DOI: 10.1002/cphc.200500369

Compression-Inhibited Pore Formation ofPolyelectrolyte Multilayers Containing WeakPolyanions: A Scanning Force Microscopy StudyBo Wang,[a] Lili Liu,[a] Ke Chen,[b] Lin Chen,[a] Jie Feng,[a] and Changyou Gao*[a]

1. Introduction

Over the past decades, electrostatic layer-by-layer (LbL) assem-bled multilayers have attracted much interest.[1] In the deposi-tion process, oppositely charged species are adsorbed on asubstrate sequentially. This simple method has several advan-tages over other surface modification approaches. Firstly, it isquite convenient to design, modify, and functionalize the surfa-ces without involvement of harsh conditions. Extensive studieshave been carried out to introduce small organic molecules orinorganic compounds, bio-macromolecules, dendrimers, andcolloids or latex particles into the layered architectures.[2] Com-pared with physical adsorption, more substances can be de-posited in the multilayers. Secondly, both the molecular organ-ization and the properties of the multilayers can be tunedeasily by assembly conditions or post-treatments, such as saltconcentration and pH value. The most typical example is thepH-controlled assembly of weak polyelectrolytes, originally de-veloped by Rubner and co-workers.[3] The substance encapsu-lation in microcapsules, with polyelectrolyte multilayers (PEMs)as walls, is also established based on tunable permeability inresponse to stimuli.[4] Thirdly, the LbL assembly is a low-cost,environment-friendly, wet-bench technique that can be em-ployed in industrial processes. CIBA-Vision has announced thefirst commercially available product that is equipped with amultilayer coating in 2002.[1] Moreover, all proteins and mostpolysaccharides are polyelectrolytes as well. Studies of thePEMs also provide meaningful information to understand bio-macromolecules adsorbed on surfaces and interfaces. Finally,as the thickness of the multilayers can be varied from nanome-ter to micrometer scales, it is possible to fabricate three-dimen-sional structures on surfaces by the LbL technique. Patternedassembly[5] and compression[6] are two typical examples.

Recently, we have reported the irreversible compression ofpolyelectrolyte multilayers assembled with poly(diallyldimeth-

ylammonium chloride) (PDADMAC) by pressing a poly(dimeth-ylsiloxane) (PDMS) stamp against the films.[6] The compressionextent varies from several to hundreds of nanometers showinga positive correlation with the multilayer thickness, which cannormally be tuned by the layer number or salt concentration.The effect of the desiccation degree on the ability of the com-pression illustrates that the structural water content plays animportant role for the intrinsic properties of PEMs. The numberof water molecules associated with one ion pair in a polyan-ion/polycation complex has been measured independently byDecher[7] and Schlenoff.[8] The average water content inside thefully hydrated films is larger than 40% by volume. Consequent-ly, the film thickness is reduced by 30% after drying.[7]

Small hydrophilic dye molecules are further used to studythe volumetric behavior of the compressed PEMs.[9] The diffu-sion coefficient of the probe decreases dramatically and losesthe sensitivity to salt addition after compression.[10,11] It is un-derstandable that the compression leads to shrinkage of thevacancies in the PEMs. Here, the concept of free volume canbe applied to some extent.[12]

Herein, the effect of the compression on pore formation ofthe poly(acrylic acid) (PAA)/PDADMAC multilayers will be re-ported. As visualized and analyzed by scanning force microsco-py (SFM), pores with a diameter of microns in the control PAA/PDADMAC multilayers were identified as a result of pH varia-

[a] B. Wang, L. Liu, L. Chen, Dr. J. Feng, Prof. Dr. C. GaoDepartment of Polymer Science and Engineering, Zhejiang UniversityHangzhou 310027 (China)Fax: (+86) 571-87951948E-mail : cygao@mail.hz.zj.cn

[b] K. ChenDepartment of Radio Engineering, Southeast UniversityNanjing 210096 (China)

Morphological changes of poly(acrylic acid)/poly(diallyldimeth-ylammonium chloride) multilayers induced by low pH were inves-tigated by scanning force microscopy. The weakened interactionbetween the charged polymer chains in the protonation processis believed to be the reason for this variation. Kinetic studieshave shown that during protonation phase separation and disso-ciation of the multilayers took place successively. The compres-sion of the multilayers, however, caused a transition of the multi-

layers from a rubbery state to a glassy state. As a result, theclosely compacted multilayers lost their sensitivity to pH change.An increase of electrostatic and hydrophobic interactions, can de-crease the free energy of the multilayers, and stabilize the films.By compression of the multilayers with a rubber stamp havinggeometric patterns, films with spatially localized pores were pro-duced.

