutions was 1 mM, although most of the hydroxamicacids were not soluble in isooctane to this extent.Substrates were left in solution for 1 to 2 days at roomtemperature. After removal from solution, the sampleswere rinsed with heptane and ethanol and then dried ina stream of ni t rogen.

Contact Angles on Monolayers of HydroxamicAcids. Table 1 summarizes advancing and recedingcontact angles of rvater and hexadecane on monolayersobtained by the adsorption of octadecanehydroxamic acid,stearic acid, octadecanephosphonic acid, and 16-hydroxy-hexadecanehydroxamic acid on the six substrates surveyedin this study. We also present contact angles for mono-layers of octadecanethiolate on copper. silver, and goldfor comparison, and contact angles for surfaces obtainedafter attempting adsorptions of the hydroxamic acid ontogold. For consistency, all of the data in Table 1 were takenduringthis study, although several of these systems havebeen studied previously.'2 1;

The presence of a closest-packed, methyl-terminatedmonolayeris often indicated by an advancing contact angleof water (surface tension Tyv

-- 72 mJlm2 at 25 oQ;2ti 2s

above 110o, and an advancingcontact angle ofhexadecane(yr": 27 mJlm2)26 28 above 45".2 6'8'15 Nthough contactangles do not provide any structural information aboutmonolayers, they are useful as indicators of their quality.A low value of contact angle, relative to a structurallywell-characterized system such as alkanethiolates on gold,indicates disorder (and/or lower coverage) in the mono-

O O - ' Mt t - ' \ o F . r ' t

# f -o t *A f .o ,Ao, ,Ad"H H

1 2 3 4

(26 )Jasper , J . J . , J . Phys . Chem. Re f . Da ta 1972 , 1 ,841-1009 .(27 ) Cognard, J. J. Chim. Phys. 1987 , 84, 357 -362, and references

therein.(28) These two probe liquids can provide different information about

the surfaces, because the origins oftheir surface tensions are different.The surface tension of water is dominated by its polar component (;'Lr,.pornr= 50 mJ/m2), and the surface tension of hexadecane originates entirelyfrom disnersive interactions.2T

SAMs of Hydroxamic Acids Langrnuir, Vol. 17, No. 3, 1995 815

Table 1, Advancing and Recedidg Contact Angles of Water and Ileradecane (HD) on Monolayers of

Octadecanehydroxamic Acid Stearic Acid, Octadecanephosphonic Acid, l6-Hydroryhexadecanehydroramic Acid, and

Octadecanethiol on Copper, Silver, Titaniun, Zirconium, Iron, Aluninum, and Gold"

contact angles on various SAMs (advancing,receding)/deg

isooctaneb ethanolh

metal adsorbate Hzo HD HzO HD

Cr-r/Cu(II)

Ag/Ag{II)

Ti/Ti(IV)

AVAI(III)

ZrlZriYl

Fe,FerII I t

Au

cHs(cHz)rocoNHoHCHr(CHz)r6COOHCHr(CHz)r ;POsHzCH;r(CHz )r;SHHO(CHz)TsCONHOHCHs(CHzhsCONHOHCHe(CHz)raCOOHCHs(CHz)rzPOsHzCHs(CHz)rrSHHO(CHz)TsCONHOHcHe(cHz)r6coNHoHCH:(CHz)raCOOHCH:r(CHz)rzPOgHzHO(CHz)TsCONHOHCHs(CHz)ToCONHOHCH:(CHz)raCOOHCHr(CHz)r;POsHzHO(CHz)TsCONHOHCH3(CH2 )T6CONHOHCH r t CH: ) r oCOOHCH3t f 11, r1 ,PO, IH:HO,CH: r ;CONHOHCH, r {CH: r r e CONHOHCHsr f [ 1 , r r . ;COOHCH:t r CH: ,1 ;POrtH:HOtCH2 )T5CONHOHCH3tQ11, )T6CONHOHCH;t CHz ) r ;SH

1 1 4 . 1 0 41 1 3 , 1 0 71 1 5 , 9 8

79 ,40113 , 1041 1 3 , 1 0 7115 , 103

82, 501 1 5 , 1 0 5

107 ,71115, 100

83, 60113 , 104113 , 1051 1 5 , 1 0 4

49 .29123 , 99

r 2 2 . 1 0 01 2 3 . 1 0 0

1 j ] 5 . 9 5

1;16. S-1l 3 ; . 9 ;;9. jl-l

106 . 87

4 7 . 1 r49. .1,143 . 35

0 . 051 , 4652,4746,2:

0 , 046.4245 ,374 2 , 3 6

0 , 0A - , t , 1. 1 , . a r

1 8 . 1 1,18. j lE

o . i )J J . ) ;

r : ' ) Q

: l l . l r r

l : i . 1 ;l i . t t

2 1 . r ,( ) . o

2 2 . t r

11 .1 . 1011 1 1 . 8 21 1 2 . 7 61 1 7 . 9 3

' ) 7 1 ' )

1 1 3 . 1 0 28 1 . 6 182. 60

1 1 3 , 1 0 268, 39

1 1 0 , 8 550, 15

1 1 0 , 9 577 , 5 r

1 1 1 , 9 31 1 0 , 8 21 1 3 , 8 37 2 , 3 6

123, 951 1 6 . 7 51 1 9 . 7 5i r ( ) . 2 0

1 r l 1 . 9 01 1 1 . 6 512-1. t r )

8-1. 63115 , 103

49 ,4 r46, 38

29 ,062,282 2 , 0

5r ,4421,00 , 0

49,430 , 0

45 ,290 , 0

41 , 360 , 0

46,4043, 35

37 , 00 , 0

43.2223 ,02 3 . 00 . 0

3 1 , 1 623 . 020 . 0

0 . 049. 39

n Metals were exposed to air and covered witl native oxides before exposure to solutions containing the ligands. t Solutions from which

monolavers wete formed. " Denotes that monolavers were not made from these solutions.

Table 2. Decrease in Advancing and Receding Contaet

Angles of Water (AOHro) and Hexadecane (AOHD) on

Changing Solvent from Isooctane to Ethanol for

Adsorptions of Octadecanehydroxamic Acid, Stearic

Acid, and Octadecanephosphonic Acid (adv'rec)a

i C H 3 t ( ' H 2 1 1 , ; C O N H O H (CH3(CH2h6COOH (CH3(CHz)uPOsHz

m e t a l h . \ ( / i l r ( ' \Fi t I I) a0H2o aoHD A{/H:o AOHI

layer; a value comparable to that of the ordered systemdoes not necessarily indicate the same structure or acorresponding degfee of order. Using contact angles, wecan infer the quality of one type of SAM relative to otherswith analogous structures on the same metal, but notrelative to SAMs on different metals, because the rough-ness of the surfaces of the various metals differ.2e'3o Toreduce the variation in roughness of structures for aparticular metal, all of the samples for each metal listedin Table 1 were taken from the same evaporation; theonly exceptions were the monolayer of HO(CHz)ro-CONHOH on silver oxide from ethanol and the two setsof monolayers on iron oxide. We infer from Table 1 thatthe quality ofmethyl-terminated SAMs obtained from eachof the three acids formed from isooctane is greater thanthat of SAMs formed from ethanol, since they have higheradvancing contact angles and lower values ofhysteresis.30

Table 2 summarizes the decrease in contact angles onchanging the solvent for adsorption from isooctane toethanol. We use these values as qualitative indicators ofthe stability of the monolayers by assuming that ethanolcompetes for polar sites on the surface but that isooctane

(29) The substrates of zirconium/zirconium oxide and iron/iron oxide,as we had prepared them, were much rougher than the other metals:advancing contact angles and values ofhysteresis in the contact anglesof water were consistently higher for monolayers on these substrates.30We have not investigated these substrates with the scanning probemicroscope, and we are not able to quantitate their relative roughness.

(30) Hysteresis in contact angles, although poorly understood, is oftenattributed to the chemical heterogeneity or the roughness ofan interface.Since we have attempted to control the relative roughnesses of oursubstrates for a given metal, we interpret the increase in hysteresis asan increase in the chemical heterogeneity of the surfaces. For examplesofexperimental and theoretical studies on hysteresis, see: Joanny, J.F. ; de Gennes, P. G. J. Chem. Phys.1984,81,552-562. de Gennes, P.G. Reu . Mod. Phys. 1985, 57, 827 - 863,and references therein. Schwartz,L. W.; Garoff, S. Langmuir, 1985, 1.219-230. Chen, Y. L.; Helm, C.A.; Israelachvili, J. N. J. Phy.s. Chem. 199L, 95, 10736-1'0747.

