Water Interface versus Reaction in a Flask: Tuning

Methyl−methyl interactions observed in CoStn samples suggest hierarchy of supramolecular chemistry at the air/water interface in tuning the C−O−...
4 downloads 0 Views 1MB Size
Download PDF
Langmuir 2009, 25, 3519-3528

3519

Chemistry at Air/Water Interface versus Reaction in a Flask: Tuning Molecular Conformation in Thin Films Smita Mukherjee,† Alokmay Datta,*,† Angelo Giglia,‡ Nicola Mahne,‡ and Stefano Nannarone‡,§ Surface Physics DiVision, Saha Institute of Nuclear Physics, 1/AF Bidhannagar, Kolkata 700 064, India, TASC-INFM National Laboratory, Area Science Park, BasoVizza (Trieste), Italy, INFM and Dipartimento di Ingegneria dei Materiali ed Amb., UniVersita` di Modena e Reggio Emilia, Modena, Italy ReceiVed July 22, 2008. ReVised Manuscript ReceiVed December 13, 2008 Atomic force microscopy and X-ray reflectivity studies of cobalt stearate Langmuir-Blodgett (LB) films (CoStp) deposited from a preformed bulk sample on quartz substrates showed formation of a Volmer-Weber type monolayer but no multilayers as compared to the excellent multilayers of cobalt stearate films (CoStn) deposited at the air/water interface by the usual LB technique, in spite of both showing bidentate bridging type coordination of cobalt ions with the carboxylate group. The difference is attributed to existence of different headgroup conformers, observed from Fourier transform infrared (FTIR) studies. The CoStp films had a higher energy ‘boat’ conformation with linear O-Co-O linkage, whereas CoStn formed a low energy conformer with a bent O-Co-O configuration (bond angle of 105°). Present results support the necessity of bidentate bridging coordination in multilayer deposition, but rejects its sufficiency by bringing out the crucial role played by air/water interface. Differences in surface pressure-molecular area isotherms and hydrocarbon tail-tail interactions (evident from FTIR spectra) of the films support the above statement. Methyl-methyl interactions observed in CoStn samples suggest hierarchy of supramolecular chemistry at the air/water interface in tuning the C-O-Co bond angle essential to satisfy the wetting condition with the substrate and subsequently form LB multilayers.

Introduction The unique role that air/water interface can play in preparation of substances is exemplified quite emphatically by deposition of Langmuir-Blodgett (LB) films1,2 of long chain fatty acid salts of divalent metals, especially those having considerable electronegativity, such as Co, Mn, Ni, Cd, Zn, Cu and Pb.3-8 Bulk preparation of these salts is a two-step process, since they do not interact directly with these acids, involving preparation of salt of the fatty acid with an alkali metal and then an exchange reaction whereby the divalent metal forms the salt.9,10 Purification of the salt is again a nontrivial task. Thin film deposition of this salt can then be carried out by vertically dipping a suitable substrate through a salt monolayer formed at the air/water interface of a Langmuir trough from a solution of the bulk prepared salt.11 On the other hand, these fatty acids, when floating in a Langmuir * Corresponding author. E-mail: [email protected]. † Surface Physics Division, Saha Institute of Nuclear Physics. ‡ TASC-INFM National Laboratory. § INFM and Dipartimento di Ingegneria dei Materiali ed Amb., Universita` di Modena e Reggio Emilia. (1) Talham, D. R.; Yamamoto, T.; Meisel, M. W. J. Phys.: Condens Matter 2008, 20, 184006. (2) Petty, M. C. Langmuir-Blodgett Films: An Introduction, Cambridge University Press: Cambridge, U.K., 1996. (3) Schwartz, D. K.; Viswanathan, R.; Garnaes, J.; Zasadzinski, J. A. J. Am. Chem. Soc. 1993, 115, 7374. (4) Luo, X.; Zhang, Z.; Liang, Y. Langmuir 1994, 10, 3213. (5) Geue, Th.; Schultz, M.; Englisch, U.; Stommer, R.; Pietsch, U.; Meine, K.; Vollhardt, D. J. Chem. Phys. 1999, 110, 104. (6) Ando, Y.; Hiroike, T.; Miyashita, T.; Miyazaki, T. Thin Solid Films 1996, 278, 144. (7) Kumar, N. P.; Major, S. S.; Vitta, S.; Talwar, S. S.; Gupta, A.; Dasannacharya, B. A. Colloid Surf. A 2005, 257, 243. (8) Erokhina, S.; Erokhin, V.; Nicolini, C. Colloid Surf. A 2002, 198 - 200, 645. (9) Kundu, S.; Datta, A.; Hazra, S. Langmuir 2008, 24, 9386. (10) Datta, A.; Kundu, S.; Sanyal, M. K.; Diallant, J.; Luzet, D.; Blot, C.; Struth, B. Phys. ReV. E 2005, 71, 041604. (11) Datta, A.; Sanyal, M. K.; Dhanabalan, A.; Major, S. S. J. Phys. Chem. B 1997, 101, 9280.