590 B 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ChemPhysChem 2006, 7, 590 – 596

tion. However, the compression of the multilayer film com-pletely inhibits the emergence of the pores. It shall be demon-strated that the compression induces a transition from a rub-bery state to a glassy state by desolvation. A thermodynamicanalysis based on a free energy model for compression hasbeen adopted to explain the results accordingly.

2. Materials and Methods

2.1 Materials

Polyethyleneimine (Aldrich, PEI, Mn 60000, Mw 750000), PDAD-MAC (Aldrich, Mw 200000–350000) and PAA (Fluka, Mw 20000,if not otherwise indicated) were used as received to prepare10�4

m aqueous solutions, supplemented with 0.3m sodiumchloride (99+ %). The pH value of the solution was adjustedto 11 by adding trace amounts of sodium hydroxide which didnot affect the ionic strength. Under these assembly conditions,thickest multilayers were obtained.[13] Poly(dimethylsiloxane)(PDMS) prepolymers were obtained from Dow Corning, Sylgard184. Silicon wafers were cleaned in 70% H2SO4 (concentrated)/30% H2O2(aq.) (“piranha”, caution: piranha is a strong oxidizerand should not be stored in a closed container) and then in hotH2O2/ammonia/water, 1:1:5 (vol/vol), rinsed in water, and driedin a nitrogen stream.

2.2 Methods

Assembly

To obtain a homogeneous growth of the multilayers, a precur-sor layer of PEI was deposited on the pre-treated substrate. Se-quential adsorption of the polyelectrolytes was performed bymanual dipping. Between alternate exposures to two kinds ofpolymer solutions for 20 min, there were 3 rinses, each for3 min. Then the films were dried under the desired relative hu-midity for 12 h.

Compression

A PDMS stamp with holes each measuring 50 mm in diameter,10 mm in space and 4 mm in height, was molded from litho-graphically prepared masters.[14] The stamp was then pressedon the multilayers with a normalized pressure of 200 gcm�2.The pressing time was 20 min.

SFM Imaging

Topographic images as well as phase images were collected si-multaneously using a SPI3800N Probe Station and a SPA400SPM Unit (Seiko Instruments Inc.) in dynamic force mode. Sili-con tips with a resonance frequency f0 of 150 kHz and a springconstant of 20Nm�1 were utilized. The scanning frequency was0.5 Hz. The contact force between the tip and the samples waskept as low as possible (<2.5 nN).

Film Thickness Measurements

To measure the film thickness, parts of the multilayers werescratched off using a razor blade. The SFM probe tip was thenscanned across the newly exposed regions to verify that verylittle, if any, material was left. The height of the film above thesubstrate was thus obtained directly. To verify the reliability ofthe method, a fresh silicon wafer was similarly scratched andexamined.[15] No visible scratch or damage was detected. Thethickness measurement of PEI(PAA/PDADMAC)7 is shown inFigure 1 as an example.

Lateral Force Measurements

To investigate the relaxation behavior of the multilayer surface,frictional force measurements were carried out at 20 8C in aclosed chamber with certain humidity. The lateral force curvewas obtained in contact force mode under a repulsive force of25 nN. A triangle 100 mm cantilever with a force constant of0.09 Nm�1 and an integrated silicon nitride (Si3N4) tip were ap-plied. The magnitude of the lateral force was evaluated withthe line scan mode. The scan rate is equal to 2 K (scan length)K (scan frequency). Six friction loop cycles were averaged forevery measurement to improve the signal-to-noise ratio.[16]

Surface Spectrum

The topographic images were transformed into digital data byFIS ASCII converte (Seiko Instruments Inc.).

The “height spectroscopy” data, that is, the histograms ofthe number of image pixels in incremental height intervalswere calculated by SPSS (Statistical Program for Social Scien-ces).