C u_ _ h

f iAIZr

- 2 . ' 2 . 2 , 2 5 3 , 6 3 , 2 2 1 4 , ' 3 5 ' i' 2 . . 2 32 .46 37 , > -47 33 , 43 > -46 , >28. 2 . 13 57 . 56 > -45 . >_37 5 , 5 <2 , <2. ' 2 . 1 3 . 2 3 5 , 9 < 2 , 2 1 1 1 , > 3 8' ' 2 . 5 6 .25 24 , 229 4 ,25 14 , : : 20

" Contact angles rvere taken from Table 1. AO is the difference

in contact angie between those monolayers prepared from isooctane

and those prepared from ethanol r Al/ : g(isooctane) - 0(ethanol)).b Iron is not included because the substrates examined were not

prepared in the same evaporation. ( " <2" is used to indicate that

changes in their contact angles of 2o or less are within the

uncertainty of the measurements. d ">" indicates that the contact

angle of the monolayer formed with ethanol was zero, and the

increase in surface free energy could be greater than that measured

by the decrease in the contact angle.

does not. In general, differences in the contact angleswere lowest for monolayers ofthe hydroxamic acid. Thesedata thus suggest that monolayers from hydroxamic acidsare more stable than monolayers of either carboxylic acidsor phosphonic acids on all of the substrates examinedexcept titanium oxide. The smallest changes in contactangles were observed for monolayers of hydroxamic acidson copper oxide and silver oxide, suggesting a high stabiiityfor the hydroxamic acid overlayer formed on lhesesubstrates. By this criterion, of the three acids, thecarboxylic acid forms the least stable monolayers on allof the substrates. Although phosphonic acids are oftenused for the formation of self-assembled monolayers onzirconium oxide, the results in Tables 1 and 2 suggest

. 2 . i l- 2 . ' 2

) . 2 0. 2 . 1 I. \ ) 1

8i6 Langmuir , Vol . 11, No. 3, 1995

that hydroxamic acids form monolayers of even higherstability on this substrate: this inference remains to betested by direct competition. Fortitanium oxide, we inferthat the phosphonic acid forms the most stable monolayers.

Hvdroxamic acids do not form closest-packed mono-lavers on gold, although the measured contact anglesindicate some partial layer is adsorbed from isooctane.The contact angles were only slightly lower than thatexpected for a methyl-terminated monolayer, but resultsfrom X-ray photoelectron spectroscopy (XPS) showed thatthe "monolayer" was approximately 11 A thinner than amonolayer ofoctadecanethiolate on gold.3l The adsorptionof hydroxamic acids to gold presumably occurs by phys-isorption on the high-energy gold surface,32 as gold doesnot form a stable oxide under ambient conditions; we havenot determined the structure of this layer.

Adsorption of alkanethiols on copper modifies its surfacemore than does the adsorption of any ofthe acid adsorbatesused in this study, as judged by the larger values of thecontact angle hysteresis measured. Since the compositionsof the interfaces present in the two monolayer classes aresimilar, we attribute the differences in hysteresis tovariations in the microscopic roughness ofthe interface.;r0This roughening is presumably a direct consequence ofthe removal of the native oxide on copper that occursduring the adsorption of the alkanethiol, as has beenshown byXPS.{ 2( The native oxide is not removed duringthe adsorption of the three acid adsorbates.33 Theadsorption of alkanethiols on the native oxide of silverdoes not roughen the surface in the same way, even thoughthe thin native silver oxide layer is lost during theadsorption.4'34

The hydroxyl-terminated hydroxamic acid formed in-complete monolayers on most of the substrates studiedhere, with the single exception of the monolayer formed oncopper oxide by adsorption from ethanol. We expected tomeasure an advancing contact angle of water below 30"for a complete monolayer of a hydroxy-terminatedligan6''r* 'r5 the observed value is gu r 27' (Table 1). Mostsubstrates gave advancing contact angles ofwater between50" and 80o, suggesting the existence of nonpolar groupsat the interface and possibly some "looping" ofthe hydroxylgroup to the surface. Similar loop structures have beenobserved by Allara et al. with a long-chain cr,a;-dicarboxylicacid adsorbate on several oxide surfaces.ll X-ray photo-electron spectroscopy showed that the mass coverages ofthe hydroxyl-terminated hydroxamic acid monolayerswere 6- 10 A less than those of monolayers of CH:(CHrlru-CONHOH (see below).

When a monolayer of the hydroxyl-terminated hydrox-amic acid was formed on copper oxide by adsorption fromethanol, the advancing contact angle of water indicated

t31 r A "thinner" or "shorter" monolayer as determined by XPS impliesincomplete formation of the monolayer. XPS is used to measure theattenuation of a signal from a substrate as a function of the amount ofh.l 'drocarbon on that substrate; the thickness of the monolayer is thendetermined after making several assumptions about this piocess (seee q s 1 - 3 t .

t32 t Weak physisorption of amphiphilic molecules such as dodecanolfrom solution onto gold has been previously observed; see: Yeo, Y. H.;NlcGonigal, G. C.; Yackoboski, K.; Guo, C. X.; Thomson, D. J. J. Phys.Chen t . f 992 , 96 , 6110 6111 .

t33 t The hydroxamic acids probably to dissolve some copper lonsfrom the oxide: when SAMs of hydroxamic acids were formed on silveroxide from ethanolic solutions that had been used previously for copperoxide, copper was evident in the XPS, but samples on silver made fromfresh solutions showed no copper (Folkers, J. P.; Whitesides, G. M.Unpubi ished resul ts) .

r34t Laib in is, P. E. ; Whi tesides, G. M. J. Am. Chem. Soc. lgg2. 1 14.1990 1995 .

t35t Bain, C. D. ; Whi tesides, G. M. J. Am. Chem. Soc. 1988. 110.6560 6561. Bain, C. D. ; Eval l , J . ; Whi tesides, G. M. J.Am. Chem. Soc.1 9 8 9 . 1 1 1 . 7 1 5 5 - 7 \ 6 4 .

Folkers et al.

the formation of a well-defined hydrophilic structure. Theadvancing contact angle of hexadecane was nonzero; thisvalue is consistent with the formation of a hydroxyl-terminated monolayer, on which a thin, adsorbed layer ofwater prevents the spreading of the contacting hydro-carbon liquid.:r6 The measured thickness ofthis monolayerwas intermediate between that of films formed byCHg(CHz)IaCONHOH and CHs(CHz)ToCONHOH (see Fig-ure 3 and the accompanying text below), suggesting thatthe adsorption proceeds to high mass coverages in anisostructural manner. Solvent plays a very importantrole in this assembly, however. We found, for example,that the hydroxyl-terminated hydroxamic acid forms anincomplete monolayer on copper oxide when adsorbed fromisooctane (the coverage was less than 70Vc).

Contact Angles on Monolayers of HydroxamicAcids as a Function of Chain Length. Figure 1presents contact angles ofwater and hexadecane on SAMsof hydroxamic acids with the general formula CHr-(CHz),CONHOH on the native oxides of copper, silver,titanium, and aluminum as a function of n (the numberof methylene groups ). Monolal, 'ers on copper oxide andtitanium oxide were formed br- adsorption from bothisooctane and ethanol : monolavers on si lver oxide andaluminum oxrde were formed onlv using isooctane as asolvent.

The contact angles measured on copper oxide wereindependent of n and ofthe solvent used for the adsorption.These results suggestthat monolayers ofhydroxamic acidson copper oxide have high stability. The trend in the datawas similar for the monolayers on silver oxide preparedby adsorption from isooctane. On titanium oxide andaluminum oxide, little variation was observed in the chainlength dependence of the contact angles for SAMs formedby adsorption from isooctane for n > 10; below this chainlength, the contact angles decreased and the hysteresisincreased markedly. Binding of alkanehydroxamic acidsto the surfaces of aluminum oxide and iron oxide thusappears to be u'eak. as judged by the ability of ethanol (orimpurit ies such as rvater)to interfere with the formationof high coverage monolavers.

The contact angles for the monolayer with n : 15 onsilver oxide were significantlr ' lower than those of the twoeven-carbon chain lengths adjacent to it. The observeddecrease is not due to an incompletell ' formed monolayer(as confirmed by XPS r. but seems to be due to a changein the structure of the methl'l gloups at the interface (seebelow). This result is similar to the "odd-even" effectsin contact angles which have been observed previouslyfor monolayers of carboxt'lic acidslt and alkanethiolateson silver.3 We note, however. that the structural originsin the other examples may be very different from thosewhich peftain here.

Formation of Mixed Monolayrs on Copper Oxide.We formed mixed monolayers by the competitive adsorp-tion of HO(CHz)r sCONHOH and CHs(CH2)r ICONHOH oncopper oxide (Figure 2). These data demonstrate thatthese mixed monolayers can be used to tailor the wetta-bility (and, by extrapolation, other properties) of thesesurfaces and illustrate the utility of hydroxamic acids informing complex organic surfaces on copper oxide withoutsignificantly roughening the surface. These materialsthus may have a significant advantage for the modificationofthe surfaces over the previously described alkanethiolateSAMs.

Relative Thicknesses of Monolayers of Hydrox-amic Acids from X-ray Photoelectron Spectroscopy

(36) Adamson, A. Physical Chemistry of' Surfa<,t:s, 4th ed.: Wilev:New York, 1982; Chapter 3.

