DOI: 10.1021/la902979m 3077Langmuir 2010, 26(5), 3077–3083 Published on Web 10/09/2009

pubs.acs.org/Langmuir

© 2009 American Chemical Society

Role of an Amide Bond for Self-Assembly of Surfactants

Romain Bordes,*,† Juergen Tropsch,‡ and Krister Holmberg†

†Department of Chemical and Biological Engineering, Chalmers University of Technology, SE-412 96G€oteborg,Sweden, and ‡BASF, 67056 Ludwigshafen, Germany

Received August 11, 2009. Revised Manuscript Received September 18, 2009

Self-assembly in solution and adsorption at the air-water interface and at solid surfaces were investigated for twoamino-acid-based surfactants with conductimetry, NMR, tensiometry, quartz crystal microbalance with monitoring ofthe dissipation (QCM-D), and surface plasmon resonance (SPR). The surfactants studied were sodium N-lauroylgly-cinate and sodium N-lauroylsarcosinate, differing only in a methyl group on the amide nitrogen for the sarcosinate.Thus, the glycinate but not the sarcosinate surfactant is capable of forming intermolecular hydrogen bonds via the amidegroup. It was found that the amide bond, N-methylated or not, gave a substantial contribution to the hydrophilicity ofthe amphiphile. The ability to form intermolecular hydrogen bonds led to tighter packing at the air-water interface andat a hydrophobic surface. It also increased the tendency for precipitation as an acid-soap pair on addition of acid.Adsorption of the surfactants at a gold surface was also investigated and gave unexpected results. The sarcosine-basedsurfactant seemed to give bilayer adsorption, while the glycine derivative adsorbed as a monolayer.

1. Introduction

Surfactants self-assemble in aqueous solution as a result of thehigh cohesive energy density of water.1,2 A watermolecule prefersto be surrounded by other water molecules and expels mostsolutes from the bulk solution. The amphiphilic nature ofsurfactants governs their packing into monolayers or bilayers atinterfaces and subsequently, when all interfaces are occupied, intomicelles and other organized structures in the bulk phase.3,4

Besides the entropic contributions, there are twomain compet-ing interactions in the process of self-assembly. van der Waalsinteractions between the hydrocarbon tails of the surfactantsfavors micellization and other self-assembly processes. Repulsionbetween headgroups as they come close together opposes self-assembly. The opposing effect can be significant; it is the reasonwhy ionic surfactants typically have a 2 orders of magnitudehigher value of the critical micelle concentration (CMC) thannonionic surfactants.

Other attractive interactions than van der Waals interactionsmay contribute to the self-assembly process, but this is not muchdiscussed in the general literature on micellization and other self-assembly processes. Groups that may act as both donors andacceptors of hydrogen bonds may increase the intermolecularattraction considerably. The amide bond is probably the primeexample of such a group, and amide bonds are not uncommon insurfactants as a linker between the hydrophobic tail and the polarheadgroup. This is, for instance, the case in monoethanolamideethoxylates of fatty acids, and it is also the case in surface activeN-acylated amino acids. N-Acylglycinate and N-acylglutamateare examples of such amino-acid-based surfactants.

The aim of the present work is to investigate the role ofhydrogen bonding in the self-assembly of surfactants. We have

used two surfactants for the purpose, sodium N-lauroylglycinateand sodium N-lauroylsarcosinate.5,6 As shown in Figure 1, theseare almost identical anionic surfactants with an amide bondconnecting the hydrophobic tail and the polar headgroup.7,8

The difference between them is that the amide nitrogen ismethylated in the sarcosinate but not in the glycinate. This meansthat the amide bond can be a hydrogen bond donor only for theglycinate; that is, hydrogen bonding can give a contribution to theintermolecular attraction for the glycinate but not for the sarco-sinate.

Self-assembly of the two surfactants was investigated in aque-ous solution by tensiometry and by conductivity measurementsand at solid surfaces by quartz crystal microbalance with dissipa-tion monitoring (QCM-D) and by surface plasmon resonance(SPR).

2. Materials and Methods

Materials. Lauroyl chloride (Aldrich, 98%), L-glycine ethylester (Aldrich, 99%), 1-hexadecanethiol (Fluka,>95%), sodiumhydroxide (Fluka, g98%), hydrochloric acid (Riedel de Ha€en,37%), and sodium N-lauroylsarcosinate (C12SarcNa) were usedas purchased. Pyridine (Aldrich, 99%)was used freshly distilled invacuum. The acidic and the alkaline solutions used for thetitrations were prepared from stock solutions of 10 M sodiumhydroxide (Fluka) and 10 M hydrochloric acid (Fluka). Milli-Qwater (resistivity >18 M Ω) was used for preparation of theaqueous solutions.