Figure 1. a) The thickness of PEI(PAA/PDADMAC)7 was measured by SFMafter scratching off all the polyelectrolyte multilayers from one area. The rec-tangular box depicts the area where the average cross-section (indicated bythe dashed line a) was taken along the direction of the arrow. b) Theaverage cross-section which is more representative than a simple linearcross-section, since every point in the profile is an average of all the pointson the dashed line. It is worth noting that the high ridge with a width of10 mm at the step edge should be ignored for thickness measurements,since it is the buildup of materials which was removed from the substrate.Because neither the film surface nor the scratched region was perfectly flat,the film thickness was averaged from several samples.

ChemPhysChem 2006, 7, 590 – 596 B 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.chemphyschem.org 591

Pore Formation of Polyelectrolyte Multilayers

Based on the analysis in ref. [17] , power spectral density(PSD) functions were determined as follows by a homemadecomputer program. A two-dimensional fast Fourier transforma-tion algorithm was applied to the SFM image data, followedby calculations of the squares to obtain density values. Thenall the values were normalized to the scanned area. The math-ematical description of this process is presented in Equa-tion (1):

I fx; fy

� �¼ 1

L2

XN

m¼1

XN

n¼1

zmne�2piDL fx mþfy nð Þ DLð Þ2

" #2

ð1Þ

where I denotes the two-dimensional PSD, L2 is the scannedarea, N is the number of data points per line and row, zmn isthe height value at position (x=m, y=n), fx and fy are the spa-tial frequencies in the x and y directions, and DL is the sam-pling distance. The transformation was followed by a transitionto polar coordinates for the frequencies. The final PSD functionwas obtained by angular averaging [Eq. (2)]:

I fð Þ ¼ 12p

Z2p

0

I f ;fð Þ ð2Þ

3. Results and Discussion

At an assembly condition of 0.3m NaCl and pH 11, PEI(PAA/PDADMAC)7 multilayers exhibit flat and featureless surfacemorphology as visualized by SFM (Figure 2a). The thickness ofthe films is on the order of 64�5 nm (Figure 1). Since the ioni-

zation degree of PAA is pH sensitive, it should, in principle, bepossible to modulate electrostatic interaction of the PAA/PDADMAC chains by controlling the solution pH. By immersingthe film assembled at pH 11 into a solution of pH 5 for 3 min,large granular features with dimensions from several tens tohundreds of nanometers were formed (Figure 2b). Moreover,retuse pits with depths of nanometers and diameters ofmicron scale were observed. However, the flat regions still

kept their original thickness without detectable variations. Byprolonging the time to 15 min, a highly porous microstructurewas observed as depicted in Figure 2c. The length scale of thepores is 1-5 mm with a characteristic depth around 50 nm,which is only slightly lower than the original film thickness. Onthe other hand, the thickness of the protuberant parts remainsalmost as constant as the original PEM thickness. When thetreatment time was extended to 30 min, a vermiculate surfacemorphology was observed (Figure 2d). This interesting mor-phology has been reported for poly(4-styrensulfonic acid)(PSS)/PDADMAC multilayers assembled in high salt concentra-tions, and is probably caused by aggregation of polymers.[15]

However, in our case, the thickness of the film decreased to17�3 nm, only �25% of the original thickness. This observa-tion is consistent with the ellipsometry results measured bySchlenoff and co-workers.[13] Even though the possibility ofsome shrinkage of the multilayers cannot be ruled out abso-lutely, the loss of the polyelectrolytes at low pH values wasconfirmed by this series of measurements.

This process can be analyzed via the PSD function (Figure 2)as well. The PSD gives the power of each of these constitutedwaveforms, and hence, allows the analysis of the most fre-quent wavelength and conveys the greatest influence to thesurface topography. The spatial frequency is the inverse of thein-plane spatial wavelength of the roughness features. Themost interesting result is that the power of the surface, whichwas treated with a pH 5 solution for 15 min (Figure 2c), con-centrates in the low-frequency region, indicating that the sur-face presents a periodicity to some extent. The intensity at ahigh spatial frequency increased upon acidic solution treat-ment (Figure 2b). This shows quantitatively that the roughnessincreases dramatically, mainly due to random features of verysmall length scale. However, the PSD functions retained the ta-pered profiles regardless of the duration of the treatment,which is characteristic of isotropic surfaces. These plots confirmthat the images consist of a composite of multiple waveforms.