SAMs of Hydroxamic Acids

lsooctane: .6f,o o e:t,o r elo D elDEthanol : .e l t 'o oel ,o

"el 'o *0lo

0.2

-cos 0 -0.2

120

0.2

-cos 0 -0.2

-0.6

-1.0

8 1 0 1 2 1 4 1 6n in CH3(CH2)nCONHOH

Figure l. Advancing (6,,) and receding (9,) contact angles ofwater and hexadecane (HD) on monolayers derived from theadsorption of hydroxamic acids of the general formula CH3-tCH2 I ,CONHOH. where h:8 ,L0,12,14, 15, 16, onto the nat iveoxides of copper. si lver, t i tanium, and aluminum. Monolayerson copper and titanium were formed from isooctane and ethanol;monolayers on silver and aluminum were formed from isooctane.As a guide to the eye, the dashed lines in the plot for copperrepresent the average values ofthe contact angles for all valuesof z and both solvents: 118" for 0 ̂ Hro , 102' for O,Hro , 48' for6oHD, and 35" for g.HD. For the monolayers on silver, the contactangles for n:15 were not used to calculate the average contactangles represented by the dashed lines; these values are ll2"for guH:o, 104'for 0,Hro,51'for O'HD, and 46" for 0,HD.

(XPS). We have used the attenuation of an XPS signaldue to various substrate core levels (specific details aregiven in the Experimental Section) to estimate relativethicknesses of the SAMs formed by the adsorption ofhydroxamic acids with different chain lengths.a'34.37 3e Inthis analysis, the intensity, Isana, of an XPS signal from

Langmuir , Vol . 11, No. 3, 1995 817

s37 532 527Binding Energy in ev

-OH 0.2 0.4 0 '6 0.8 CHg

X.rr,.o"Figure 2. Contact angles of water on monolayers formed bythe compet i t rve: rdsorpt ion o f CH:(CHz)r+CONHOH and HO-tCHz TT.,CONHOH tinto copper, plotted against the composit ionof the SAll presented as the mole fract ion of the methyl-terminate-d component rn the SAII t /( I t s.\rr t . The composit ionof the SA\l r i ' rrs dett ' rmined br.XPS using the oxygen 1s signalfor the te t 'n t inu l hr -c l r ' , rx r l gr 'oup nn HOTCHTI ;CONHOH.\lonolavers \\ ' r ' l ' r , t i rrnrecl ir t room temperature from ethanolicso lu t rons u ' r th i i to tu l concent ra t ion o f hydroxamic ac id o f 1m}l fr ir 5 duvs. Adr-ancing contact angles are shown as f i l ledcrrcles: recedrng contact angles are shown as open circles. Linesthrough the data are provided as guides to the eye. The errorbars represent the approximate accuracy of acquiringand fittingthe XPS data. The error in the measurement of the contactangles is smaller than the size of the circles. Above the plot,XPS spectra are presented for the oxygen ls region of mono-layers of HO( CHz ) l5CONHOH ( left ) and CHs( CHz )14CONHOH(right) on copper. The peak at lower binding energy (-530 eV)is due to the oxide on the copper and the oxygens in theheadgroup of the molecules. The peak at higher binding energy(-532.5 eVt is due to the terminal OH.

a substrate covered with an alkyl-chain-terminated SAMis given b1' eq 1

1r,\\r : 1,/ exp( -mdll sin 0) (1)

where 1,, is the intensity of the signal without themonolayer.l/ scales the intensity for attenuation throughthe head group of the ligand, m is the number of carbonatoms in the aikyl chain (.m : n * l), d is the thicknessfor a methylene group, z, is the attenuation length (orescape depth) of the photoelectron through the hydro-carbon layer, and I is the angle between the plane of thesurface and the detector (35o).3? 10 To obtain the value ofd for a set of monolayers, we compare the intensity of theXPS signal from a SAM to that from the shortest chainlength SAIV[3e''11

t 37 t Ba in , C . D . :Wh i tes ides , G . M. J . Phys . Chem.1989 ,93 , 1670-1673 .

(38tLaib in is, P. E. ; Bain, C. D. ; Whi tesides. G. M. J. Phys. Chem.1 9 9 1 . . 9 5 , 7 0 r 7 7 0 2 r .

(39)Laibin is, P. E. ; Fox, M. A. ; Folkers, J. P. ; Whi tesides, G. M.Langmuir 1991. 47. 3167-3173. Folkers, J. P. ; Laib in is, P. E. :Whitesides, G. M. Langmuir 1992,8, f330-1341.

(40) Seah, M. P. In Practical Surface Analysis by Auger and X-rayPhotoelectron Spectroscopy; Briggs, D., Seah, M. P., Eds.;John Wiley:Chichester , 1983; Chapter 5.

r4 l rCa lcu la t i on o f abso lu te th i cknesses us ing th i s me thod requ i resdetermination of10 and /. The primary reason we have not determinedthese quantities is that cleaning metaVmetal oxide substrates bysputtering the surfaces with Ar* ions in the XPS chamber often removespart of the oxide and changes the measured value of1s.

120

120

300

300

300

300

300

120

o;- -t- -c - -O o. C-

f - B - 4 - € 0 $ -

S- -F- { - - r. s-- -E - + - € + =

._ _o_ _. _ : _ :_

cr --- -J - 't, F \-

lAslagol

l - - I - - l - - l - -t_- E I-

L-F

n

o

o

o

T

a

a-

o

I!

a o t

o o g

^ o

m,o;lT I NN D E

a o a g o ao o o o Q

lAyArF;l

! : !

818 Langmuir , Vol . 11, No. 3, 1995

(m - m36)d - -A sin 0ln(I-l/*"n)

In eq 2, nt36 is the number of carbon atoms in the chainof the shortest molecule. We use an empirical equationto estimate the attenuation lengths (A) ofthe photoejectedelectrons through the hydrocarbon layer (eq 3)

A : 0.0228k + 9.0 (3)

where Ep is the kinetic energy (eV) of the photoelectrons;the validity of this equation was established for SAMs ofalkanethiolates on copper, silver, and gold.38

Results for monolayers ofhydroxamic acids on the nativeoxides of copper, silver, and titanium are summarized inFigure 3. On copper oxide, a small decrease in thicknesswas observed for monolayers formed by adsorption fromethanol on to copper oxide relative to those formed usingisooctane. Both sets of data yield the same value for theaverage increase in thickness per methylene group, 1.1A/CH2. The value expected for a trans-extended alkylchain oriented perpendicular to the surface is 1.36 NCH2.From this value, we estimate that the average tilt of thechains in a monolayer of hydroxamic acids on copper oxideis approximately 20o, in agreement with results frominfrared spectroscopy (see below).

On monolayers of hydroxamic acids on silver oxideformed by adsorption from isooctane, we calculate a valueof d of 1.2 AJCHz from the data. This value suggests acant angle for the alkyl chain of about 18" from the surfacenormal; the uncertainty in this value is large (*15" or+0.1 A/CH2 for d), however, because ofthe limited numberof points measured.

On titanium oxide, we observed differences in thethicknesses of monolayers formed by adsorption fromisooctane or compared to those formed using ethanol asthe solvent;32 the data suggest lower coverages for theSAMs prepared using ethanol. These data are poorly fitby our analytical model and, thus, our ability to estimatea value of d is limited. For the isooctane samples, thedata are linear (two separate sets of data for isooctane areincluded in Figure 3t, but our analysis leads to animplausible estimated average increase in thickness permethylene group of 1.4 NCH2. Several experimentalissues. especially the roughness of substrate, can lead toa large error in the calculated value of the attenuationlength. Variations in the thickness of the native oxideamong sampies also are believed to contribute to thisdiscrepancy. Similar systematic errors for SAMs onaluminum, iron, and zirconium oxides precluded mea-surement of values of d on these substrates.

Deprotonation of the Hydroxamic Acids on theVarious Oxides from )(PS. In the XPS data formonolayers ofhydroxamic acids, the N ls core level spectraprouded the most information on the nature of bondingthat occurs between the hydroxamic acid and the varioussurfaces. Figure 4 presents N 1s spectra for octadecanehydroxamic acid as a solid and for monolayers adsorbedonto the native oxides of copper, silver, aluminum, andtitanium. We calibrated the binding energies of thesespectra by frxing the C 1s binding enerry of the mainGaussian component (due to the CH2 groups) at 284.6eV.12'13 The N 1s spectrum of the solid hydroxamic acidcould be frt with a single peak centered at 400.7 eV. Thespectra of the monolayers showed multiple componentsin the N 1s region: one centered at -400.5 eV (inagreement with the peak for the solid hydroxamic acid)

(42iMuilenberg, G. 8., Ed., Handbook of X-ray PhotoeLectronSpectroscopy; Perkin Elmer Corp.: Eden Prairie, Mn, 1979. Wagner,C. D.: Gale. L. H.; Raymond. R. H. Anal. Chem. 1979, 51, 466-482.