Synthesis of Sodium N-Lauroylglycinate (C12GlyNa).L-Glycine ethyl ester (10 g) was dissolved in pyridine (20 mL) ina round-bottomed flask under reflux. Lauroyl chloride (15.7 g)was added dropwise under stirring, leading to a white precipitate.The solvent was evaporated. The product was dispersed in water,stirred, and then filtered and recrystallized from ethanol. The N-lauroyl ester formed was dissolved in ethanol (100 mL), andaqueous sodium hydroxide (2 M, 35 mL) was added dropwiseunder stirring, leading to a white precipitate of the sodium salt of

*Corresponding author. E-mail: bordes@chalmers.se.(1) Moy�a,M. L.; Rodrıguez, A.; delMar Graciani, M.; Fern�andez, G. J. Colloid

Interface Sci. 2007, 316, 787–795.(2) Sjoeberg, M.; Henriksson, U.; Waernheim, T. Langmuir 1990, 6, 1205–1211.(3) Israelachvili, J. Intermolecular and Surface Forces, 3rd ed.; Academic Press:

London, 2007.(4) Evans, D. F.; Wennerstr€om, H. The Colloidal Domain, 2nd ed.; Wiley-VCH:

New York, 1999.

(5) Sakamoto, K.; Hatano, M. Bull. Chem. Soc. Jpn. 1980, 53, 339–343.(6) Shinitzky, M.; Haimovitz, R. J. Am. Chem. Soc. 2002, 115, 12545–12549.(7) Oshimura, E.; Yamashita, Y.; Sakamoto, K. J. Oleo Sci. 2007, 56, 115–121.(8) Yahagi, K.; Tsujii, K. J. Colloid Interface Sci. 1987, 117, 415–424.

3078 DOI: 10.1021/la902979m Langmuir 2010, 26(5), 3077–3083

Article Bordes et al.

the surfactant in a yield of 46%. NMR data of the product aregiven in the Supporting Information.

Methods. pH Measurements and Titrations. The titra-tions were made by using an automatic titrator, Titrino 785DMP(Metrohm), using the software Tiamo (Metrohm). Mixtures(25 mL) were prepared and acidified with the required amountof 1 M HCl and then titrated with a 1 M sodium hydroxidesolution using incremental additions of 5 μL and awaiting time of10 s between additions.

Determination of the Krafft Point. The Krafft temperaturewas determined visually by heating the solution in a thermostatedbath at a rate of 2�/min. The experiments were repeated at leasttwo times.

NMR. NMR analyses were carried out on a Jeol 400 MHzspectrometer at 25 �C. Deuterated solvent was purchased fromArmar Chemicals. The solubility measurements by NMR weremade using acetonitrile as a reference (5 μL for 1mLof surfactantsolution). The samples of surfactant were prepared in D2O, andthe required amount of a concentrated solution of CaCl2 or HClwas added. The samples were then stirred for several hours.

Surface Tension Measurements. Surface tension measure-ments were carried out on a Sigma 70 tensiometer (KSV) usingthe du No€uy ring method. The temperature was kept at 25 �C((0.01 �C) by using a cryostat Neslab RTE-200 instrument. Theglassware was cleaned with chromosulfuric acid, and the ring wasburned prior to use. The CMCs for C12SarcNa at different pHvalues were determined using a solution adjusted to the specificpH with concentrated HCl or concentrated NaOH solution.

Conductivity. The CMC and the ionization degree weredetermined by conductivity measurements. The conductivitymeter was a CDM 210 (Radiometer Copenhagen), and thesolutions were thermostated with a water bath Haake FisonDC1 ((0.1 �C). The surfactant solution at a concentration abovethe CMC was diluted by addition of 200 μL of water every 20 swith an automatic buret 765 Dosimat (Metrohm). The conduc-tivitymeter and the automatic buret were controlled by computervia an interface RS232, and a homemade software written inPython (version 2.4).

Quartz Crystal Microbalance with Monitoring of the Dis-sipation (QCM-D). A QCM-D instrument (model D300) fromQ-sense AB (G€oteborg, Sweden) was used. The measurementswere done under nonflowing conditions, in order to avoid crystalperturbations during the shear oscillation of the crystals. TheAT-cut crystals coated with a 100 nm thin gold layer were also fromQ-sense. Prior to use, a cleaning procedure was conducted asfollows: UV-ozone treatment for 10 min, immersion in a 5:1:1mixture of H2O/ammonia (25%)/H2O2 (30%) at 75 �C for 5 min,rinsing with Milli-Q water, drying with N2, and finally 10 min ofUV-ozone treatment.

The hydrophobic self-assembledmonolayers (SAM-CH3) wereprepared by immersing the cleaned gold surfaces in alkanethiolsolution (2mMin ethanol) for at least 16 h.Then the crystalswererinsed with ethanol and sonicated in ethanol for 5 min to removeloosely adsorbed alkanethiols. Finally, the surfaces were driedwith nitrogen prior to use.

The measurements were carried out at 20 �C with a baselinecorresponding to Milli-Q water. The crystal was exposed to thesolutionusing ahomemadeadditiondevice, going fromthe lowestto the highest concentration without rinsing the surface betweenthe additions. The crystal was finally rinsed withMilli-Q water toremove poorly adsorbed surfactants.

The automatic addition device is based on electric valves thatopen different flow channels on a collector connected at the inlet

of theQCM-Dcell via an electronic interface using remote controlby the com port of a computer. The software was developed inPython (version2.4). Prior touse, all the flowchannelswere rinsedwith ethanol and water.