By magnifying the pore structure (Figure 2c) as shown inFigure 3, the remaining polyelectrolytes at the bottom of thepores showed up clearly. Furthermore, the morphology at thepore bottom was not uniform. Instead, some small pits werefound as indicated by the arrows in Figure 3b, which were im-printed by the released polyelectrolyte complexes. The heightspectrum shows that pores occupy �20% of the surface area(Figure 3c). This pore structure is quite interesting and regular/reproducible. In the sections below, it is considered as a meas-ure of the pH sensitivity of the multilayers.

These results indicate that in acidic solution the change ofthe multilayer morphology develops gradually from nanometerscaled granular features to micrometer-scaled pores, while thethickness of the multilayers (unchanged regions) remainsrather constant. As for the transition of PAA/poly(allylamine hy-drochloride) (PAH) multilayers reported by Rubner and co-workers,[18] the reason for the morphological changes of thePAA/PDADMAC multilayers described here should be attribut-ed to the chain rearrangement and/or the phase separation(formation of granular features),[19] followed by dissolution ofthe polyelectrolytes. The whole process should be caused by a

Figure 2. PSD functions and SFM topographical images of PEI(PAA/PDAD-MAC)7 multilayers assembled at pH 11, 0.3m NaCl. a) As-prepared films aftertreatment in a pH 5 solution for b) 3, c) 15 and d) 30 min. The scanningareas are 15 mmK15 mm.

592 www.chemphyschem.org B 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ChemPhysChem 2006, 7, 590 – 596

C. Gao et al.

decrease in the degree of ionization. Following the proceduredescribed by Granick and co-workers,[20] the degree of ioniza-tion of the carboxylic groups in the PEMs was measured andactually decreased from 100 to 40% after acidic treatment for15 min, which is consistent with the results reported byRubner and co-workers.[21]

Basically, below a critical charge density, the enthalpicenergy required to maintain the chains in a more extended flatconformation as in the multilayers is no longer able to over-come the entropic penalty associated with this conformationalstate.[21,22] The destruction of the multilayers may actually yieldvery loosely associated complexes. These complexes have in-sufficient structural integrity to maintain a layer but do notfully dissociate. Classical polymer adsorption theory also sug-gests a small loss of conformational entropy (of order 0.2kTper segment) when a polymer adsorbs on a surface.[13] A com-plex in solution might be a little more stable than a complexwithin a multilayer film since an additional degree of freedomis afforded in the former. Moreover, the formation of a morehydrophobic hydrogen bonded PAA complex in acidic environ-ment can expedite the process of forming the final structure.[23]

Additionally, accommodating excess positive charge withinmultilayers by rearrangement into an elevated surface area isthe reason for the rougher surface, such as pores and a vermic-ulate morphology.[24]

It is worth mentioning that very thick films can in fact be ob-tained when PAA/PDADMAC multilayers are assembled atpH 5.[21] However, dissolution of the film with a pH 5 solution

during post-treatment was observed in our study. Furthermore,only a very thin PAA/PDADMAC film was formed under suchconditions, which is consistent with results from Schlenoff andco-workers.[13] This inconsistency might be caused by the dif-ferent molecular weight of PAA adopted in these studies. Infact, we assembled a film using PAA with higher molecularweight (Mw�100000, Sigma–Aldrich), no such transition wasobserved after the same treatment. The cooperative chargecoupling effect is crucial for multilayer formation and stabili-ty.[25] For example, longer polyelectrolytes are more tightlybound and thus restricted in their positions. Sukhishvili andco-workers also observed that, controlled by the phase dia-gram, smaller amounts of polyacids were adsorbed on a sur-face or precipitated in a polyelectrolyte complex when the mo-lecular chains were shorter, especially when the molecularweight of the oppositely charged polymer was large.[26] It isalso much easier for polymers of small molecular weight tomove and, as a result, to be stripped from the surface, becausefew attachment points exist on the shorter chains.[27] This mo-bility is quite important for our studies, which is the reasonwhy we chose such small polyacids.