Thicknessfrom XPSrelal ive toshortestacid (A)

8

A

4

2

Ag/AgO

o lsooctane

d = 1 .2 A /CH2

Thicknessf rom XPSrelat ive toshortestacid (A)

Figure 3. Relative thickness of monolayers formed fromadsorption of hydroxamic acids of the general formulaCHr(CHz),CONHOH and HOTCH: )r;CONHOH onto the nativeoxides of copper. si lver. and t i tanium. Al l thicknesses weredetermined re la t iye to the monolayer o f CH,r (CH:)gCONHOHfrom isooctane using the attentuation of the XPS signals fromthe subst rAtes { se( ' t he tex t {br deta i ls i . F i l led r O t and open (O)circles represent nrethr ' l - tet-minated monolavers formed fromisooctane and ethanol. respectivel l ' . The thicknesses of mono-layers obtained b1'the acisorption of HOtCHztr;CONHOH fromisooctane (O ) and from ethanol tC t are included at the positionof 15 methylene units in the alkyl chain. Lines through thedata were determined by least-squares fits to the data andwere determined without the data for the hydroxyl-terminatedmonolayers.

and a second at -398.9 eV.aa The relative intensities ofthese peaks varied with the substrate used (as shown inFigure 4).aa From these data, we conciude that on copperoxide and silver oxide the predominant species bound tothe surface is the monodeprotonated hydroxamate, while

on aluminum oxide and titanium oxide, the predominant

species bound to the surface is the hydroxamic acid.

Data from the C 1s region also suggest varying extentsof deprotonation in the different monolayers (Figure 5).45The peak due to the carbonyl occurs at 287.6 eV for thesolid hydroxarnic acid. In the monolayer, this peak isshifted to287 .0 and 286.5 eV on titanium oxide and copper

(43) Blements in monolayers on thick layers of oxide, such a-s thaton aluminum, have higher binding energ'les than those on thin layersof oxide due to the inhomogeneous charging that occurs on the thickeroxide substrate. To correct for these dif 'ferences. u'e have referencedthe binding energies of elements in the monolayer relati ve to C 1 s bindingenergy of the hydrocarbon at 284.6 eV.a:

G4)The measured binding energies of the peaks in the N ls regionfor the monolayer on silver are sigr-ri{icantlv lt iu'er than those for theother substrates (400.1 and 398.6 eVr These peaks are super imposedon the high-energy background following the Ag 3d peaks, which mightcomplicate accurate determination of the binding energies. The noisein this spectrum limited our abilitv to determine an accurate ratio ofacid to monoanion, although the monoanion is clearly favored.

(45) We used monola-n-ers of CHrrr CH:)ICONHOH to improve oursensitinty by decreasingthe attenuation of the signal due tcl the carbonylby the alkyl chain.

Folkers et al.

Q )

Thicknessfrom XPS 6relative toshortest 4acid (A)

2

-2

1 2

1 0

10 1 '2 tc oH

n in HONHOC(CH2)nCH3

Metal I I Ratio

Cu i l I 1 :8

A g t J 1 . 1 : 1 . 7

SAMs of Hydroxamic Acids

405 403 401 399 397 395Binding Energy in eV

Figure 4. X-ray photoelectron spectra forthe nitrogen 1s regionofsolid CHe(CHz)TaCONHOH (bottom) and monolayers obtainedby the adsorption ofit onto the native oxides ofcopper, silver,aluminum, and titanium. Binding energies of the peaks arereferenced to the carbon 1s peak for the alkyl chain at 284.6eV. Relative intensities could not be compared directly as theywere not referenced to a standard. "Ratio" refers to the relativeintensities of the peaks ascribed to -CONHOH and -CONHO.

oxide, respectively. The observed line width is verv broadfor the titanium oxide supported monolayers and less sofor the monolayers on copper oxide. We take this asevidence of a heterogeneous environment present at theheadgroup on titanium oxide and, based on the cleardistinctions observed in the N 1s region, attribute thisline width to a heterogeneous state of protonation.

Orientation and Order in SAMs of HydroxamicAcids from Polarized Infrared External Refl ectanceSpectroscopy (PIERS).'16'47 Inspection of the PIERSdata (Figure 6, Table 3) reveals that the monola;'ers formedby hydroxamic acids on many of the substrates examinedin this study (i.e. those on copper, silver, aluminum, andtitanium) are characterized by highl) organized struc-tures. Quantitative evaluation, however, indicates thateach differs in the orientation adopted by the chain. Themode assignments for the C-H stretching region followfrom a similar analysis as that described for alkanethiolateSAMs on golda and alkanecarboxylic acid SAMs onaluminum oxide.15 In the discussion that follows, weanalyze each system briefly in terms of an averagestructure defined by the model shown in Scheme 2. Thismodel assumes that a best fit to the data involves a specificaverage orientation of an all-trans chain. Such a structureis described uniquely by the two angles, a and p, shownin Scheme 2. The discussions that follow are qualitative

(46) Greenler . R. G. J. Chem.(47 ) Snyder , R .G. ;S t rauss , H .

86 . 5145-5150 .

Phys. 1966, 44, 370-315.L.; Ell iger, C. A. J. Phys. Chem. 1982,

Longmuir, Vol. i1, No. 3, 1995 819

2s1 "r:".,::tr";;

" l'

Figure 5. X-rav phritoeiectrc-rn spectra for the carbon 1s regionof monolavers derived from the adsorption of CH,r(CH2)s-CONHOH onto copper oxide and titanium oxide, and of solidCHr(CHz)TcCONHOH. Binding energies of the peaks arereferenced to the carbon ls peak for the alkyl chain at 284.6eV. The dashed line through the spectra allows comparison ofthe binding energies of the peaks due to the carbonyl. Relativeintensities could not be compared directly as they were notreferenced to a standard.

in that we have not performed specific calculations foreach of these SAMs but. rather, used the well-establishedstructures and optical functions of the alkanethiolatemonolavers on copper, silver, and gold as benchmarks forcomparison. The C-H stretching region in the infraredspectra of hydroxamic acid SAMs on copper oxide dem-onstrates an intriguing similarity to those of alkanethi-olates on gold. Both the relative and absolute intensitiesof the methyl and methylene modes suggest that the alkylchains are canted from the surface normal direction. Thisstructure involves a chain tilt, a, of -20o (in agreementwith the XPS results presented above) and a rotation ofthe plane of the trans-zigzag,0, of -50-55". Significantdifferences in structure are also evident between the twosystems. The frequency of the antisymmetric methylenestretch for the hydroxamic acid SAM on copper oxide ishigher than that found for alkanethiolate SAMs on gold(2921cm I versus 2919 cm 1). This observation suggeststhat the population of gauche conformations in the alkylchains must be relatively high in this system. If so, thevalues of cr and B are at best approximate given that themodel is based on an all-trans conformation of the chain.Still, the presence of a canted chain structure is stronglysuggested by the data. The hydroxamic acid SAMs oncopper oxide differ from the alkanthiolate SAMs on goldin another important respect. The data shown in Figure6 show very little, if any, sensitivity to chain length asevidenced by alternations in the intensities of the methylsymmetric (r-) and in-plane antisymmetric (r, ) stretchingmodels. Alkanethiolate SAMs on gold show pronouncedodd-even chain length progressions in the intensities ofthese modes, a feature consistent with a preferred absolute

o

.i

G

6

oo

o

jL

tE

o

5oo

H O N H O C ( C H z ) s C H g

l l / l l U 2 +

H O N H O C ( C H z ) e C H g

( cH2) r5cH

820 Longn tu t r , \ ' o l . 11 , No .3 , 1995 Fr , l ke rs e t a l .

r@Ttgo.oo1 r lo.oor

- -

o ' , n , \ .A - - r - / \ / \ l t

( J , l / \ i V l , a r y '

F -^-nJ'' ! \,.- *-'i '

€ i i r I \o J l " l \

{ ^^/tl J\t'

2900 3000 2800

Wavenumbers (cm-1)

z lIL---}

x

Scheme 2

z

HONHOC(CH2)"CH3

n

.-----l-rTr T__-.2900 3000 2800 2900 3000

Wavenumbers (cm'1)

Figure 6. Polar ized infrared external ref lectance spectra l i ) r 'the C-H stretching region of monolaS'ers of octadecant, - .heptadecane-, and hexadecanehydroxamic acid on copper. siiver'.t itanium, and aluminum. Scale bars are presented in thefigures for comparison.

Table 3. Frequencies of Peaks in the Infrared Spectrafor Monolayers of Octadecanehydroxamic Acid on

Copper, Silver, Titanium, and Aluminum,and Octadecanethiolate (RS ) on Gold

(all Frequencies in cm r)

mode description" Cu Ag Ti Al RS /Au'

CH2, sym id. ) 2851CH3. sym u ' - ) 2880CHz, asvm rd ) 2920CH3, sym tFRC tr ' . '2938

CH;l. ?SVITI l l ' , , )" 2967

MOx (.x

1 2 0 ' l- 1 4

I 8 ' l

at the surface is indeed fluxional, however. The densityof gauche confbrmations. which we expect to segregate tothe chain termina,t 1; is sufficiently high in this systemto prevent us from reaching a definitive conclusion onthis point. For example, relaxation of the chain ends viathe gauche-trans isomerism might lead to similar surfacestructures (e.g. equivalent average orientations of themethyl groups)in an odd-even chain-length series whichform a series of SAMs characterized by a single, absolutevalue of a.