Surface Plasmon Resonance (SPR). The measurements ofthe variations of the surface plasmon resonance weremade with aSPR Biacore X instrument from Biacore SIA. Details about thetechnique and the apparatus can be found elsewhere.9 Thesurfactant solution was injected (40 μL) in the analysis chamberunder a continuous flow of 25 μL/min. The gold chips fromBiacore SIA were cleaned prior to use by using the followingprocedure: UV-ozone treatment for 10 min, immersion in a 5:1:1mixture of H2O/ammonia (25%)/H2O2 (30%) at 75 �C for 5 min,rinsing with Milli-Q water, drying with N2, and finally 10 min ofUV-ozone treatment.

3. Results and Discussion

The section is divided into twoparts. The first part concerns thebehavior of the two surfactants, sodium N-lauroylglycinate(C12GlyNa) and sodium N-lauroylsarcosinate (C12SarcNa), inaqueous solution. The secondpart concerns adsorptionof the twosurfactants at surfaces, studied by the two complementary meth-ods, QCM-D and SPR.3.1. Solution Behavior. CMC and Degree of Ioniza-

tion.TheCMCof the surfactantswas determined by conductivityand surface tension measurements. The CMC values were takenat the break of the conductivity versus concentration curve10,11

and at the point on the surface tension versus concentration curveatwhich the surface tension reaches a constant value, respectively.The surface tension plots did not exhibit a dip at the break point,indicating that the surfactants were free from hydrophobicimpurities. The values, reported in Table 1, show that bothtechniques give a slightly lower CMC for C12SarcNa than forC12GlyNa. The values obtained by conductivity are somewhathigher than those determined by tensiometry. This has beenobserved before for amino-acid-based surfactants.12

One may compare the CMC values determined for C12GlyNaand C12SarcNa with those of sodium myristate and sodiumpalmitate, for which the values 1.5 and 0.55 mM have beenreported, measured by tensiometry.13 Both C12GlyNa andC12SarcNa have a carboxylate headgroup at the end of a straight14 atom tail; thus, a length in-between that of myristate andpalmitate, which have straight C13 and C15 hydrocarbon tails,respectively, connected to a carboxylate end group. The mean ofsodiummyristate and sodium palmitate, that is, an assumed alkylcarboxylate with a 14 carbon tail, can be expected to have a CMCaround 1 mM. This value is 1 order of magnitude lower than thevalue recorded for the amino-acid-based surfactants. The higherCMC of the amino-acid-based surfactant must be attributed tothe amide bond, which evidently imparts substantially increasedwater solubility of the unimer.

Figure 1. (left) SodiumN-lauroylglycinate and (right) sodium N-lauroylsarcosinate.

(9) Oskarsson, H.; Holmberg, K. J. Colloid Interface Sci. 2006, 301, 360–369.(10) Rosen, M. J. Surfactants and Interfacial Phenomena, 3rd ed.; Wiley:

New York, 2004; pp 139-144.(11) Zana, R. J. Colloid Interface Sci. 1980, 78, 330–337.(12) Bordes, R.; Tropsch, J.; Holmberg, K. J. Colloid Interface Sci. 2009, 338,

529–536.(13) Zimmels, Y.; Lin, I. J. Colloid Polym. Sci. 1974, 252, 594–612.

DOI: 10.1021/la902979m 3079Langmuir 2010, 26(5), 3077–3083

Bordes et al. Article

The only difference between the two surfactants is theN-methyl group that is present on C12SarcNa but absent onC12GlyNa. An extra carbon is known to give a certain contribu-tion to surfactant hydrophobicity if it is inserted in the hydro-phobic tail region. For instance, addition of one methylene groupto the alkyl chain of anionic surfactants usually reduces the CMCby a factor of 2.14,15 However, it is known that introduction of ashort alkyl group in the polar headgroup of a surfactant will giveonly a marginal effect on the CMC.16 Thus, with respect to groupcontribution to hydrophobicity, the values are according toexpectations. More interesting, and to us more surprising, is thefact that the blocking of the possibility for hydrogen bondinginvolving the amide linkage does not affect the CMC values. Theimmediate interpretation is that intermolecular hydrogen bond-ing does not play a role for the self-assembly of amide-containingsurfactants. This statement will be put into a broader contextbelow.

Determination of the CMC by tensiometry also provides anopportunity to determine the packing of the surfactant at theair-water interface using the Gibbs adsorption equation17

(although this approach of determining the area occupied by asurfactant at the air-water interface has recently been ques-tioned18). The surface excess concentration and the area permolecule were calculated using the Gibbs equation:

Γ¼ -1

2:303nRT

DγD log C

� �

A ¼ 1023

NAΓm

where γ is the surface tension in mN 3m-1, T is the absolute

temperature in K, R=8.31 (J 3mol-13K

-1), NA is Avogadro’sconstant, Γ is the surface excess inmol/1000 3m

2, andA is the areaper molecule in A2. A value of n = 2 was chosen for bothsurfactants.19

The values of surface excess and of area per molecule at theair-water interface are reported in Table 1. They indicate thatC12GlyNa packs considerably tighter than C12SarcNa. Thus,these results seem to indicate that self-assembly into a monolayeris assisted by the formation of intermolecular hydrogen bondsbetween the amphiphiles aligned at the surface. Furthermore,the steric hindrance of the N-methyl group and the potential

conformation, cis or trans, of the amide bond7 ofC12SarcNamayrender close alignment of the amphiphiles difficult.