monolayer on water surface, can react directly with ions of required divalent metal, dissolved in water in less than micromolar concentrations, to yield a nearly purified form of the salt that can be deposited as multilayers on a suitable substrate by the same vertical dipping procedure.12 Two questions that come in mind are as follows: (1) Are the two films, prepared by these processes, identical? (2) What specific role do supramolecular forces play at the air/water interface during film deposition? These questions relate to our findings about effects of different divalent metals on building supramolecular structure and hence film morphology. The occurrence of bidentate-bridged metalcarboxylate coordination in bilayer headgroups was found to drastically reduce “pinhole” defects in LB multilayers,13 and the importance of supramolecular interactions in LB multilayer formation14 was also suggested. Cobalt stearate (CoSt), which formed almost defect-free LB multilayers in these studies, is hence chosen as the model system in the present attempt to answer these questions. Incorporation of transition metals, especially cobalt in soft matter,15,16 is receiving increasing interest due to potential uses toward responsive materials. Transition metal-bearing amphiphiles have also been used to transfer and monitor the transfer of biologically interesting macromolecules to Langmuir-Blodgett multilayers.17 However, other than these application aspects, our studies relate to the specific interconnections among coordination and bonding on one hand, and supramolecular interactions and morphology on the other, since (12) Kundu, S.; Datta, A.; Hazra, S. Chem. Phys. Lett. 2005, 405, 282. (13) Mukherjee, S.; Datta, A.; Giglia, A.; Mahne, N.; Nannarone, S. Chem. Phys. Lett. 2008, 80, 451. (14) Mukherjee, S.; Datta, A. J. Nanosci. Nanotechnol. (in press). (15) Wu, N.; Fu, L.; Su, M.; Aslam, M.; Wong, K. C.; Dravid, V. P. Nano Lett. 2004, 4, 383. (16) Shakya, R.; Hindo, S. S.; Wu, L.; Allard, M. M.; Heeg, M. J.; Hratchian, H. P.; McGarvey, B. R.; da Rocha, S. R. P.; Verani, C. N. Inorg. Chem. 2007, 46, 9808. (17) Rajdev, P.; Chatterji, D. Langmuir 2007, 23, 2037. and citations therein.

10.1021/la8023502 CCC: $40.75  2009 American Chemical Society Published on Web 02/18/2009

3520 Langmuir, Vol. 25, No. 6, 2009

such interconnections among structural hierarchy over many length scales provide the key to understand complex dynamics.18 In the present work, we have tried to answer the first question through identifying the specific molecular conformers created by the processes of bulk reaction and reaction at air-water interface. To answer the second question, we have indicated the dependence of the specific conformer formed at water surface on supramolecular interactions. We have also shown the correlation between the specific conformer and morphology of multilayers comprising molecules in that conformer. We prepared thin films of CoSt in the two distinct ways mentioned before. In the first, cobalt stearate (CoStb) prepared in bulk, was spread on water as the Langmuir monolayer. In this case head-tail ratio of amphiphiles in the monolayer is predetermined. However, for such preformed salts, exact molecular configuration at the air/ water interface is not generally known. These thin films, deposited from CoStb are named CoStp, throughout the paper. In the second way, the usual LB technique of vertical deposition19 with stearic acid as the monolayer on subphase water containing divalent cobalt ions, was employed. Henceforth for clarity, these CoSt films will be called CoStn. A comparative study of films deposited by both processes provided clues to arrive at the answers, as indicated above. X-ray reflectivity (XRR) and Atomic force microscopy (AFM) have been used to study the observed morphological changes of the LB films of cobalt stearate deposited in these two ways. Information about the metal ion-headgroup conformations have been obtained from Fourier transform infrared (FTIR) spectroscopy. The structure of the paper is as follows. After this introductory section, the methods of preparation of CoSt films are described. Details of their probing alongside the relevant methods of analysis are also discussed. It is then shown that both films have same metal-headgroup coordination but widely different film structure. From a study of surface energy relations, it is indicated that they belong to two different growth modes, thereby proving the insufficiency of metal-organic coordination in deciding supramolecular association. The different conformations of the molecules of CoSt prepared from these two ways are established and it is shown that a particular conformer exhibits supramolecular bonding. The results are followed by a conclusion and an outlook.

Experimental Details I. Preparation of Cobalt Stearate in Bulk (CoStp Film Deposition). CoStb was prepared through stepwise chemical reaction. In the first step, preparation of sodium stearate was carried out by addition of sodium hydroxide (Merck, 99%) in Milli-Q water (resistivity 18.2 MΩ-cm) containing stearic acid (C17H35COOH, Sigma-Aldrich, 99%) in appropriate amounts at about 80-90 °C. Addition of sodium hydroxide was continued until the medium was slightly alkaline (pH 7.0-7.5). Sodium stearate was completely soluble in hot water. In the second step, measured amount of cobalt chloride (Merck, 99%) solution was added to the freshly prepared sodium stearate solution in hot condition, whence cobalt stearate was formed. Cobalt stearate is insoluble in water and forms a blue precipitate, which was filtered out. In the third step, cobalt stearate was repeatedly washed with hot Milli-Q water to filter out any unreacted sodium stearate and other water soluble impurities, followed by a repeated wash with benzene (SLR, 99.8%) to remove any unreacted stearic acid and other organic impurities. Formation of the stearate salt from stearic acid was confirmed by FTIR spectroscopic (Spectrum GX, Perkin-Elmer) measurements20 on bulk (18) Whitty, A. Nature Chem. Biol. 2008, 4, 435. (19) Sanyal, M. K.; Mukhopadhyay, M. K.; Mukherjee, M.; Datta, A.; Basu, J. K.; Penfold, J. Phys. ReV. B 2002, 65, 033409. (20) Smith, B. C. Fundamentals of Fourier Transform Infrared Spectroscopy, CRC Press: New York, 1996.