Following the process we reported previously,[6] partially de-hydrated polyelectrolyte multilayers (dried under 75% relativehumidity) were compressed by a PDMS stamp with patternedfeatures (Figure 4). From the profile shown in Figure 4b, the

height difference between the compressed and the uncom-pressed regions was measured to be 45�4 nm. Hence, thecompression ratio is an estimated 75% of the original thick-ness. The remaining thickness of the multilayers after homoge-neous compression with a flat PDMS film has also been exam-ined by SFM using the scratching method (Figure 5). The re-

Figure 3. a) SFM topographical image showing the morphology of PEI(PAA/PDADMAC)7 multilayers after immersion in a pH 5 solution for 15 min. b)Magnified view of the region enclosed in the box in (a). c) Height spectrumof (a). The peaks 1 and 2 correspond to the concave and protuberant re-gions in (a), respectively. The peaks are fitted with a Gaussian function andthe areas are calculated correspondingly.

Figure 4. a) SFM topographical image and c) phase image showing the pat-tern-compressed (PAA/PDADMAC)7 multilayers. b) and d) are line profilescorresponding to the lines in (a) and (c), respectively.

ChemPhysChem 2006, 7, 590 – 596 B 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.chemphyschem.org 593

Pore Formation of Polyelectrolyte Multilayers

sulting height (�15 nm) measured from the profile (Figure 5b)matches roughly the remaining value of the compressed re-gions shown in Figure 4b (65 nm-45 nm). It is worth notingthat the dustlike particles in Figure 5a are residual materialsproduced during the scratching procedure. Compared with thecase shown in Figure 1 and the extremely flat compressed sur-face (Figure 4a),[6] one can conclude that the compressed mul-tilayers are more fragile and easily pulverized.

The compression has also been characterized by the phaseimage of SFM (Figure 4c). The phase lag between the reso-nance signal and the cantilever response is a parameter thatcontains relevant information about the interaction betweentip and sample. The topographic component has been separat-ed from material-dependent contributions to the phaseshift.[28] It has been proved that the topographic component ofthe phase lag is almost constant when the height changesgradually.[29] For elastic materials, the absolute value of thephase decreases from 908 to 08 as the cantilever approachesthe sample. This is a consequence of the resonance frequencydisplacement of the cantilever to higher frequencies due tothe tip–sample contact. The shift is smaller for softer samplesbecause the frequency change is smaller.[30] Variations of thecapillary force in samples with regions of different hydrophilic/hydrophobic properties could also be a source of phase con-trast.[31] The phase contrast profiles presented in Figure 4dshow that the compressed regions display harder and hydro-phobic properties, which is consistent with the results of othercharacterization approaches.[32] The deviating low absolutevalue of the phase lag on the boundaries should be attributedto the heterogeneous stress distribution in these regions.[9]

When the pattern-compressed multilayers were immersed ina pH 5 solution for 15 min, the porous structure emerged onlyin the uncompressed regions (Figures 6a and 6b for a magni-fied view). The length scale and the depth of the pores werequite consistent with the uncompressed films (Figure 6c forline profile). This locally porous film might find wide applica-tions in areas of controlled release, personal care, sensors, andfiltration.[33]

It is quite hard to imagine that the compression closes upvoids and channels in the polymer matrix that facilitate small

ion diffusion. As prepared, the polyelectrolyte multilayers con-tain no mobile small ions within their bulk.[34] Actually, uponcompression the oppositely charged polyelectrolyte chains canapproach each other much closer to produce a higher densityof ion cross-linking and physical entanglement, which in turncan effectively enhance their stability against chain rearrange-ment and/or phase separation.[35] A compact packing of themolecules might result in reduced diffusion. Consequently, thedissolution of the films needs longer time. Furthermore, uponcompression the positive charges approach the polyacids,which moves the barrier to ionization of a carboxylic acidgroup to lower pH than for the uncompressed multilay-ers.[20,21, 36] However, when we immersed these locally porousfilms in a pH 1 solution for 48 h, no change was found. Actual-ly, in contrast to the uncompressed films,[37] compressed oneswere not swollen in 2m NaCl.[6] All these proofs point out thatthe compressed multilayers become thermodynamically andnot just kinetically much more stable. The ionization degreealso remained as high as 82% after incubation in a pH 5 solu-tion for 2 h. Electrostatic couplings possibly provide a sufficientsource of stabilization energy (or rather enthalpy) to counter-balance unfavorable entropy changes. An increased hydropho-bic interaction in the compact polymer complex can be anoth-er source of stabilization energy as well.[21,38]