Table 1 also gives values for degree of ionization of the twosurfactants. The values are obtained from the plots of conducti-vity versus surfactant concentration. The ionization degree, R, isderived from the ratio between the slopes above and below theCMC.11 The counterion binding, β, is defined as β=1 - R.

The considerably higher degree of ionization of the sarcosinatesurfactant than that of the glycinate surfactant is in accordancewith a less tight packing in a self-assembly of the former. Theeffect of proximity of charged headgroups on theR value has beenstudied systematically for cationic gemini surfactants in which thelength of the spacer unit has been varied. It was found that the Rvalue increased with increasing spacer length, that is, withincreasing distance between the charged groups.20-23

In summary, the values of CMC, packing at the air-waterinterface, and degree of ionization, all reported in Table 1,indicate that the possibility for formation of intermolecularhydrogen bonds between the surfactants in self-assemblies leadsto tighter packing of the amphiphiles both when aligned as amonolayer on the surface andwhen present asmicelles in aqueoussolution. However, the CMC is the same with and without theN-methyl group; that is, the energy ofmicellization is not affected.

Effect of pHonBulk and Interfacial Behavior.TheN-methylgroup in the sarcosine-based surfactant can be expected to affectthe solubility at low pH, whichmay be of practical importance forsome applications. In order to study this in some detail, theKrafftpoint of solutions ofC12GlyNa,C12SarcNa, and a 1:1mixture ofthe two surfactants were evaluated as a function of pH. Allsolutions were made at 5 times the CMC. Figure 2 shows thevariation of Krafft temperature as the molar ratio of HCl tosurfactant is varied.

As can be seen from the figure, the two surfactants behave verydifferently when exposed to acid and the mixture of the tworesembles the behavior of C12GlyNa. Neither C12GlyNa norC12SarcNa has a detectable Krafft point before addition of acid(expressed in the figure as a Krafft temperature of 0 �C). Theglycine-based surfactant shows high sensitivity to acid. Addition

Table 1. Values of CMC, Surface Excess, and Area Per Molecule at

the Air-Water Interface, and Degree of Micelle Ionization for the

Surfactants Sodium N-Lauroylglycinate (C12GlyNa) and Sodium

N-Lauroylsarcosinate (C12SarcNa)a

surfactant

CMC byconducti-vity (mM)

CMC bytensiome-try (mM)

surfaceexcess, Γ

(mol/1000m2)

area permolecule(A2)

ionizationdegree,

R

C12GlyNa 14 11 2.75� 10-3 60 0.45C12SarcNa 13 9.5 2.24� 10-3 74 0.53

aThe pH of the solutions of C12GlyNa and C12SarcNa is the same(pH = 8.5).

Figure 2. Change of Krafft temperature as the molar ratio of HClto surfactant is varied.

(14) Holmberg, K.; J€onsson, B.; Kronberg, B.; Lindman, B. Surfactants andPolymers in Aqueous Solution, 2nd ed.; Wiley: Chichester, 2003; p 44.(15) Rosen, M. J. Surfactants and Interfacial Phenomena., 3rd ed.; Wiley:

New York, 2004; p 121.(16) Klevens, H. J. Am. Oil Chem. Soc. 1953, 30, 74–80.(17) Rosen, M. J. Surfactants and Interfacial Phenomena, 3rd ed.; Wiley:

New York, 2004; p 60-64.(18) Menger, F. M.; Shi, L.; Rizvi, S. A. A. J. Am. Chem. Soc. 2009, 131,

10380-10381.(19) Tsubone, K.; Rosen, M. J. J. Colloid Interface Sci. 2001, 244, 394–398.

(20) Zana, R.; Benrraou, M.; Rueff, R. Langmuir 1991, 7, 1072–1075.(21) Menger, F. M.; Keiper, J. S.; Mbadugha, B. N. A.; Caran, K. L.; Romsted,

L. S. Langmuir 2000, 16, 9095–9098.(22) Wettig, S. D.; Verrall, R. E. J. Colloid Interface Sci. 2001, 235, 310–316.(23) Hirata, H.; Hattori, N.; Ishida,M.; Okabayashi, H.; Frusaka,M.; Zana, R.

J. Phys. Chem. 1995, 99, 17778–17784.

3080 DOI: 10.1021/la902979m Langmuir 2010, 26(5), 3077–3083

Article Bordes et al.

of only small amounts of HCl results in a Krafft temperatureabove room temperature. The sarcosine-based surfactant behavesdifferently. The Krafft temperature remains below 0 �C until aHCl to surfactant ratio of above 0.5 is reached. After that, theKrafft temperature immediately exceeds 100 �C (expressed in thefigure as a Krafft temperature of 100 �C.)