Mukherjee et al. stearic acid (StA) and CoStb in transmission mode at a resolution of 4.0 cm-1. For this, both StA and CoStb solutions in chloroform (CH3Cl, Sigma-Aldrich, 99.9+%) were drop cast on KBr substrates. These were the bulk samples. Results of these measurements are discussed in detail later. A 0.5 mg/mL chloroform solution of CoStb was spread in the Langmuir trough (KSV instruments, KSV-5000) containing Milli-Q water and its surface pressure (π) vs specific molecular area (A) isotherm2 was recorded by compressing the monolayer at a barrier speed of 3 mm/min. Two CoStp films viz. CoStp-3 and CoStp-9, were deposited on fused quartz substrates by three and nine subsequent passages of the substrate through the monolayer respectively starting from water to air, by the Langmuir-Blodgett technique.21 Fused quartz substrates were made hydrophilic by keeping them in a solution containing ammonium hydroxide (Merck, 98%), hydrogen peroxide (Merck, 98%) and Milli-Q water (H2O: NH4OH: H2O2 ) 2:1:1 by volume) for 10-15 min at 100 °C. II. Preparation of Cobalt Stearate at Water Surface (CoStn Film Deposition). Divalent cobalt ions were introduced in the Langmuir trough containing Milli-Q water through addition of 0.5 mM cobalt chloride solution. pH was maintained at 6.0 by addition of sodium bicarbonate (Merck, 99%). Stearic acid monolayers were spread from a chloroform solution of concentration 0.5 mg/mL. The π - A isotherm was recorded by compressing the monolayer at a barrier speed of 3 mm/min. Two films, viz. CoStn-3 and CoStn-9, were deposited on fused quartz substrates by three and nine subsequent dips respectively, of the substrate through air/water interface starting from water, by the same Langmuir-Blodgett technique. The films were deposited at a monolayer pressure of 30 mN/m (measured by Platinum Wilhelmy plate) for CoStn films, and at 18 mN/m for CoStp films, at 22 °C at a dipping speed of 3 mm/min. Drying time after first stroke was 10 min. and the same was 5 min. from next cycle. All films were deposited more than once and were checked for reproducibility by noting their consistent XRR profiles after each deposition. XRR22 of the deposited LB films was carried out using the versatile X-ray diffractometer (VXRD; Bruker AXS) with X-ray wavelength λ ) 1.54 Å (Cu KR line). Surface morphology of the deposited LB films was studied by AFM (Autoprobe CP, Park Scientific) in contact mode. Scans of 10 × 10 µm2 and 5 × 5 µm2 were performed in constant force mode over several regions of the film. A low force constant (1.3 mN) was used to minimize damage. FTIR measurements of all LB deposited films were carried out in transmittance, attenuated total reflection (ATR), or reflection modes depending on the frequency region of interest. A detailed discussion of these measurements will be provided at the relevant places.

Results and Discussion I. Cobalt-Carboxylate Coordination. FTIR spectra of CoStb and StA are shown in Figure 1. Assigned peaks are given in Table 1. Presence of strong bands corresponding to carboxylate asymmetric (1542 cm-1) and symmetric (1398 cm-1) stretch modes23,24 in the CoStb spectra (Figure 1a) indicate large conversion of fatty acid to its metal salt. Presence of very weak bands corresponding to CdO stretch (around 1700 cm-1) and C-OH deformation (around 1300 cm-1), and absence of OH stretch band (around 3400 cm-1) in the sample indicate negligible amount of free acid and hydroxyl group,26 respectively, compared to strong acid peaks in StA spectra (Figure 1b). In order to find coordination of metal ion with carboxylate group,27 difference in symmetric (νs) and asymmetric (νa) COO stretching frequencies (21) Schwartz, D. K. Surf. Sci. Rep. 1997, 27, 241. (22) Daillant, J.; Gibaud A. X-Ray and Neutron ReflectiVity: Principles and applications; Springer: New York, 1999. (23) Kimura, F.; Umemura, J.; Takenaka, T. Langmuir 1986, 2, 96. (24) Kemp, W. Organic Spectroscopy, Palgrave: Hampshire, U.K., 2002. (25) Corkery, R. W.; Hockless, D. C. R. Acta Crystallogr. 1997, C53, 840. (26) Rabolt, J. F.; Burns, F. C.; Schlotter, N. E.; Swalen, J. D. J. Chem. Phys. 1983, 78, 946. (27) Nakamoto, K. Infrared and Raman Spectra of Inorganic and Coordination Compounds; Wiley and Sons: New York, 1970.

Chemistry at Air/Water Interface

Langmuir, Vol. 25, No. 6, 2009 3521

Figure 1. FTIR spectra of (a) bulk cobalt stearate (CoStb) and (b) bulk stearic acid (StA). Methyl stretching regions (2800-3600 cm-1) are shown in respective insets. Table 1. Infrared Spectra of Bulk Cobalt Stearate and Stearic Acid (ν, Stretch; a, Asymmetric; s, Symmetric)a vibrational frequencies (cm-1) CoStb 1699 (M) 1684 (M) 1542 (S) 1398 (VS) a

StA

assigned to

3400 (S) 1703 (VS) 1688 (VS)

ν(O-H) ν(CdO); split due to dimer formation νa(COO) νs(COO)

VS: very strong. S: strong. M: medium.

(∆ν) were measured. For CoStb, ∆ν ) 144 cm-1, which corresponds to bidentate bridging coordination,15 having one metal ion per carboxylate group. It must be mentioned here that CoStn LB multilayers were reported to form same bidentate bridging coordination of cobalt ion with COO group.13 Thus we see that identical metal ion carboxylate coordination exists in cobalt stearate samples prepared by the two routes. In general for bulk prepared samples, multiple metal ion coordinations can occur with sufficiently high subphase pH28 and coordinations vary with chemical environment.29 As such coordinations give well separated characteristic carboxylate stretch bands.29 Absence of other strong bands from 1550 cm-1 - 1600 cm-1 and from 1350-1450 cm-1 in our bulk CoSt sample, however, clearly shows that samples are quite free of such multiplicities and an almost pure bidentate bridging coordination is obtained. The fact that same coordination is obtained in the sample prepared at air/water interface points to two aspects. On one hand it shows that, bonding being same, the two systems CoStp and CoStn, can behave differently only by having different conformations, thereby bringing in the role of supramolecular forces in deciding their behavior. On the other hand it shows the selectivity of air/water interface toward a particular coordination of Co ions with carboxylate groups when the latter form a twodimensional lattice at the interface, and Co ions associate with this lattice and distort it.30 II. Morphology and Surface Energy. AFM topographic images (2 µm × 2 µm) depicting surface morphology of CoStp-3 and CoStn-3 LB films are shown in Figure 2. Presence of islandlike formations are evident in CoStp-3 (Figure 2a) whereas CoStn-3 (Figure 2b) shows no such features. On the contrary, the latter shows a mesoscopically smooth morphology with (28) Larabee, J. A.; Leung, C. H.; Moore, R. L.; Thamrong-nawasawat, T.; Wessler, B. S. H. J. Am. Chem. Soc. 2004, 126, 12316. (29) Kremer-Aach, A.; Kla1ui, W.; Bell, R.; Strerath, A.; Wunderlich, H.; Mootz, D. Inorg. Chem. 1997, 36, 1552. (30) Kmetko, J.; Datta, A.; Evmenenko, G.; Dutta, P. J. Phys. Chem. B 2001, 105, 10818.