Structural arguments for hydroxamic acid SAMs onsrlver oxide are simpler. The mode assignments for theseSAfIs fbllow similariy from those described above. Thelori ' f i 'eque,no' of the d mode Q9IT cm 1) suggests astructure n'rth a f 'eri ' gauche conformations. The relativeand absoiute intensrtres of the d- and d modes alsosuggest ;, i unique structure fbr this SAM: the averageva lues o f ' the charn r ) l ' ren t i r t ion parameters , a and p , andunlike those found in unr- class of SAII studied to date.Using the models described above. we estimate that a ry-I4" and p ry 38" in this system. The odd-even progres-sion of intensities of the methyl r,,, and l is pronounced,providing definitive evidence that the bonding geometrypresent at the headgroups is conserved. The single signof cr and the absence of signif,rcant surface reconstructionsdue to the population of gauche conformations must resultin a chain length dependence of the orientations of themethyl groups that terminate the chains. It is importantto note that hydroxamic acid SAMS on both copper oxideand silver oxide are canted chain structures and that thegauche conformation densities must depend on more thanjust the presence ofthis structural feature. It seems likelythat surface roughness (which we expect to be significantlyhigher for the copper oxide surface) must contributesignificantly as well. It is particularly intriguing thatthe wetting data for the hydroxamic acid SAMs on silveroxide (Figure 1t also show odd-even progressions. Thisobservation provides one ofthe best documented examplesof this structural influence on wetting to date.

The hydroxamic acid SAMs on aluminum oxide exhibitstructural features that appear to be intermediate betweenthose of SAMs on silver oxide and on copper oxide. Thedensity of gauche populations is likely to be large (d at-2922 cm 1). Although the average chain tilt is small,we estimate an approximate structure with lal ry 8" andfi = 47". This description of the chain structure is, atbest, approximate because of the complex chain confor-

Cu

Ag

A I

T i

54'-'

3B'

47'

1 6

t 5

oocG I-oooo

TE|ooo, /1| /1 .4 / \

*?v "L..-il

^^J\td\'^*il

*_/A^\^* 1 4

2851 28502879 28822917 29t6

-2937 2953'12965 297r

2851 28512879 28782922 2919

-2938 -29362967 2964

" Assignments of peaks from ref 4. b FRC : Fermi resonancesplitt ing component.l ' For consistency, this spectrum was takenduring the present study. '1 The assignment ofthis peak is uncertain;it may'correspond to an out-of-plane antisymmetric meth5rl stretch.No peak is visible for the expected Fermi resonance peak at -2935

cm 1. ' ' The frequencies listed are for the in-plane antisymmetricstretch; the frequencies of the out-of-plane mode (r,6 ) are uncertaintfor most SAMs except as noted in d) given the weak appearanceof this band as an unresolved shoulder on the r" band at lowerfrequencies.

sign of the tilt angle, cr. The magnitude of the intensityprogressions is "softened", however, by the presence ofgauche conformers at the chain termina. In this respect,then, the hydroxamic acid SAMs on copper oxide appearto differ significantly from alkanethiolates on gold. Onemight assume, based on the model defined above, thatthe absolute sign ofthe chain cant angle, G, must alternatein an odd-even series. There exists great uncertainty asto whether the head group bonding of the hydroxamate

oy

I.*'

SAMs of Hvdroxamic Acids

-CHs

0,8

0.6

0.4

0.2

-oH

0 0,01 0.1 1 10 100 -

a [CH3(CH2)lo 'HG']I rsorn - txb (cnr)rucoruttoxl

Figure 7. Competitive adsorptions of HO(CHz)TsCONHOHagainst CH3( CHzhsSH (open circles), against CHs(CHz)r4COOH(open squares), and against CHa(CHzh+CONHOH (filled circles)onto copper from ethanol. Ail mixed monolayers were formedfrom ethanolic solutions containing a total concentration ofadsorbate of 1 mM at room temperature for 1 day. Surfacecomposition was determined by relating the measured advanc-ing contact angle of water to the surface compositions deter-mined in Figure 2 (see text). Lines through the data are presentas guides to the eye.

mations noted above. It initially seems contradictory thatthe chains are relatively uncanted yet still contain a largegauche population. Whether surface roughness or someother defect structure is responsible for this combinationof features is unclear at present. Nevertheless. the datacompel a model of a highly oriented charn. albeit one witha complex conformational architecture.

We cannot estimate the orientation of the alkyl chainsin the monolayers of hydroxamic acids on titanium oxide,because the modes seen in the C-H stretching region areexceedingly complex and are superimposed on a nonlinearsubstrate background (by necessity, a gold mirror wasused as a reference). The organization of the alkyl chainsin these SAMs is unclear to us. One might assume thatthe alkyl chains are predominantly in a trans zigzagarrangement, given that the stretching frequency of thed- band occurs at 29t6 cm-l. The broadness andcomplexity of the lineshapes, however, strongly suggestthat these structures are disordered.

Examination of the spectra of the various SAMs in thelow-frequency region below 2000 cm 1 was relativelyuninformative with one significant exception. Modescorresponding to the stretching vibrations of the carbonylgroups were not observed in any ofthe spectra. This resultimplies that either the transition moment of the carbonylgroup, and therefore the C:O bond, is oriented parallelto the surface or that the bond between the carbon andthe oxygen has substantial single-bond character and asa result is shifted to a much lower frequency. For themonolayers on silver oxide, a peak is observed at about1395 cm 1. By comparison with studies on the infraredspectra of metal hydroxamates, we assign this band asbeing due predominantly to the C -N stretching motion.a8The presence of this band, in conjunction with the absenceof intensity in the carbonyl reg'ion, suggests that the C:Obond is oriented largely parallel to the surface and thatboth oxygens are not bound to the surface in the form ofa chelate.

Competitive Adsorption of Hydroxamic Acids andOther Adsorbates on Copper Oxide. Figure 7 quali-tatively examines the relative affinities of different headgroups for copper oxide. We used wettability to follow thecompetitive adsorptions of HO(CH2)15CONHOH against

(48) Brown, D. A.; McKeith, D.; Glass, W.K. Inorg. Chim. Acta lg7g,35,57-60. Brown, D. A. ; Roche, A. L. Inorg. Chem. Ig83,22,2lgg-2202.

Langmuir, Vol. 11, No. 3, 1995 827

Scheme 3. Qualitative Ranking of Strengths ofAdsorption

R a n k i n g o f s u b s t r a t e s f o r a f f i n i t r e s t o h y d r o x a m i c a c i d s :

A 1 O x

.9.:ooCLEooooG

5a

R a n k i n g o f

o x i d e s :

e o x > T L O X > > A u

r n r r c r l r r o r

- i P f : i I :" . r : i

' - ! l . r l J l { : : i > P a l a C I l

z t r c o n L u m , a n o

r l i : l : r : n e t a l o x i d e

CHr(CHz)rr"HG", where the head group ("HG")was eitherCOOH or CHzSH; we include "HG" : CONHOH forcomparison. We approximated the composition of theSAMs by relating the advancing contact angles of wateron the monolayers to the surface composition measuredbyXPS (Figure 2). Although the advancingcontact anglesare not linear in the composition of the SAM, we assumethat the wettabilities of the mixed SAMs depend only onthe amount of each tail group in the SAM and not on thehead group ol'the methvl-terminated molecule.ae We havenot established n'hether these results reflect kinetic orthermodvnamrc af f in i t ies.

These results show that the hydroxamic acid is favoredon the surface over the carboxylic acid, by a factor ofapproximately 4:1 (this value was determined from thecompositions of the solution needed to form monolayerswith normalized advancing contact angles of 0.5). It alsoshows that the thiol has a higher affinity for the surfaceof the copper oxide than does the hydroxamic acid; theIast data point, for example, is for a solution compositionof 30:1 which gives a monolayer that is approximately25% rn hydroxamic acid.

Discussion

Long-chain alkanehydroxamic acids form stable, ori-ented monolayers on all ofthe substrates with native metaloxides surveved in this study. Scheme 3 summarizesrankings of the interactions we have investigated in thisstudy. These conclusions were reached from the variousexperiments described in this paper. We discuss each ofthese rankings in turn.

Basic metal oxides are better substrates forhydroxamic acids than acidic metal oxides. Hy-droxamic acids form monolayers of higher stability onmetal oxides with isoelectric points (IEPs) greater thanthe pK^ of the acid (pK. of acetohydroxamic acid:9.3)2tthan on metal oxides with IEPs lower than the pK, (Table4).so-ss Most of the values in Table 4 were not measuredon native oxides. but thev are nontheless useful for

(49) We are assuming that the distribution of molecules in the SAMis independent ofthe head group, and, therefore, the structure ofthealkyl chains. This assumption becomes invalid when the structuresand the packing of the alkyl chains of the single-component SAMs aresigaificantly different.

(50 ) Parks , G . A . Chem. r?eu . 1965 ,65 ,177- I98 .(51)Kal lay, N. : Torbfc, Z. ; Gol ic , M.; Mat i jev ic. E. J. Phvs. Chem.