Addition ofHCl leads to protonation of the carboxylate group.Thus, a mixture of the anionic surfactant and its correspondingacid is generated and the ratio of anionic surfactant to aciddecreases linearly with the HCl to surfactant ratio. For normallong-chain alkyl carboxylates, such a mixture is often referred toas an acid-soap and it is well-known that several properties thatrelate to the efficiencyof surfactant adsorptionare atmaximumata 1:1 mixture of the protonated and the unprotonated species.24

The self-assembled amphiphiles are believed to align with alter-nate packing of the two moieties, enabling hydrogen bondsbetween protonated (donor) and unprotonated (acceptor) head-groups and leading to tight packing. Further addition of acidusually leads to precipitation. The behavior of C12SarcNa issimilar to that of a normal alkyl carboxylate in this respect. Thus,the effect of pH on the solution behavior is not affected by thepresence of the N-methylated amide bond between the hydro-phobic tail and the polar headgroup. C12GlyNa, on the otherhand, forms a precipitate as soon as the acid-soap mixture isgenerated. This indicates that intermolecular hydrogen bondinginvolving the amide bond favors the formation of acid-soapcrystals. It is interesting that the behavior of the 1:1 mixture ofC12GlyNa and C12SarcNa resembles that of C12GlyNa only. Itseems that the effect of hydrogen bonds involving nonmethylatedamide linkages is decisive of the tendency for precipitation also forthe surfactant mixture.

NMRwas used to quantify the precipitation at 25 �Cof the twoamphiphiles for increasing ratios ofHCl to surfactant.NMR is anadequate tool to monitor surfactant precipitation because it onlyrecords the dissolved species. Acetonitrile was used as an internalstandard to quantitatively determine surfactants in solution. Theamount of surfactant in solution is determined from the ratio ofthe integrals from the peaks from the surfactant and the peakfrom the acetonitrile methyl group. The integrals of the 1H peaksfrom the surfactants were used rather than the intensity of thepeaks. The reason for this was tominimize the problems related tothe decrease in relaxation time that can occur for surfactants inthe aggregated state. Such a decrease in relaxation time maybroaden the signal and reduce the intensity. Measurements weredone below and above the CMC, and the results are presented inFigure 3.

The figure shows that for both surfactants the tendency toprecipitate is larger when the surfactant is at a concentrationbelow the CMC than when present above the CMC. This isobviously due to the favorable alternate packing of protonatedand unprotonated species inmicelles and other self-assemblies, asdiscussed above.

The figure also illustrates the difference between the twosurfactants in resistance to precipitation on addition of acid thatwas also evident from the Krafft point values of Figure 2.Micellized C12SarcNa has much less tendency than micellizedC12GlyNa to precipitate at 25 �C in spite of the fact that the extramethyl group on the former should make this surfactant morehydrophobic, at least in a formal sense. The results are anotherillustration of the role of intermolecular hydrogen bonding of theglycine-based surfactant.

The effect of the pH on the CMC and on the surface tension atthe CMC for C12SarcNa is shown in Figure 4. (Due to thetendency for precipitation, the corresponding experiments withC12GlyNa could not be performed.) The effect of pH on CMC isaccording to expectations. The gradual protonation of the car-boxylate group below pH 8 leads to a reduction in hydrophilicityand to a lowering of the CMC. The minimum in surface tensionseen at pH 7 most likely reflects a perfect alternate packing ofprotonated and deprotonated sarcosinate species. The extremelylow surface tension values obtained at the optimum pH is a goodillustration of how efficiently an acid-soap pair aligns at theair-water interface.3.2. Adsorption at Solid Surfaces. The adsorption study

was made with naked gold surfaces or with gold surfaces madehydrophobic by the self-assembled monolayer technique (SAM-CH3). Two complementary techniques were used to monitor theadsorption, QCM-D and SPR. QCM-D is a technique thatdirectly measures the amount of material adsorbed on a surface.The value obtained is the sum of adsorbed solute and solventbound to the adsorbed solute. This is often important whenstudying adsorption from aqueous solutions because the water of

Figure 3. Relative peak integral versus molar ratio of added HClto surfactant. The integrals refer to the sumofall 1Hpeaks fromthesurfactant, and the value of 1.0 corresponds to 100% of the addedsurfactant beingpresent indissolved form.Themeasurementsweremade at 25 �C.

Figure 4. CMC and surface tension at the plateau versus pH forC12SarcNa. pH was adjusted with hydrochloric acid and sodiumhydroxide.

(24) Meade, E. J. Am. Oil Chem. Soc. 1962, 39, 235–237.

DOI: 10.1021/la902979m 3081Langmuir 2010, 26(5), 3077–3083

Bordes et al. Article

hydration may constitute a large fraction of the total adsorbedamount. For surfactants, the polar headgroup, and in particularthe oligooxyethylene chain of nonionic surfactants, is known tobring along a considerable amount ofwater on adsorption. SPR isan optical technique that monitors the adsorption as a change inrefractive index close to the surface. The refractive index change isproportional to the mass of substance adsorbed. SPR onlyrecords adsorption of the solute; adsorbed solvent is not includedin the values. Thus, by combining QCM-D and SPR in studies ofadsorption from a water solution, one can at least semiquantita-tively assess also the amount ofwater of hydration in the adsorbedlayer.