negligible “pinhole” defects, consistent with our previous work.13 Line profiles drawn through these images (Figure 2, parts a and b (insets)) show heights of these islands (pinholes for CoStn-3) to be ∼26 Å (30 Å for CoStn-3) which corresponds to a typical monolayer height. Estimated rms roughness is low (∼5.8 Å for CoStp-3 and 1.9 Å for CoStn-3). Hence, AFM results show that in spite of bidentate bridging coordination, the CoStp-3 film did not show smooth, LB-like morphology. This clearly indicates that even with same salt having identical headgroup coordination deposited in the same way, films do not show identical morphology, illustrating importance of air/water interface in depositing LB multilayers. Indication of such a possibility is evident from π-A isotherms (Figure 3) of CoStb monolayer on water (I) and StA monolayer with cobalt ions in water (II). The former showed a lower collapse pressure compared to the latter. However, both show a region corresponding to the stable condensed phase, although at different pressure regions (18 mN/m for CoStb and 30 mN/m for StA+ Co2+ system) and the films were deposited at these phases. Here, it must be mentioned that due to the low area per molecule and surface pressure values observed from the CoStp isotherm, existence of a true monolayer at the air/water interface may be questioned. To investigate that, two more CoStp films were deposited under the same conditions. One was lifted horizontally (CoStp-H) by the modified inverted Langmuir-Schaefer (MILS) technique31 at a speed of 0.5 mm/min, and the other by lifting the substrate vertically (CoStp-V) through the air/water interface. The AFM topographic images of CoStp-H (Figure 4a) and CoStp-V (Figure 4b) show fairly uniform coverage along with some “pinhole” defects. Line profiles (insets) of both show the heights of the “pinholes” to be 25 Å, corresponding to monolayer heights. CoStp-H showed clearly that even at a lower area/molecule and surface pressure value, preformed cobalt stearate was present as a stable monolayer at the air/water interface and did not form multilayered structures. Furthermore, deposition of CoStp-V showed that this preformed system can be successfully transferred to a solid substrate as a true monolayer. In terms of surface free energy per unit area σ, we know from Bauer’s criteria,32 the value of ∆σ ) σo + σi - σs where suffixes o, i and s denote overlayer, interface and substrate respectively. ∆σ e 0 denotes Frank van der Merwe (F-M) or “layer-by-layer” growth whereas ∆σ > 0 gives Volmer-Weber (V-W) or “islandlike” growth.33 In this case, s is SiO2 with σs ) 30.95 mJ/m2, obtained by dividing its Hamaker constant value (reported as (31) Kundu, S.; Datta, A.; Hazra, S. Langmuir 2005, 21, 5894. (32) Van der Merwe, J. H. Interface Sci. 1993, 1, 1. (33) Daruka, I.; Barabasi, A. Phys. ReV. Lett. 1997, 79, 3708.

3522 Langmuir, Vol. 25, No. 6, 2009

Figure 2. AFM topographic images (2 µm × 2 µm) of (a) CoStp-3 and (b) CoStn-3 LB films with corresponding line profiles (insets).

Figure 3. Surface pressure π vs molecular area A isotherms of (I) CoStb monolayer on water and (II) StA monolayer with cobalt ions in water.

(6.4-6.6) × 10-20 J)34 by a factor 2.1 × 10-21 m-2.35 The overlayer o is the hydrocarbon tail of the monolayer, with σo ) 47.4 mJ/m2 for CoStb (17.7 mJ/m2 for StA+ Co2+ system),

Mukherjee et al.

calculated from the basic relation π ) σw - σo, where π ) 25.4 mJ/m2 (55.1 mJ/m2) is measured surface pressure of CoStb (StA + Co2+) Langmuir monolayer just before collapse and σw ) 72.8 mJ/m2 is that of pure water at 20 °C. The reason for choosing collapse value of surface pressure for cobalt stearate is that it corresponds to the most closed packed condensed phase of monolayer on water. Therefore, for CoStb (StA + Co2+), ∆σ ) 16.5 + σi (∆σ ) -13.3 + σi), last term being interaction energy of cobalt ions with substrate, a small and often positive quantity. Above analysis indicates that as long as interaction energies do not differ drastically, ∆σ values predict F-M growth for CoStn films and V-W growth for CoStp films. Thus, LB films of CoStp deposited at π ) 18 mN/m will essentially show island-like formations resembling V-W growth mode. It must be mentioned here that a study of morphological evolution of CoStn films indeed showed F-M type growth from very first monolayer.14 Thus the obtained morphology of CoStp is a result of its behavior at air/water interface. However it must be bourne in mind that surface energy of overlayer was obtained from the Langmuir monolayer, which may differ from that on substrate, due to development of some amount of strain during transfer of monolayers onto substrate. At this point, it appears that CoStp contains growth defects, which stems from the fact that its surface energy on water at the deposited pressure characterizes V-W growth for this system. It must also be noted that AFM profile analysis indicated monolayer heights for both CoStp-3 and CoStn-3 films. For CoStn-3, with its smooth surface, this is an indication of very good film coverage along film thickness with pinholes strictly confined within topmost layer. For CoStp-3, this value should have been ∼75 Å (height of a trilayer) considering presence of huge number of defects in the in-plane image. Although height values obtained from AFM data include tip convolution effect and may not give true height variations, yet the above result seems puzzling. So to obtain precise out-of-plane information of deposited films XRR was carried out. XRR data of CoStp-3 and CoStp-9 films are shown (open circles) in Figure 5a parts I and II, respectively. Both films indicate only a monolayer feature along with some ‘humps’. The data has been analyzed using Parratt formalism,36 where each film is divided into layers of fixed thickness (d), average electron density (F) and interfacial roughness (σ), and used as fit parameters.37 Solid lines are best-fit curves. Electron density profiles (EDPs) for films (electron density as a function of film thickness z, where z ) 0 is air), obtained from fits, are shown in respective insets. Values of the fit parameters of all films are given in Table 2. CoStp-3 was fitted by a monolayer of thickness 20.8 Å and a 39.8 Å thick bilayer of negligible electron density (0.02 electrons per Å3). Monolayer roughness estimated was 7 Å (consistent with AFM). CoStp-9 data was fitted with three arbitrary layers of thickness 80, 48.3, and 44.5 Å having F ) 0.68, 0.11 and 0.06 electrons per Å3 respectively, i.e., islands varying from three to seven monolayers, suggesting an increase in coverage but no LB multilayer features were observed, consistent with AFM results. XRR data of CoStn-3 and CoStn-9 films are shown (open circles) in Figure 5b, parts I and II, respectively. Both films show clear indication of multilayer formation. For CoStn-9, 5 Bragg peaks, with prominent Kiessig (interference) fringes are obtained. Prominent Kiessig fringes are an indication of coherence between air/film and film/substrate interfaces.22 From the number of Bragg peaks it is evident that LB multilayers have good outof-plane crystalline growth. Corresponding EDPs of CoStn (insets) show no considerable decrease in film coverage along z. This indicates presence of very few pinholes in films. Total