1991,95,7028-7032.(52) Kal lay, N. ; Torbic, Z. ; Barouch, E. ;Jednacak-Bisan, J. J . Col lo id

Inter face Sci . 1987, 118,43\-435.(53) Chau, L.-K.; Porter, M. D. J. Colloid Interface Sci.Iggl, 145,

283-286.

8 2 2 L o n g m u i r . \ i t l . / 1 . N o . 3 . 1 9 9 5

Table -1. Publ ished Values of Isoelectr ic Points ( IEPs) of

] leta l Oxides and Degree of Deprotonat ion ofSurface-Bound Hvdroxamic Acid"

F , i l he rs e t a l .

and perhapS n lo l 'e , , t 'derec i t l r r f l r r ] . lv t - l - : I n z : I 'C( . )n Ium ox ide

and therefore also be r:setul t I l : , ,ntt- ' : . . ' . 'hn:crt l r ippl icat ionson z i rcon ium oxtde for r i 'h rch phL r :D l l , , f l le . lc lc iS have been

used . l 6Self-assembied monolavers hrrve not been previousiy

studied on the surface of t i tanrum u\rde. Our results

suggest that on t i tanium oxide. phosphonlc acids form

more stable monolayers than either h1'drorirnl ic or car-

boxyl ic acids (Scheme 3). From the PIERS data. rt seems

unlikely that well-ordered monoiayers of alkane h.r'drox-

amic acids are obtained on t i tanium dioxide. The contact

angles show that their stability is not as great as that ofphosphonic acid SAMs. Carboxylic acids do not form

complete monolayers on titanium oxide from ethanol.

Experimental Section

General Information. Decanoy'l chloride (98ci t, dodecanoylchloride $8%t. tetradecanoyl chloride (971.r t . hexadecanoylchloride Q8%), heptadecanoyl chloride (98%), and octadecanol ' ' lchloride QgE) were purchased from Aldrich and used as received;16-hydroxyhexadecanoic acid (99+Vc), 1-ethyl-3-(3-(dimethyl-amino )prop1'l tcarbodiimide. and O-benzi'lhydroxylamine hydro-chloride rvere obtitined from Sigma and used without furtherpurif icat ion. Steanc acid and palmit ic acid were avai lable inthe laboraton' and u'ere not puri f ied prior to use. Octadeca-nephosphonic acrd u'as s1-nthesized using the Michaelis-Arbuzovreaction.;; Isooctane r99*ci. Aldrich) and hexadecane (anhy-

drous. Aldrich ) were precolated twice through neutral alumina(Fisher); ethanol (Pharmco) was used as received. Copper(Aldrich ), silver (Johnson Matthey Electronics), titanium (Aesar),

aluminum (Alfa). iron (Alfa), and zirconium (Aldrich) used forevaporations were of 99.99% or higher purity. Single-crystalsilicon wafers oriented in the [100] direction were obtained fromSil icon Sense (Nashua, NH). For infrared spectroscopy, wafersu'ere 2 in. in diameter and 0.01 in. thick; for all other uses, wafersu'ere 100 mm in diameter and -500 .am thick.

()e n e ral Synthes is of O -Benzylhydroxamates. A solution,,t ' tht, rre id chloricl t ' r 20 mmol t . O-benzylhydroxylamine (21 mmol),rrncl pr nci rr-rt. ') I tt'tnrol , u'its stirred in dry THF for t h at roomtenlper i r t - i r t ' ru t r l - t t I rsec l t t t 'n t lv re f luxed for t h . The so lu t ion\ \ ' i l : c ( )n( ' ( .n l l ' . r l ( ' ( l . i t l t i r .h t ' t ' t ' . idut ' $-a 's red isso lved in 200 mL ofe lhv l i i ce t l r l t ' I ' i t . . . , lL i l t r l t \ \ ' i1s ( 'x t l ' l tc ted tw ice wi th 100-mLpor t ions r r t ( - ) . l i \ l K I { : r ) : . ( ) l - } r ' r 'u ' r th 1 t lO mL of saturatedNaHCO, r and once u - t t h l ( ) t )mL b r tne . The o rgan i c l aye r wassubsequently'dried over Na:SOr. The solr-ent wa's evaporatedand the res idue was red isso lved in CH:C1: 'hexane t1"1r . Am-monia was bubbled through to precipitate traces of the free acidas the ammonium salt. The solution was filtered. Afterevaporation of the solvent, the corresponding O-benzylhydrox-amate was obtained (92-987c crude yield).

N-(Benzyloxy)decanamide. 1H NMR (300 MHz, CDCI3) d8 . 1 7 ( b r s , 1 H ) , 7 . 3 5 ( b r s , 5 H ) , 4 . 8 8 ( b r s , 2 H ) , 2 . 0 1 ( b r , 1 . 6 H ) ,1 .57 (m , 2H) ,1 . . 23 (b r m , 12 H ) . 0 .85 ( t . 3 H , J : 7 Hz ) ; HRMS(FAB) calcd for CrHzxNO: tN{ * H ' t 278.2120. found 278.2100.

N-(Benzyloxy)dodecanamide. rH N\IR t250 NIHz. CDCl,r rd 8 . 3 7 ( b r s . 1 H ) , 7 . 3 8 ( b r s . 5 H r . ' 1 . 9 0 r b r s . 2 H r , 2 . 3 5 ( b r . 0 . ' l

H ) . 2 . 0 3 f t , 2 H , J : 7 H z ) , 1 . 5 9 ( m , 2 H t . 1 . 2 5 ( b r m , 1 4 H ) , 0 . 8 8( t . 3 H. J :7 Hz) ; HRNIS (FAB) ca lcd for CrgHszNOz (M + H+)306.2433, found 306.2428. Anal. Calcd for CrsH:lNOz: C, 7 4.7 l;H . 10 .23 : N . 4 .59 . Found : C ,74 .77 ; H , 10 .64 ; N ,4 .51 .

N-(Benzyloxy)tetradecanamide. 1H NMR (400 MHz,C D C l , r i ) 7 . 9 5 r b r s . 1 H ) . 7 . 3 6 ( b r s , 5 H t , 4 . 9 0 ( b r s , 2 H ) , 2 . 0 1(b r . 2 H r . 1 .58 rm . 2 H ) . 1 .23 (b rm . 20 H ) , 0 .86 ( t , 3 H , J : 7 Hz ) ;HRMS IFABr c : r lcd {br C: rH: raNO2 (M t Na*) 356.2565, found356.2550. Anal . Calcd f t i r C: rH ' r ;NOz: C, 75.63; H, 10 '58; N,4 .2 . Found : C , 75 .51 ; H . 10 .97 ; N , 4 .21 .

N-(Benzyloxy)hexadecanamide. lH NMR (250 MHz,CDCIaI d 8 .02 (br s . 1 H) , 7 .36 (br s , 5 H) , 4 .88 (br s , 2 H) , 2 .35(br s , 0 .4 H) , 2 .01 (s , 2 H) , 1 .60 (m, 2} j ) , I .23 (br m, 24 H) , 0 .86(t, 3 H, J :7 Hz); HRMS (FAB) calcd for CzsHaoNOz (M + H+)

(57 )Koso lapo f l G . M. J . An t . Chem. Soc . 1945 , 67 , 1180-1182 .Bhattacharya, A. K. : Thyagarajan, G. Chem. Reu. 1981, 81,415-430.

(58 ) Microanalyses were performed by Oneida Research Services inWhitesboro. NY.

l ' . r . ' . . i l

publ ishedva lues o f IEP

IRCONHO 1", 'r ltRCONHOHl,,.r

9 .51 '1 0. .1'1

'2 .h 1 -7 '7 - 106 .j ( . { 'a

5 -9 ,

87 . 70.430.40

\ ' . r iu . . . t i rk t 'n I i 'om Figure 4. i 'Values taken f rom refs 51 andr l \ ' . r lLrrs t r tken f i -om ref 50. ' l Value taken f rom ref 53. " This' , r i lu t ' i . l , r ' th€ natural ly occurr ing mineral . Other values l is ted

in r t . : i t t r l 'ere measured on synthet icZrOz made under strongly

i l r t ( l . l ' : t ronglv basrc condi t ions; the condi t ions of preparat ion

r i | ) | ) t , . i r t ,d t ( ) rnf luence the measured value of the IEP. / Ni t rogen 1s

>l) t , r ' l l ' i i s ' r re not taken for these samples.

ratronal iz ing the t rends we have observed.5a According

to the data in Table 4, the most stable monolayers shouldform on silver oxide and copper oxide. We note that thereis onlv a ver)' loose correlation between values of IEP andthe ratios of deprotonated hydroxamic acid to protonatedhr-droxamic acid inferred from the N 1s spectra (Figure4 t .