Prior to use, the gold surfaces were thoroughly cleaned andthe contact angle was determined as a way to characterize thesurface. The value obtained for the contact angle was 88 ( 3�,which is in accordance with literature values for gold. Thisvalue indicates a rather hydrophobic surface. In the literature,there is some ambiguity with respect to the nature of the surfaceof gold. Gold kept under ultrahigh vacuum has been reportedto be hydrophilic.25 Under normal conditions, however, goldmust be regarded to be hydrophobic. It is known that a goldsurface can bind halides, in particular bromide and iodide,strongly by chemisorption.26,27 This gives the surface a nega-tive charge and triggers adsorption of cationic amphiphiles,which adsorb as micelles or as a double layer. Cationicsurfactants with a halide as counterion have been found togive this adsorption pattern, while cationic surfactants withhydroxyl as counterion adsorb as a monolayer.28 Anionicsurfactants have been shown to adsorb at the surface withthe tail down and the polar headgroup oriented away from thesurface, either as a monolayer or as hemimicelles.28 This is thenormal mode of self-assembly of an amphiphile on a relativelyhydrophobic surface.

The hydrophobic SAM-CH3 surface was prepared by immer-sing the gold surface in a solution of alkylmercaptan in ethanol,a procedure known to anchor the solute to the surface byformation of a gold-sulfur bond and to give a tightly packedmonolayer.29,30 An alkanethiol of 16 carbon atoms was used inorder to ensure formation of a stable and homogeneous mono-layer. Prior to use, the contact angle at the surfacewas determinedand a value of 110( 3� confirmed the formation of a hydrophobicfilm on the surface.

QCM-D. As described in detail elsewhere,31 QCM-D is agravimetric method that directly senses the amount of materialadsorbed at the surface through the change of the oscillationfrequency at different overtones, and the loss of energy due to theadsorbed layer by measuring the dissipation.

The variation of frequency was monitored at the third, fifth,and seventh overtone. The values obtained were within 5%deviation when normalized versus the overtone number, whichsuggests an adsorption without slipping on the surface. Thisallows a direct determination of the mass adsorbed using theSauerbrey relation, assuming homogeneous coverage on thesurface and the amount adsorbed being less than 2% of the mass

of the crystal:32

Δm ¼ -ΔfC

n

where Δm is the adsorbed mass, Δf is the observed variation offrequency at overtone n, andC=17.7 ng 3 cm

-23Hz-1, a constant

characteristic of the equipment.The dissipation was also monitored during the measurements.

It corresponds to the decay of the intensity of the oscillationversus time and is due to loss of energy. The dissipation can bedefined as

D ¼ Edissipated

2πEstored

where Edissipated and Estored are the dissipated and the storedenergy of the adsorbed layer, respectively. Adsorption at thesurface will add to the total mass, thus giving a decrease ofthe oscillation frequency. The adsorption will also increase thedissipation factor due to an increase of the viscoelasticity of theformed film.

The baseline was first recorded in pure water, and thensolutions of surfactant of increasing concentrations were injectedinto the measurement cell. Figure 5 shows the stepwise change infrequency that occurs during an experiment. Each step downwardindicates injection of a surfactant solution of increasing concen-tration. The peaks on the curve are due to slight variations of thetemperature and the pressure in the measurement cell when theinjection occurs and can be disregarded. Finally, water wasinjected in the measurement cell to rinse the surface. Bothsurfactants were readily desorbed during the rinsing. This ap-proach of replacing one surfactant solutionby another solutionofhigher concentration was preferred instead of the commonly usedapproach of rinsing between each surfactant solution.

It can be seen fromFigure 5 that the adsorption is fast. One canalso see that, at a certain concentration, corresponding to thethird addition, the surface seems to be saturated. The fourth andfifth addition (corresponding to the fourth and fifth peak) do not

Figure 5. QCM-D measurements of adsorption at 20 �C ofC12GlyNa at a gold surface using successively higher surfactantconcentrations followed by rinsing with water. The change infrequency (Δf) corresponds to the third overtone.

(25) Smith, T. J. Colloid Interface Sci. 1980, 75, 51–55.(26) Tao, N. J.; Lindsay, S. M. J. Phys. Chem. 1992, 96, 5213–5217.(27) Wandlowski, T.; Wang, J. X.; Magnussen, O. M.; Ocko, B. M. J. Phys.

Chem. 1996, 100, 10277–10287.(28) Jaschke, M.; Butt, H. J.; Gaub, H. E.; Manne, S. Langmuir 1997, 13, 1381–

1384.(29) Sigal, G. B.; Mrksich, M.; Whitesides, G. M. Langmuir 1997, 13, 2749–

2755.(30) Bain, C. D.; Evall, J.; Whitesides, G.M. J. Am. Chem. Soc. 1989, 111, 7155–

7164.(31) Tehrani-Bagha, A. R.; Holmberg, K. Langmuir 2008, 24, 6140–6145. (32) Sauerbrey, G. Z. Phys. 1955, 155, 206.