Chemistry at Air/Water Interface

Langmuir, Vol. 25, No. 6, 2009 3523

Figure 4. AFM topographic images of (a) CoStp-H (1.5 µm × 1.5 µm) and (b) CoStp-V (4 µm × 4 µm) LB films with corresponding line profiles (insets).

Figure 5. X-ray reflectivity data (open circles) with corresponding fits (solid lines) of (a) CoStp samples, (I) CoStp-3 and (II) CoStp-9, and (b) CoStn samples, (I) CoStn-3 and (II) CoStn-9, with their corresponding electron density profiles (EDPs) shown in respective insets.

film thickness is 72 Å for CoStn-3 and 221 Å for CoStn-9. Air/ film σ-values are 6.1 Å (CoStn-3) and 6.0 Å (CoStn-9). Thus, for CoStn we have a compact multilayer growth on fused quartz substrates. XRR analysis clearly shows absence of LB multilayers in case of CoStp. Hence, island-like formations observed in AFM topographic image of CoStp are not “pinhole” defects arising

during multilayer deposition. They merely represent a V-W type monolayer. Thus, we can definitely say that metal ion coordination, although a necessary condition for defect-free morphology, as mentioned in ref 13, is not sufficient to yield proper LB multilayers. In other words, bulk growth and interfacial growth produce cobalt stearates, which though are same as far as

3524 Langmuir, Vol. 25, No. 6, 2009

Mukherjee et al.

Table 2. Fitting Parameters of X-ray Reflectivity Data for CoStp and CoStn Samples (h, Head; t, Tail; tt, Tail-Tail Interface) sample

layer

d (Å)

F (eÅ-3)

σ (Å)

sample

layer

d (Å)

F (eÅ-3)

σ (Å)

CoStp-3

t3 h2 t2 tt t1 h1 SiO2 layer3 layer2 layer1 SiO2 t3 h2 t2 tt t1 h1 SiO2

17.0 3.8 17.0 2.0 17.0 3.8 1 × 107 44.5 48.3 80.0 1 × 107 18.4 6.9 22.0 3.5 18.2 3.2 1 × 107

0.02 0.03 0.02 0.00 0.30 0.36 0.67 0.06 0.11 0.68 0.67 0.14 0.18 0.14 0.03 0.33 0.37 0.67

4.0 0.5 0.5 0.5 7.0 0.0 6.8 1.0 8.6 15.5 8.5 6.1 1.7 1.1 2.7 4.2 1.8 5.7

CoStn-9

t9 h8 t8 tt t7 h6 t6 tt t5 h4 t4 tt t3 h2 t2 tt t1 h1 SiO2

20.9 7.8 19.2 1.3 21.0 6.8 21.0 1.6 21.0 5.0 20.6 1.9 21.0 5.8 20.8 1.3 18.8 5.0 1 × 107

0.20 0.33 0.21 0.14 0.20 0.34 0.23 0.15 0.21 0.42 0.27 0.18 0.26 0.44 0.33 0.23 0.35 0.41 0.67

6.0 4.7 4.4 1.5 2.1 4.7 4.1 0.0 1.0 4.5 4.1 1.5 2.3 5.0 4.5 1.3 3.0 4.0 8.0

CoStp-9

CoStn-3

coordination is concerned, yet behave differently in supramolecular organization. As suggested before, the next step is to check their conformation, in particular, conformations of respective headgroups, where the metal is bonded to the organic amphiphile. III. Molecular Conformation and Supramolecular Interactions. In order to investigate that, a comparative FTIR study of both CoStp-9 and CoStn-9 was carried out which focused on large amplitude vibrational modes associated with Co-bearing headgroups. The results are discussed in the next subsection. FTIR spectroscopy was carried out in attenuated total reflectance (ATR) mode at 4.00 cm-1 resolution (Figure 6, parts a and b, for CoStp-9 and CoStn-9, respectively). ATR spectra was taken

Figure 6. FTIR spectra of (a) CoStp-9, (b) CoStn-9, and (c) stearic acid. FTIR spectra of CoStb (inset) showing presence of peak at 775 cm-1.