Although thiolates are more stable on copper andsilver, hydroxamic acids present a versatile alter-native for these metals. Our results indicate thatthiolates form the most stable system of monolayers onccpper and silr.er.;; Hydroxamic acids form more stableoverlal 'ers than carboxylic acids or phosphonic acids onthese substrates (Scheme 3 i and provide an importants]'stem for the formation of thin organic f,rlms on the nativeoxides of copper and si lver. There al 'e two advantages tothe use of h1'droxamic acids rather than thiols on copperand silver. First, in the case of copper oxide. thehydroxamic acid forms a strong bond to the surface butdoes not etch the surface during formation as much as thethiol does. Second, hydroxamic acids are not as prone tooxidation as thiois and may be more stable to long-termexposure to ambient conditions.20'56 Monolayers of thi-olates photooxidize on the surface, forming labile metal-sulfonate linkages with the surface.20':'6 The oxidativestability of the hydroxamic acids suggests their use as anaiternative to thiols as protective coatings. Potential usesof these monoiayers will likely be found in lithography,corrosion resistance, and tribology.le'20

Hydroxamic acids present an alternative andimprovement to other organic acids for the forma'tion of self-assembled monolayers on most acidicmetal oxides. Hydroxamic acids form stable monolayerson "acidic" metals oxides (i.e., those with isoelectric pointslower than the pK, of the hydroxamic acid),:'a althoughthey are bound predominantly as the parent acid. On thenative oxides of aluminum, zirconium, and iron, hydrox-amic acids appear to form more stable monolayers thaneither carboxylic acids or phosphonic acids (Scheme 3).The smaller size of the hydroxamic acid compared to thephosphonic acid may allow the formation of more coherent

ro.1 r The most relevant set of IEPs were determined using metalbead-. b)' measuring the uptake of charged latex particles' althoughzirconium. i ron, and s i lver were not studied.rr l s2 Jhs IEP of the nat iveoxide of silver was measured by measuring contact angles as a functiono{ ' the pH of ' the drop.53

r 55 r A 1:1 solution ofoctadecanethiolate and octadecane hydroxamicacid in ethanol yielded a monolayer on silver oxide that was 9:1 in favorof thiolate by XPS: Folkers, J. P.; Whitesides, G. M. Unpublished results'

r 56 i Monolayers of thiolate are prone to oxidation even on gold. Forexamples on gold, see: Tarlov, M. J.; Newman, J. G. Langm-uir 1992,7. 1398 1405. Huang, J. ; Hemminger, J. C' J. Am. Chem. Soc. 1993,r I 5. 3342-3343.

SAMs of Hydroxamic Acids

362.3059, found 362.3036. Anal. Calcd for CzrHrgNOz: C,76.4;H , 10 .87 ; N ,3 .87 . Found : C ,76 .25 : ' H , 11 .29 ; N , 3 ' 54 .

N-(Benzyloxy)heptadecanamide. 1H NMR (250 MHz'C D C I 3 ) 6 8 . 1 2 ( b r s , 1 H ) , 7 . 3 6 ( b r s , 5 H ) , 4 . 8 8 ( b r s , 2 H ) , 2 . 3 5(b r s , 0 .4 H ) , 2 .01 (b r , 2 H ) , 1 .58 (m , 2 H ) , 1 .23 (b r m ,26 H ) ' 0 .86(t, 3 H, J :6Hz);HRMS (EI) calcd for CzaH+zNOz (M*) 375.3138,found 375.3126. Anal. CalcdforCz+HarNOz: C,76.75; H, 11.0;N ,3 .73 . Found : C ,76 .74 ; H , 11 .14 ; N , 3 .65 '

N-(Benrylory)octadecanamide. lH NMR (250 MHz, CDCIg)d 8 .3? (br s , 1 H) , 7 .38 (br s , 5 H) , 4 .90 (br s , 2 H) , 2 .03 (br , 2 H) ,1.60 (m, 2H), L25 (br m, 28 H), 0.88 (t , 3 H, J :7 Hz); HRMS(FAB) calcd for Cz;H++NOz (M + H+) 390.3372, found 390.3358.Anal . Calcd for Cz;H.rsNO2: C, 77.07;H, 11.12;N,3.59 ' Found:C . 7 7 . 1 7 : H . 1 1 . 0 9 ; N , 3 . 4 9 .

Synthesis of 16-Hydroxy'N'(benzyloxy)hexadecana'mide. 16-Hydroxy-.1/-(benzyloxy)hexadecanamide was preparedby a slight variation of a method for synthesizing trihydroxamicacids that can tolerate hydroxyl groups.23 The pH of a solutionof 1 g (3.67 mmol) of 16-hydroxyhexadecanoic acid and 0.604 g(3.85 mmol) of O-benzylhydroxylamine hydrochloride in THF/lHzO (.2:l) was adjusted to 4.8 by adding 1 N NaOH. A 1.41-9(7.34 mmol) portion of 1-ethyl-3-(3-(dimethylamino)propyl)-carbodiimide was added in small amounts to the stirred solutionand the pH was held approximately constant by adding 1 N HCl.After being stirred for an additional 6 h, the reaction mixturewas extracted twice with 200-mL portions of ethyl acetate. Thecombined organic phases were extracted, washed with 150 mLof saturated NaHCOg, 150 mL of 0.5 M citr ic acid. 150 mL ofu'ater. and 150 mL of brine. and dried over NazSO.r ' The solut ion\vas concentrated and the residue rvas purif ied bv columnchromatographr' , si l ica: hexane ethf i acetate methanol 7 6 2Y ie ld : 0 .8 g .60e . rH N l lR 250 \ IHz . CDCI , , ) 7 .95 b r s . 1 H .7 . 3 8 ( b r s , 5 H . 4 . 9 0 , b r s . 2 H . 3 . 6 3 t . 2 H . J - - ; H z . 2 . 3 5 b r .0 . 4 H ) . 2 . 0 2 r b r . 1 . 6 H r . 1 . 5 9 r m , 4 H ' . 1 . 2 5 ' b r m . 2 2 H : H R \ I S(FAB) ca lcd for CzsHroNOs r I I - H- ' 378.3008. found 378.2983.Anal . Calcd for CzsHseNOs: C, 73. I7 H. 10 '41: N. 3 .71 ' Found:C , 73 .01 ; H , 10 .29 ; N , 3 .57 ,

General Synthesis of Hydroxamic Acids. The O-benzyl-hydroxamates (20 mmol) were dissolved in 200 mL of ethanoland hydrogenated for 6-12 h at a H2 pressure of 1 atm using 50mg of Pd on carbon as catalyst. The solutions were filtered overCelite and concentrated, yielding 95-I00Vc of the free acid. Uponrecrystallization from CHzClzlhexane (1:1) 70-80Vc of the hy-droxamic acid were recovered.

N-Hydroxydecanamide.ss 1H NMR (400 MHz, DMSO-do)d 10 .32 ( s , 0 .9 H ) , 9 .67 ( s , 0 .1 H ) , 8 .97 ( s , 0 .1 H ) , 8 .65 ( s , 0 .9 H l ,2 . 2 5 ( t . 0 . 2 H , J = 7 H z ) , L 9 2 ( t , 1 . 8 H , J : 7 H i l , I . 4 7 ( q , 2 H .

J : 7 Hz) , 1 .3- 1 .1 (br m, 12 H) , 0 .86 ( t , 3 H, J : 7 Hz t ; HRNIS(FAB) ca lcd for CroHzzNOz tM * H* t 188.1650, found 188.1626.Anal . Calcd for CroHzrNOz: C.64.13: H. 11.3: N. 7 .48 Found:C , 63 .91 ; H , 11 .38 ; N , 7 .4 .

N-Hydroxydodecanamide. 1H NIIR r 250 llHz. DIISO-d6 r

d 1 0 . 3 2 ( s , 0 . 9 H ) , 9 . 7 1 ( s , 0 . 1 H l , 8 . 9 7 r s . 0 . 1 H r . 8 . 6 5 r b r s . 0 . 9H ) , 2 . 2 4 ( b r , 0 . 2 H ) , 1 . 9 2 ( t , 1 . 8 H , J : i H 2 , . 1 . 5 5 - 1 . 4 t m , 2 H t ,1 .3-1.1 (br m, 16 H) , 0 .86 ( t , 3 H, J : i Hz r : HRI IS rFAB r ca lcdfor CrzHzoNOz (M + H+) 216.1963, found 216.1948. Anal . Calcdfor CrzHzsNOz: C, 66.93; H, 11.7; N, 6 .5 . Found: C. 67.00: H '11 .37 ; N , 6 .4 .

N-Hydrorytetradecanamide. lH NX'IR t250 N{Hz. DMSO-d e ) d 1 0 . 3 ( s , 0 . 9 H ) , 9 . 7 4 ( s , 0 . 1 H ) , 8 . 9 7 ( s . 0 . 1 H t . 8 . 6 4 ( s , 0 . 9

H ) , 2 . 2 3 ( b r , 0 . 2 H ) , 1 . 9 1 ( t , 1 . 8 H . J : 7 H 2 t . 1 . 5 5 - 1 ' 4 ( m , 2 H ) ,1 . 3 5 - 1 . 1 5 ( b r m , 2 0 H ) , 0 . 8 4 ( t . 3 H , J = i H z ) ; H R M S ( F A B )

calcd for Cr+HgoNOz (M -f H- I244.2276, found 224.2261. Anal 'Calcd for Cr+HzgNOz: C, 69.09; H, 12.01; N, 5.75. Found: C'68.95; H, 11.88; N, 5 .60.