With stronger interactions among polymer chains, theenergy barrier of polymer rearrangement should becomemuch higher. The molecular surface motion can be measuredby the frictional force on the surface. A typical example of thelateral force loop is shown in Figure 7a. The frictional force,being nonconservative, is quantified as the difference of theaverage lateral force scanned in opposite directions.[16] Thedata at the extreme left and right ends of the scanning intervalcorrespond to the static friction region, where the tip has notyet attained a sliding state.[39] They should thus be neglectedin the calculation. Since the magnitude of the frictional force isalmost proportional to the loss modulus,[40] the frictional force

Figure 5. Thickness measurement of the homogeneously compressed film.The rectangular box in (a) depicts the area where the average cross-section(indicated by the dashed line a) is taken along the direction of the arrow.The average cross-section is shown in (b).

Figure 6. a) SFM topographical image showing the pattern-compressed(PAA/PDADMAC)7 multilayers after immersion in pH 5 solution for 15 min. b)Three-dimensional magnified image corresponding to the region enclosedin the box in (a). c) Line profile corresponding to the line in (a).

594 www.chemphyschem.org B 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ChemPhysChem 2006, 7, 590 – 596

C. Gao et al.

is larger at the glassy–rubbery transition than in either theglassy or rubbery state. The friction of the multilayers com-pletely dried under 15% relative humidity (Figure 7b) does notshow any frequency dependence (the frequency shows a posi-tive correlation with the scanning rate), illustrating that thefilm is in a glassy state. In contrast to this, as seen from thefrictional force curve displayed in Figure 7c, the multilayersdried under 75% relative humidity show a glass transition withan increase of the scanning rate. This demonstrates that thepartially dehydrated film is in the rubbery state. The differentbehaviors confirm the lower modulus of a wetting film com-pared with a fully dried one.[41] However, the rubbery film be-comes glassy after compression, as evidenced by the frequen-cy independent friction curve (Figure 7d). The distinct transi-tion from a rubbery to a glassy nature of the multilayers is fur-ther supported by the following experiment: small areas of theuncompressed and the compressed film (both dried under75% relative humidity) were scanned at a larger load force(100 nN). The uncompressed multilayers could be abraded me-chanically, while the compressed films could not. This phenom-enon is comparable with the result of Cohen-Stuart and co-workers for an ionic induced glass transition of multilayers.[42]

So far, the mechanism of the compression has been partiallyunveiled (Figure 8). In the process of compression, the watermolecules (left) which function as plasticizers in the partiallydehydrated films are squeezed out. Consequently, the polymerchains approach each other closer (right) and the free spacefor molecular motion is diminished. Moreover, due to the for-mation of more temporary ionic bonds and microhydrophobiccomplexes, the motions of the polymer chains are retarded tosome extent. Thus, a pressure induced glass transition has oc-

curred. This is the reason for the property variation, such asstability, permeability, stiffness, wettability and so on.

4. Conclusions

Driven by weakened interactions between charged polymerchains, low/high pH values induce morphological variation ofPAA/PDADMAC multilayers. The multilayers present granularfeatures, micropores, or a vermiculate morphology tuned bypH and incubation time. The kinetics shows that in the processof protonation phase separation and film dissociation occursuccessively. However, after compression the closely compact-ed multilayers lose their sensitivity to pH changes, and under-go a glass transition. It has been demonstrated that electro-static or hydrophobic interactions, can decrease the freeenergy of the multilayers, thus stabilizing the film. Based onthis study and our previous results, the molecular mechanismof compression is revealed. The locally microporous films mayfind potential applications in areas of controlled release, per-sonal care, sensors, and filtration.

Acknowledgments

We thank J. C. Shen and H. Mçhwald for their continuous sup-port and stimulating discussions. We also thank Z. Wang for hishelp for the statistical data analysis. J. L. Duda is greatly appreci-ated for his advice on free energy volume theory. G. Haugstad isgreatly appreciated for his advice on frictional force microscopy.This study is financially supported by the Natural Science Foun-dation of China (No. 20434030 and 90206006) and the NationalScience Fund for Distinguished Young Scholars (No. 50425311).