using a zinc selenide plate, which has a transmittance range above 650 cm-1. So for the frequency range below 650 cm-1, separate spectra was taken in reflection mode, using silicon as substrate (since fused quartz gave low signal in reflection mode). Assigned peaks of the spectra (after ATR and Krammers Kronig corrections20) are given in Table 3. A. The Headgroup Conformation. As discussed earlier, both films showed bidentate bridged Co-COO coordination. This requires two cobalt ions to be incorporated between two carboxylate groups, such that each cobalt ion is linked with two oxygen atoms (of different COO groups) on either side of it (Figure 7a). The Co-O stretch band lies in the range 665-703 cm-1 in cobalt oxide samples.38 Since the COO stretching vibrations lie in the region 1400-1500 cm-1, i.e, far off from the O-Co-O stretch frequency range, O-Co-O and COO groups can be considered to be effectively decoupled and O-Co-O can be treated as isolated symmetric triatomic system giving rise to two stretching frequency modes, asymmetric (νa) and symmetric (νs), on either side of the diatomic X-Y stretching frequency, the asymmetric stretch being at a slightly higher value.39 For a theoretical diatomic system Co-O, the stretching frequency comes out to be 640 cm-1. Thus in both CoStp-9 and CoStn-9 samples, νa and νs are expected to lie above and below 640 cm-1 respectively. Comparing the two spectra and also with that of StA in this region (Figure 6 (c)), we find that CoStp-9 (CoStn-9) has two distinct peaks at 773 cm-1 (650 cm-1) and 616 cm-1 (616 cm-1), which we tentatively assign to νa and νs respectively. It is worthwhile mentioning that presence of the peak at 775 cm-1 in CoStb (Figure 6c (inset)) indicates that headgroup conformers of CoStp-9 and CoStb are similar, and hence presumably remain unchanged at air/water interface. In other words, bulk CoSt sample formed during chemical reaction exists in a particular headgroup conformer state and it remains in that conformer state throughout. For a symmetric triatomic molecule X-Y-X, the asymmetric (νa) and symmetric (νs) stretching frequencies are given by the relationship νa, s ) (2πc)-1k1/2[m-1 + M-1(1 - cos 2R)]1/2, where (34) Bollinne, C.; Cuenot, S.; Nysten, B.; Jonas, A. M. Eur. Phys. J. E 2003, 12, 389. (35) Israelachvili, J. N. Intermolecular and Surface Forces, Academic Press: London, 1991. (36) Parratt, L. G. Phys. ReV. 1954, 95, 359. (37) Basu, J. K.; Sanyal, M. K. Phys. Rep. 2002, 363, 1. (38) Ahmed, S. R.; Kofinas, P. J. Magn. Magn. Mater. 2005, 288, 219. (39) Colthup, N. B.; Daly, L. H.; Wiberly S. E. Introduction to Infrared and Raman Spectroscopy; Academic Press: New York, 1965.

Chemistry at Air/Water Interface

Langmuir, Vol. 25, No. 6, 2009 3525 Table 3. Infrared Spectra of CoStn and CoStp Samplesa

vibrational frequencies (cm-1) CoStn

CoStp 1220 (M)

1213 (C) 1165 (VS)

1163 (W)

1112 (VS) 1068 (VS) 982 (M) 881 (VS)

1117 (W) 1100 (VW) 1053 (M) 883 (W) 773 (VS)

StA 1219 (W) 1201 (W) 1186 (W) 1121 (VW)

650 (VS) 616 (M)

1102 (W) 1074 (VW) 880 (VW)

616 (S)

assigned to δ(CH2), wag δ(CH2), wag δ(CH2), wag δ(CH2), wag, and/or ν(C-O-Co), stretch ν(C-C), skeletal ν(C-C), skeletal, and/or ν(C-O-Co), stretch ν(C-C), skeletal ν(C-C), skeletal skeletal plane and/or δ(C-O-Co) deformation νa(O-Co-O) νa(O-Co-O) νs(O-Co-O)

a

VS: very strong. S: strong. M: medium. W: weak. VW: very weak. C: convoluted. ν: stretch. δ: deformation/bend (includes scissor, rock, wag, twist). a: asymmetric. s: symmetric.

Figure 7. Cartoon showing (a) the bidentate bridged headgroup along with (b) the headgroup conformations of (I) CoStp and (II) CoStn samples.

m and M are masses of X and Y atoms respectively, 2R is bond angle, c is the velocity of light and k is the force constant of the X-Y bond.40 The value of k was calculated from the empirical relation for an approximate force constant of a single bond between atoms X and Y as kXY ) 7.20 × 102(ZXZY)(nXnY)-3 N/m, where Z and n are atomic number and principal quantum number of valence electrons, respectively.41 The theoretical k value came out as 3.04 × 102 N/m. From the above two simultaneous equations, experimental values of k and R of CoStp-9 and CoStn-9 are determined (Table 4), using observed νa and νs values for (40) Goldstein, H. Classical Mechanics; Addison-Wesley: Upper Saddle River, NJ, 1980. (41) Schrader, B. Infrared and Raman Spectroscopy: Methods and Applications; VCH Publishers: NewYork, 1995.