N-Hydroxyhexadecanamide. I H NMR (250 MHz, DM SO -

do ) d 10 .3 ( s , 0 .9 H ) , 9 .71 ( s ,0 .1 H ) , 8 .96 ( s , 0 .1 H ) , 8 ' 64 ( s , 0 .9H) ,2 .22 (br , 0 .2 H) , 1 .90 ( t , 1 .8 H, J : 7 Hz) , 1 .55-7.4 (m, 2 H) ,1 .35-1.15 (br m, 24 H) ,0 .84 ( t , 3 H, J : 7 Hz) ; HRMS (FAB)

(59) Two sets of peaks in the lH NMR are observed for the hydrogenson the oxygen and nitrogen in the hydroxamic acid and the hydrogensin the methylene group o. to the hydroxamic acid. These two setscorrespond toE andZ conformations of the hydroxamic acid. For moredetail. see: Brown, D. A.; Glass, W. K.; Mageswaran, R.; Girmay, B.Magn. Reson. Chem. t988,26,970-973. Brown, D. A. ; Glass, W. K. ;Mageswaran, R.; Mohammed, S. A. Magn. Reson. Chem. 1991, 29, 40-45.

Langmuir,Vol. 11, No. 3, 1995 823

calcd for CraHs+NOz (M + H') 272'2589, found 272.2573' Anal.CalcdforCreH3sNO2: C, 70.8; H, 12.25;N, 5.16. Found: C,70.74;H . 12 .48 : N , 5 .10 .

N-Hydroxyheptadecanamide. 1 H NMR CI50 MHz, DMSO-dn ) d 10 .3 ( s , 0 .9 H ) . 9 .71 ( s , 0 .1 H ) , 8 ' 97 ( s , 0 .1 H ) , 8 ' 64 ( s , 1 H ) ,2 . 2 3 ( b r . 0 . 2 H ) , 1 . 9 1 ( t , 2 I H , , J : 7 H z ) , 1 . 5 5 - 1 . 4 ( m , 2 H ) , 1 . 3 5 -1.15 (br m. 26 H) , 0 .84 ( t , 3 H, J :7 Hz) ; HRMS (FAB) ca lcd for

Cr;HrsNO,:\a rN{ * Na* }308.2565, found 308'2575. Anal. Calcdfo r C r ;H : ;NO: : C . 71 .53 : H . 12 .36 ;N , 4 .9 i . Found : C ,7 l ' 72 ;H ,12.35: N. 4 .80.

N-Hydroxyoctadecanamide. 1H NMR (250 MHz, DMSO-da ) d 10 .31 { - s . 0 .9 H r , 9 .71 1s . 0 .1 H r . 8 .97 ( s , 0 .1 H ) , 8 .64 ( s , 1 H ) ,2 .22 (b r , 0 .2 H r . 1 .91 r t . 1 .8 H . J : 7 Hz ) ,1 .6 -1 .1 (b r m , 30 H ) ,0.84 (t . 3 H. J : 7 Hz); HRMS tFAB t calcd for CraHgeNOz (M +

H*) 300.2902. found 300.2895. Anal. Calcd for CrsHazNO2: C,72 .19 :H .12 .45 : N ,4 .68 . Found : C .72 .23 ; H , 12 .68 ; N ,4 .6 .

l6JV-Dihydroxyhexadecanamide. 1H NMR (300 MHz'MeOH-d+) d 3.43 (t , 2 H, J : 7 Hz), 1.97 (t , 2 H, J : 7 Hz'),1 .55-1.35 (m, 4 H) , 1 .3-1.1 (br m, 22H); HRMS (FAB) ca lcd for

CroHaaNOa (M -r H*) 288.2539,found 288.2541. Anal. Calcd forCroHssNOs: C, 66.86; H, 11.57; N, 4 .87. Found: C, 66.95; H,11 .58 ; N , 4 .75 .

Preparation of Substrates. The copper, silver, zirconium,iron, and aluminum films (1000 A; thinner films, generally 600-800 A. were used for aluminum, because of the high power neededto evaporate the metal) were prepared by"evaporation of themetals onto silicon wafers primed with 50 A of titanium as anadhesion iaver: t i tanium {1000 Atrvas evaporated direct ly ontothe si l icon u-afers, Al l nietals were evaporated at a rate of 3 A/s.The evaporatr(rns n'ere performed in a cr1'ogenical ly pumped

electron-be:tI ' r^ e\- i lporatol base pressure < 1 x 10-; Torr). Aftercvaporatlon the chrlmber r i 'as backfi l led with nitrogen. The

san-rples \\ 'ere tr i tnsf 'erred through air to the solut ions of thehvdroxamic ac ids r i ' i th rn 5-10 min.

Preparation of Monolayers. Monolayers were formed fromsolutions in either isooctane or ethanol with concentrations of1 mM in adsorbate. The hydroxamic acids, however, were notsoluble in isooctane at 1 mM, but enough dissolved to formmonolayers. Monolayers for infrared spectroscopy were formedin glass Petri dishes, which had been cleaned with "piranhasolution" (7:3 concentrated HzSO,tl3j% HzOz) at -90 'C for 30min. rinsed with deionized water, and dried in an oven at 125"C prior to use.

WARNING: Piranha solution should be handled withcaution; it has detonated unexpectedly.6o For all other uses,monolavers were formed in single-use glass scintillation vialsr20 mL r. The slrdes were removed from solut ions after 1-2 days,nnsed u' i th ethanol and heptane, blown dry with N2, andcharactenzed. The method for forming solutions for mixedmonolavers has been described previously'34 35'3e

Measurement of Contact Angles. Contact angles weremeasured on a Rame-Hart } lodel 100 goniometer at roomtemperature and ambient humidity. Water for contact angleswas deionized and dtst i l led in a glass and Teflon apparatus.Advancing and receding contact angles were measured on bothsides of at least three drops of each l iquid per sl ide; data in thefigures represent the average of these measurements' Thefollowing method was used for measuring contact angles: A dropapproximately 1-21rL in volume was grown on the end of thepipette tip (Micro-Electrapette syringe; Matrix Technologies:Lowell, MAt. The tip was then lowered to the surface until thedrop came in contact with the surface. The drop was advancedby slowly increasing the volume of the drop (rate -l pUs)'Advancing contact angles of water were measured immediatelyafter the front of the drop had smoothly moved a short distanceacross the surface. Receding angles were taken after the drophad smoothly retreated across the surface by decreasing thevolume of the drop.

X-ray Photoelectron Spectroscopy (XPS). X-ray photo-electron spectra were collected on a Surface Science SSX-100spectrometer using a monochromatized Al Kct source (hv:1486.6

eVi. Th" spectra were recorded using a spot size of 600 ,am anda pass energy on the detector of 50 eV (acquisition time for one

(60) Several warnings have appeared concerning "piranha solu-tion": Dobbs, D. A.; Bergman, R. G.; Theopold. K. H. Chem. Eng.I'{ews1990 .68 (17 \ .2 . Wnuk , T .Chem.Eng .News 1990 ,68 (26 ) ,2 . Ma t low.S. L. Chem. Ens. News 1990, 68 (30). 2.

824 Langmuir, Vol. 11, No. 3, 1995

scan was approximately 1.5 min). For the monolayers, spectrau'ere collected for carbon, nitrogen, and oxygen using the 1s peaksiit 285. ,400. and 530 eV, respectively; the binding energies forelement,s in the monolayer were referenced to the peak due tohvdrocarbon in the C 1s region, for which we fixed the bindinge.nergl at 284.6 eV.a2 Spectra for the solid hydroxamic acid werecollected using an electron flood gun of 4.5 eV to dissipate chargein the sample. The following sigaals were used for the sub-strates: Cu 2p,r .z at 932 eV for Cu(O), and at -934 eV for Cu(II);Ag 3d;,. 3d:;: at 368 and 374 eV for al l oxidation states of si lver;Ti 2p'r, . 2pr: at 454 and 460 eV for Ti(O), and at 459 and 464.5eV fbr TirI \rr: Al 2p at 73 eV for Al(0), and at 75 eV for AI(I I I) ;Zr i )d ; , .3d, r : a t 179 and 181 eV for ZrQ) , and at 183 and 185.5

Folkers et al.

eV for Zr(Nl Fe 2plz, 2pnat 706.5 eV and 719.5 eV for Fe(0),and at 710 and 723.5 eV for Fe(II I t .12 The binding energies forthe substrates were not standardized to a reference sample. Allspectra were fitted using an 807r Gaussian/20% Lorentzian peakshape and a Shirley background subtraction.6l

Infrared Spectroscopy. The experimental apparatus fortaking infrared spectra has been described previously.a 8 Theangle between the incident beam and the surface normal wasapproximately 85".

L49407 47r

(61r Shir ley, D. A. Pfrv.s. Rer ' . B 1972.5,4709 4774