Keywords: compression · multilayers · polyelectrolytes ·scanning probe microscopy · self-assembly

[1] for a review, please see Multilayer Thin Films: Sequential Assembly ofNanocomposite Materials, (Eds. : G. Decher, J. B. Schlenoff), Wiley-VCH,2002.

[2] a) G. Mao, Y. Tsao, M. Tirrell, H. T. Davis, V. Hessel, H. Ringsdorf, Langmuir1993, 9, 3461; b) Y. Lvov, G. Decher, G. Sukhorukov, Macromolecules1993, 26, 5396; c) A. C. Fou, M. F. Rubner, Macromolecules 1995, 28,7115; d) D. L. Feldheim, K. C. Grabar, M. J. Natan, T. C. Mallouk, J. Am.

Figure 7. a) Lateral force measured on the multilayers (dried under 75% hu-midity) as a function of the tip position during a single left-to-right/right-to-left cycle (averaged from six independent measurements). The scanning rateis 6400 nms�1. Insets are intended to show the tip state in a scanning cycle.Frictional force versus scanning rate measured on b) uncompressed filmdried under 15% humidity, c) uncompressed film dried under 75% humidity,and d) compressed film dried under 75% humidity.

Figure 8. Schematic illustration of the molecular mechanism of compression.The conformation of the molecules has been optimized by Chem3D throughenergy minimization.

ChemPhysChem 2006, 7, 590 – 596 B 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.chemphyschem.org 595

Pore Formation of Polyelectrolyte Multilayers

Chem. Soc. 1996, 118, 7640; e) M. Olek, J. Ostrander, S. Jurga, H. Mçh-wald, N. Kotov, K. Kempa, M. Giersig, Nano Lett. 2004, 4, 1889.

[3] a) D. Yoo, S. S. Shiratori, M. F. Rubner, Macromolecules 1998, 31, 4309;b) S. S. Shiratori, M. F. Rubner, Macromolecules 2000, 33, 4213.

[4] a) L. DOhne, S. Leporatti, E. Donath, H. Mçhwald, J. Am. Chem. Soc.2001, 123, 5431; b) S. Leporatti, C. Y. Gao, A. Voigt, E. Donath, H. Mçh-wald, Eur. Phys. J. E 2001, 5, 13; c) C. Y. Gao, H. Mçhwald, J. C. Shen,ChemPhysChem 2004, 5, 116.

[5] a) S. L. Clark, E. S. Handy, M. F. Rubner, P. T. Hammond, Adv. Mater. 1999,11, 1031; b) X. P. Jiang, P. T. Hammond, Langmuir 2000, 16, 8501.

[6] C. Y. Gao, B. Wang, J. Feng, J. C. Shen, Macromolecules 2004, 37, 8836.[7] M. Lçsche, J. Schmitt, G. Decher, W. G. Bouwman, K. Kjaer, Macromole-

cules 1998, 31, 8893.[8] T. Farhat, G. Yassin, S. T. Dubas, J. B. Schlenoff, Langmuir 1999, 15, 6621.[9] B. Wang, C. Y. Gao, L. L. Liu, J. Phys. Chem. B 2005, 109, 4887.

[10] N. Karibyants, H. Dautzenberg, Langmuir 1998, 14, 4427.[11] R. V. Klitzing, H. Mçhwald, Macromolecules 1996, 29, 6901.[12] a) J. S. Vrentas, C. M. Vrentas, Macromolecules 1996, 29, 4391; b) S. Alsoy,

J. L. Duda, AIChE J. 1999, 45, 896; c) H. H. Romdhane, P. E. Price, Jr. , C. A.Miller, P. T. Benson, S. Wang, Ind. Eng. Chem. Res. 2001, 40, 3065.