O-Co-O stretch mode of the samples. No unique solution for k and R was obtained for CoStp-9 but, when k versus R curves obtained from both equations are plotted (Figure 8a), k values of both curves approached each other as R approached 90° and these values were chosen as the solutions. The value of ∆ν ) (νa - νs), which varies as 2R, was also calculated. Theoretical ∆ν values were calculated for all values of R from 0° to 90° (Figure 8c), and a plot of ∆νTh vs R (Figure 8c, inset) was obtained. The values are presented in Table 4. As seen from the table, observed and calculated k and ∆ν values of CoStn-9 are in good agreement. Most importantly, R values show that O-Co-O linkage is linear in CoStp-9 but nonlinear for CoStn-9 with bond angle of 105°. For transition metals, the 3d-2p hybridization affects bond angle,42 with the latter decreasing with an increase in former, hinting at existence of different hybridization in each case. Difference in bond angles point out that although both headgroups are bidentate bridged, they definitely exist as different conformers43 with different hybridization states, with bulk samples having higher energy, since it has higher νa value. Next, the carboxylate linkages in both samples are investigated. From carboxylate stretch bands mentioned before, similar calculations for determination of experimental values of force constant k and bond angle R (labeled as RC in Figure 7 for clarity) were carried out (Table 4). However, since carboxylate group in bidentate bridging coordination behaves similar to a carboxylate free ion,44 and since the latter exhibits a CO bonding intermediate between a pure single and double bond due to existence of the resonance hybrid structure,45 theoretical k value was computed using the expression of force constant for multiple bond as kN ) kXYN(r1/rN) where kXY is the force constant for a X-Y single bond, N is the bond order, and r1 and rN are single and multiple bond lengths.41 Taking N ) 1.5, r1 ) 1.36 Å, and rN ) 1.27 Å, kTh came out to be 8.67 × 102 N/m. Comparing experimental values for COO (which are in good agreement with theoretical ones), we find that carboxylate group is similar in both samples. From the above analysis, possibility of existence of planar metal-carboxylate headgroup geometries in both CoStp and CoStn films can be ruled out, since a planar geometry would lead to greater delocalization of π bonds of COO leading to coupling (42) Nekrasov, I. A.; Streltsov, S. V.; Korotin, M. A.; Anisimov, V. I. Phys. ReV. B 2003, 68, 235113. (43) Lister, D. G. Internal Rotation and InVersion; Academic Press: London, 1978. (44) Nakamoto, K. Infrared Spectra of Inorganic and Coordination Compounds; Wiley and Sons: New York, 1963. (45) Pauling, L. The Nature of The Chemical Bond and The Structure of Molecules and Crystals: An Introduction to Modern Structural Chemistry, 3rd ed.; Cornell University Press: Ithaca, NY, 1960.

3526 Langmuir, Vol. 25, No. 6, 2009

Mukherjee et al.

Table 4. Force Constants and Bond Angles from FTIR Analysis triatomic system

sample

νa (cm-1)

νs (cm-1)

∆ν (cm-1)

k × 102 (N/m)

R (deg)

kTh × 102 (N/m)

∆νTh (cm-1)

O-Co-O linkage

CoStp-9 CoStn-9 CoStp-9 CoStn-9

773 650 1542 1541

616 616 1398 1397

157 34 144 144

3.61 2.98 8.77 8.76

90.0 52.3 50.0 49.9

3.04 3.04 8.67 8.67

148 36 145 145

O-C-O linkage

between O-Co-O and COO systems, which is inconsistent with an isolated triatomic system. Taking into consideration the fact that CoStb forms a stable monolayer up to 26 mN/m, as seen from its π-A isotherm, we assume that the preferred conformer state is a ‘boat’ (Figure 7b (I)), with two O-Co-O groups forming its base and two carboxylate groups with their hydrocarbon tails being on same side i.e. away from water at the air/water interface with a fixed value of C-O-Co angle θ. Taking into account covalent radii of Co and O, area of the base of the boat was calculated as 15.75 Å2, which was very close to the value (12.5 Å2) of area/molecule of CoStp at 18 mN/m pressure as obtained from its isotherm. Although it is difficult to quote the exact value of θ from the above data, nevertheless it can be definitely said that θ remains frozen at the air/water interface and also in CoStp samples, as FTIR spectra of CoStb and CoStp samples show similar spectral features as mentioned earlier, suggesting a rigid headgroup structure. Similarly, from structural data obtained above, and taking into account covalent radii of Co ions, a tentative headgroup conformer is constructed for CoStn (Figure 7b (II)). As seen from the cartoon, the headgroup is a twisted conformer with O-Co-O plane inclined at some angle β (e.g.) with plane of four O atoms of two COO groups. However, it is difficult to quote the exact value of β at this stage.

Figure 8. Variation of force constant k vs R (half the bond angle) for O-Co-O bond in (a) CoStp and (b) CoStn samples. Curves I and II correspond to asymmetric and symmetric O-Co-O stretching frequencies. (c) Theoretical variation of (I) νa, (II) νs, and (inset) ∆νTh with bond angle R for a O-Co-O triatomic system.

It must be mentioned here that evaluation of force constants and bond angles for O-Co-O and O-C-O linkages were done considering harmonic vibration potentials neglecting corrections due to anharmonicity (which might be in question considering suggested molecular arrangement) and transitions between vibrational states (which might happen during film transfer). However, as mentioned earlier, due to considerable frequency separation between O-Co-O and O-C-O linkages, they are effectively decoupled systems owing to which anharmonicity can be neglected, even for the suggested two headgroup conformers. Also, the transition between vibrational states can be neglected in the above calculation, since we are comparing FTIR spectra of LB films after transfer onto substrates. Above results say that the two modes of preparation yield two different molecules. Though this difference is in the conformation of headgroups, it means that they are in different hybridization states. This answers the first question raised by us and shows the importance of water surface in facilitating growth of the alternate conformer. We shall now see how supramolecular forces play a stronger role at the water surface in selecting this conformer. B. Hydrocarbon Tail-Tail Interactions. It has been shown that, in LB multilayers comprising two-tailed amphiphiles, molecular conformation allows only tails to take part in supramolecular and interlayer coupling.19 In particular, interlayer hydrocarbon tail-tail interactions take place predominantly via methyl tops.46 The origin of such methyl-methyl interactions was attributed to hypercongujation.14 The present focus is to elucidate the effect of such interactions on molecular structure, using FTIR spectroscopy. Interaction between two methyl tops causes linking of adjacent hydrocarbon chains which may in turn affect C-C skeletal vibrations,25 another very important large amplitude vibration mode associated both with conformation and supramolecular forces. Ordinarily, C-C stretch modes are IR weak26 such that most of the 17 C-C modes in stearic acid41 remain undetected. Simple calculations show that increasing the number of C atoms from 18 (one stearic chain) to 36 (two coupled stearic chains) causes each C-C stretch mode to be doubly degenerate, thereby increasing their intensity. Also, hydrocarbon tail linkage would result in coupling of CH2 wag modes of both chains,39 causing their intensity enhancement. Thus, in our case, increased intensity of both C-C skeletal and CH2 wag modes are expected in CoStn compared to CoStp and StA. FTIR spectra of CoStp-9 (Figure 6a) and CoStn-9 (Figure 6b), showing regions of C-C skeletal vibrations (1000-1200 cm-1) in carboxylate salts,24 are clearly different. Peaks at 1112, 1068, and 982 cm-1 present in CoStn-9, attributed to C-C skeletal modes (Table 3), are enhanced in intensity as well as frequency downshifted compared to those in CoStp-9 and StA (in an all-trans C form).47 The increased intensity of these C-C modes in CoStn are definite indications of hydrocarbon chain coupling. Again frequencies at 1165 and 1213 cm-1 in CoStn-9, attributed to methylene wag, are enhanced in intensity compared to CoStp and StA, consistent with our assumption. However, a second possibility remains that the peaks at 1112 and 1165 cm-1, are indeed absent in StA, in (46) Kitaigorodskii, A. I. Organic Chemical Crystallography; Consultants Bureau: New York, 1961. (47) Holland, R. F.; Nielsen, J. R. J. Mol. Spectrosc. 1962, 9, 436460.