3082 DOI: 10.1021/la902979m Langmuir 2010, 26(5), 3077–3083

Article Bordes et al.

give a change in frequency. It is also clear from the curve thatdesorption on rinsing with water is rapid.

During the course of the adsorption, the dissipation firstincreased and then remained constant around 2.5 � 10-6,representative of a relatively rigid layer. This also demonstratesthat the oscillations of the adsorbed layer are not decoupled fromthe surface, which confirms that the Sauerbrey equation can beused for direct determination of the mass.

Figure 6 shows adsorption at the hydrophobic SAM-CH3

surface at a surfactant concentration where plateau adsorptionhas been reached, that is, corresponding to the third addition inFigure 5. A monolayer adsorption is assumed and the surfaceshould be saturated at the CMC. The continued rise of the curvesat concentrations beyond the CMC is not related to adsorptionbut is a bulk effect caused by an increase in the viscosity.The quartz crystal is very sensitive to small changes in bulkviscosity.33,34 At concentrations corresponding to the CMC foreach surfactant, C12GlyNa adsorbs slightly more than C12Sarc-Na. This is in accordance with the tighter monolayer packing dueto intermolecular hydrogen bonding of C12GlyNa as comparedto C12SarcNa, as was discussed above for the situation at theair-water interface.

The adsorption pattern on gold is very different. Figure 7shows that at the CMCC12SarcNa gives approximately twice theadsorbed amount as C12GlyNa.

One can also see that surface saturation is achieved at lowersurfactant concentration for C12GlyNa than for C12SarcNadespite the fact that the sarcosine-based surfactant has a slightlylowerCMC than the glycine-based amphiphile. The results will befurther discussed in connection to the results from the SPRexperiments.

SPR. In the SPR technique, monochromatic, p-polarized lightis directed through a prism and a glass support coated with a thinlayer of gold in contact with the solution in the flow cell. Lightdirected above a critical angle of the p-polarized incident light willbe totally reflected at the gold-prism interface, which will causean evanescent field topenetrate into the gold film.This evanescentfield can couple to an electromagnetic surfacewavewhich is calleda surface plasmon. The surface plasmon is excited at the gold-liquid interface and is measured by photodiodes as a minimum inreflected light. As surfactants adsorb at the gold surface, therefractive index will be altered, and the conditions for SPRare changed, which is monitored as a change in the angle of theminimum intensity reflected light. This change of angle is in theBIAcore terminology expressed in resonance units (RUs), with 1RU being equal to 0.0001� change in the intensity minimum. Thechange in angle of reflected light,measuredasRU, is related to themass adsorbed at the surface, ΔmSPR, according to

Δm ¼ ΔRUCSPR

β

whereCSPR is a factor containing an instrument constant and dn/dc (the variation of refractive index with concentration) of theadsorbent and β is a factor compensating for the decrease in SPRsignal with distance from the gold substrate.CSPR was calculatedto 0.094 ( 0.008 ng/cm2 using an average dn/dc for 18 differentsurfactants, and β was set to 1, which is the case for a plain goldsurface.35 The factor β will differ from 1 when the surface layer isthick. For instance, a 30 nm thick SiO2 layer on a gold substrategives a β value of 0.7. In the experiments performed in this work,the layer is very thin and the value of β can be anticipated to beclose to 1.

It is important to realize that the change in refractive indexmeasured by the SPR technique can be caused not only byadsorption of a solute at the gold surface but also by variationsof the refractive indexof the solution in close proximity (within ca.200 nm) from the surface. When surfactants are adsorbed fromsolutions of high concentrations, the change of the refractiveindex of the solutionwill give a substantial contribution to theRUvalue recorded.29 In this work, the surfactant concentrations havenot been very high, however, and we believe that the RU valuesrecorded are almost entirely due to the adsorption process.

The values of the adsorbed amount at the CMC for C12GlyNaand C12SarcNa are given in Table 2. For comparison, the valuesobtained from the QCM-D measurements are also included. Itcan be seen that the values of the adsorbed amount are lower from

Figure 6. Mass adsorbed on a SAM-CH3 surface determined byQCM-D at the third overtone at 20 �C.

Figure 7. Mass adsorbed on gold determined by QCM-D at thethird overtone at 20 �C.

Table 2. Adsorbed Amount of C12GlyNa and C12SarcNa on Gold

Determined by QCM-D and SPR

surfactantadsorbed amount byQCM-D (ng/cm2)

adsorbed amount bySPR (ng/cm2)

C12GlyNa 215 91-142C12SarcNa 417 131-162

(33) Keiji Kanazawa, K.; Gordon, J. G. Anal. Chim. Acta 1985, 175, 99–105.(34) Fawcett, N. C.; Craven, R. D.; Zhang, P.; Evans, J. A. Anal. Chem. 1998,

70, 2876–2880. (35) Larsson, C.; Rodahl, M.; Hook, F. Anal. Chem. 2003, 75, 5080–5087.

DOI: 10.1021/la902979m 3083Langmuir 2010, 26(5), 3077–3083

Bordes et al. Article

SPR than from QCM. This is often the case and is usuallyattributed to QCM taking into account also water of hydration,which SPR does not do. In addition, Table 2 shows that also SPRgives a higher adsorbed amount for the sarcosine-based surfac-tant than for the glycine analogue although the difference is not aslarge as with the QCM technique.