[13] S. T. Dubas, J. B. Schlenoff, Macromolecules 2001, 34, 3736.[14] Y. Xia, G. M. Whitesides, J. Am. Chem. Soc. 1995, 117, 3274.[15] R. A. McAloney, M. Sinyor, V. Dudnik, M. C. Goh, Langmuir 2001, 17,

6655.[16] G. Haugstad, W. L. Gladfelter, E. B. Weberg, R. F. Weberg, R. R. Jones,

Langmuir 1995, 11, 3473.[17] C. Ruppe, A. DuparrQ, Thin Solid Films 1996, 288, 8.[18] J. D. Mendelsohn, C. J. Barrett, V. V. Chan, A. J. Pal, A. M. Mayes, M. F.

Rubner, Langmuir 2000, 16, 5017.[19] A. S. Michaels, Ind. Eng. Chem. 1965, 57, 32.[20] A. F. Xie, S. Granick, Macromolecules 2002, 35, 1805.[21] J. Choi, M. F. Rubner, Macromolecules 2005, 38, 116.[22] a) S. Y. Park, C. J. Barrett, M. F. Rubner, A. M. Mayes, Macromolecules

2001, 34, 3384; b) S. Y. Park, M. F. Rubner, A. M. Mayes, Langmuir 2002,18, 9600.

[23] V. V. Khutoryanskiy, A. V. Dubolazov, Z. S. Nurkeeva, G. A. Mun, Langmuir2004, 20, 3785.

[24] Z. J. Sui, J. B. Schlenoff, Langmuir 2004, 20, 6026.[25] J. KrTz, H. Dautzenberg, J. Phys. Chem. A 2001, 105, 3846.[26] V. Izumrudov, E. Kharlampieva, S. A. Sukhishvili, Macromolecules 2004,

37, 8400.[27] Z. J. Sui, D. Salloum, J. B. Schlenoff, Langmuir 2003, 19, 2491.[28] M. Stark, C. Moller, D. J. Muller, R. Guckenberger, Biophys. J. 2001, 80,

3009.[29] a) R. Garcia, R. Perez, Surf. Sci. Rep. 2002, 47, 197; b) S. S. Tan, R. L. Sher-

man, Jr. , D. Q. Qin, W. T. Ford, Langmuir 2005, 21, 43.[30] M. Radmacher, R. W. Tillmann, H. Gaub, Biophys. J. 1993, 64, 735.[31] J. Tamayo, R. GarcTa, Langmuir 1996, 12, 4430.[32] B. Wang, J. Feng, C. Y. Gao, Colloid Surf. A-Physicochem. Eng. Asp. 2005,

259, 1.[33] a) J. Dai, A. W. Jensen, D. K. Mohanty, J. Erndt, M. L. Bruening, Langmuir

2001, 17, 931; b) S. E. Burke, C. J. Barrett, Macromolecules 2004, 37,5375.

[34] T. Farhat, J. B. Schlenoff, Langmuir 2001, 17, 1184.[35] J. B. Schlenoff, H. Ly, M. Li, J. Am. Chem. Soc. 1998, 120, 7626.[36] a) A. I. Petrov, A. A. Antipov, G. B. Sukhorukov, Macromolecules 2003, 36,

10079; b) S. E. Burke, C. J. Barrett, Langmuir 2003, 19, 3297; c) E. Khar-lampieva, S. A. Sukhishvili, Langmuir 2003, 19, 1235.

[37] J. A. Hiller, M. F. Rubner, Macromolecules 2003, 36, 4078.[38] a) J. KrTz, J. Dybal, D. KurkovV, J. Phys. Chem. B 2002, 106, 2175; b) J. W.

Ostrander, A. A. Mamedov, N. A. Kotov, J. Am. Chem. Soc. 2001, 123,1101.

[39] R. Overney, E. Meyer, MRS Bull. 1993, 18, 26.[40] T. Kajiyama, K. Tanaka, A. Takahara, Macromolecules 1997, 30, 280.[41] P. V. Pavoor, A. Bellare, A. Strom, D. H. Yang, R. E. Cohen, Macromolecules

2004, 37, 4865.[42] D. Kovacevic, S. van der Burgh, A. de Keizer, M. A. Cohen-Stuart, Lang-

muir 2002, 18, 5607.

Received: July 10, 2005

Revised: October 25, 2005

Published online on January 20, 2006

596 www.chemphyschem.org B 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ChemPhysChem 2006, 7, 590 – 596

C. Gao et al.