Although the values differ, both techniques that we have usedto measure the adsorbed amount of surfactant at the naked goldsurface indicate that C12SarcNa adsorbs much more thanC12GlyNa. With the QCM technique, there was almost a 2:1ratio, and with SPR the ratio was of the order of 1.5:1. This isunexpected considering that C12GlyNa adsorbed slightly morethan C12SarcNa on the hydrophobic SAM-CH3 surface. Thehigher adsorbed amount of C12GlyNa at the SAM-CH3 surface(and the better packing at the air-water interface) was attri-buted to a contribution by hydrogen bonds involving theamide group to the intermolecular interactions. Why do we notsee that effect whenwe compare adsorption of the two surfactantsat the untreated gold surface? A tentative explanation is givenbelow.

The sarcosine headgroup is known to interact strongly withmany metal surfaces, an observation that has practical implica-tions. For example, dodecanoyl sarcosinate is a very usefulinhibitor of corrosion of steel.36 IR investigations have shownthat the surface metal atom is chelated by carboxylate oxygenatoms and by the amide nitrogen atom. Surfactants based on thesarcosine headgroup are also common as collectors in mineralore flotation.37 A well-known example is reduction of the phos-phorus content in iron ores by removing apatite by flotation.38

Sarcosinate-based amphiphiles are stronger chelating agents thanglycine-based surfactants, most likely due to the N-methyl groupimparting an increased electron density on the nitrogen. Thestrong binding of the sarcosine headgroup to the surface can giverise to double layer adsorption. We postulate that we have thesame phenomenon on the gold surface. Due to strong head-group-surface interaction, a double layer is formed withC12SarcNa. The glycine headgroup does not interact so stronglywith the surface, and C12GlyNa gives the normal monolayeradsorption. The reason why C12SarcNa gives less than twice theadsorbed amount compared toC12GlyNamay be that the lack ofintermolecular hydrogen bonding for C12SarcNa makes thesurfactant packing in each layer less tight than the packing inthe C12GlyNa monolayer.

Calculation of area per molecule from the SPR data givesvalues of around 40 A2 and around 30 A2 for C12GlyNa andC12SarcNa, respectively, if monolayer adsorption is assumed. Ifinstead monolayer adsorption is assumed for C12GlyNa andbilayer adsorption for C12SarcNa, the values would be 40 A2 forthe glycinate surfactant and 60 A2 for the sarcosinate, which seemmore reasonable.

4. Conclusions

In this work, the role of the amide bond in self-assemblyprocesses was assessed by studying two simple N-acyl amino acidsurfactants: C12GlyNa that can form intermolecular hydrogenbonds and C12SarcNa that lacks this possibility.

The CMC was approximately the same for the two surfactantsand one order of magnitude higher than that of a normal alkylcarboxylate with the same tail length, which showed that the amidebond,N-methylatedornot, increased the solubility of the surfactant.

The area occupiedby each surfactant at the air-water interfacewas considerably smaller for C12GlyNa than for C12SarcNa, 60versus 74 A2. The tighter packing of C12GlyNa can be attributedto intermolecular hydrogen bonds, the formation of a “hydrogenbelt”. This effect is also seen in the tendency to precipitate onaddition of acid. Whereas C12SarcNa behaved like normal alkylcarboxylates, forming an acid-soap pair that gave extremely lowsurface tension values, C12GlyNa precipitated already at smalladdition of acid. We interpret this increased tendency for theacid-soap pair to precipitate to the additional intermolecularinteraction caused by the hydrogen bonds. The increased inter-molecular interaction of C12GlyNa as compared to C12SarcNacould also be seen in adsorption studies performed with theQCM-D technique. At the CMC, C12GlyNa gave higheradsorbed amount than C12SarcNa at a hydrophobic surface.

Adsorption experiments at a gold surface, performedbyQCM-D and SPR, gave unexpected results. The sarcosinate-basedsurfactant adsorbed 1.5-2 times more than the glycine-basedcounterpart. We interpret this as bilayer adsorption for C12Sarc-Na and monolayer adsorption for C12GlyNa. Amphiphilicsarcosine derivatives are known to form chelates to metal sur-faces, and such a strong interaction between the polar headgroupand the surface may induce formation of a bilayer. Glycinederivatives do not form as strong chelates as sarcosinates, andthe glycine-based surfactant therefore prefers to adsorb with thetails down at the relatively hydrophobic gold surface.

Supporting Information Available: NMR data of C12Gly-Na. This material is available free of charge via the Internetat http://pubs.acs.org.

(36) Salensky, G. A.; Cobb, M. G.; Everhart, D. S. Ind. Eng. Chem. Prod. Res.Dev. 1986, 25, 133–140.(37) Ranjbar, M.; Schaffie, M.; Ranjbar, M. World Metall.;Erzmet. 2006, 59,

34–39.(38) Hellsten, M.; Ernstsson, M.; Idstr€om, B. U.S. Patent 4732667, 1988.