Chemistry at Air/Water Interface

Langmuir, Vol. 25, No. 6, 2009 3527

Figure 9. C K-edge NEXAFS spectra of 1 ML CoStn film for (a) normal and (b) grazing angles of incidence. Experimental data (circles) have been fitted with voigt function (dotted lines). Solid lines are the composite fits. Table 5. Results of NEXAFS Analysis of 1 ML CoStn Filma peak 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

xc

AG

AN

284.1 285.1 286.1 286.6 287.4 287.8 292.5 294.8 296.5 300.3 305.8 314.0 291.0 295.9 298.6

0.08 0.10 0.17 0.42 0.14 0.42 1.43 0.17 1.40 1.20 3.80 0.30 0.09 0.08 0.17

0.06 0.04 0.17 0.41 0.11 0.42 1.55 0.10 1.40 1.04 2.80 0.29

a xc ) peak position. AN (AG) ) peak amplitude at normal (grazing) incidence. xc is in eV, A is in au.

which case, their origin may be attributed to metal-carboxylate interactions in CoSt samples. Even then, CoStn spectra shows clear signatures of enhanced ordering of tails due to supramolecular interaction, and the fact that corresponding peaks follow an ascending intensity scale from StA to CoStp to CoStn indicate that strength of supramolecular forces follow the same scale. This is consistent with interlayer coupling through the end methyl tops, where the tops are staggered with respect to each other, a configuration that has been found to promote hyperconjugation within a molecule.48 CoStn spectra show a prominent band at 881 cm-1 that appears as very weak in spectra of CoStp and StA. This band, also present in the spectra of cadmium stearate LB film at 880 cm-1,13 is characteristic of LB multilayers. It was tentatively assigned to the metal-carboxylate deformation frequency. Although the C-O-Co deformation is different in the CoStp and CoStn (48) Pophristic, V.; Goodman, L. Nature 2001, 411, 565.

samples, which supports the above observation, nevertheless presence of even a very weak peak near this frequency in CoStp raises some amount of uncertainty in this assignment. This is because the methyl rocking frequency lies close to this value as observed in C form of stearic acid.47 The StA spectra shows presence of two very weak peaks at 893 and 880 cm-1, attributed to methyl rock. Thus a second, and probably more justified possibility is that the peak at 881 cm-1 indeed corresponds to methyl rocking mode, with an increased intensity in case of normal LB multilayers. This is because interlayer supramolecular coupling between adjacent methyl tops freezes free rotation of the C-C skeletal planes of adjacent hydrocarbon chains in the CoStn multilayer, compared to that of StA, thereby enhancing the methyl rock mode. Although a detailed analysis is necessary for precise assignment of this band, it can be said that enhancement of this peak, along with the enhancement in skeletal vibration, suggests the enhancement of interlayer forces. Carbon k edge NEXAFS spectra (open circles) of 1 monolayer CoStn deposited on silicon at normal and grazing (20° to substrate) angles of incidence (Figure 9, parts a and b, respectively) give supportive evidence regarding the hydrocarbon tail orientation wrt substrate.49 The experiment was conducted using synchrotron radiation at the BEAR beamline, ELETTRA synchrotron. Details of the experimental mode and analysis are discussed elsewhere.13 The spectra is fitted with voigt function (solid line). Position and amplitude of fitted peaks are given in Table 5. Although individual peak assignments are not done, the spectra is broadly classified as having prominent features around 287 eV, 295 and 305 eV, as observed in LB monolayers of fatty acid salts.50 The two sharp features around 287 eV, characteristic of hydrocarbon chains (Fitted by peaks 1 to 6 in Table 5) are usually assigned to C 1s * transitions.50 However, the same have been attributed f σC-H to C 1s f Rydberg transitions51 and mixed Rydberg/valence * orbitals (of antibonding σC-H ) transitions,52 and so there remains (49) Stohr, J. NEXAFS Spectroscopy; Springer-Verlag: New York, 1996.

3528 Langmuir, Vol. 25, No. 6, 2009

ambiguity regarding their assignment. Features around 295 and 305 eV are shape resonances typically assigned to C 1s f σ*C-C * transitions and C 1s f σC-C′ transitions.50 The amplitudes of all shape resonances (except peak at 292.5 eV), are found to increase in the grazing incidence spectra. From here, we suggest that the hydrocarbon tails are oriented more toward the substrate normal. The C 1s f σ*C-O and C 1s f σ*CdO transitions (typically observed at 296.5 and 303.5 eV in formic acid spectra respectively), are generally weak compared to the C-C transitions as amplitude of carboxylate resonances are about 5% of the same for hydrocarbon resonances. Although our spectra show the presence of weak peaks, which may be due to carboxylate resonances, it is not possible to precisely assign them at this point. Thus we find that the molecules of cobalt stearate formed at the air-water interface show considerably higher supramolecular bonding than those formed by the reaction in bulk. The enhancement in supramolecular interaction comes from more flexibility in orientation of the chains, which in turn, is brought about by the flexibility in the C-O-Co angle or θ. In case of CoStn deposition, θ is not fixed beforehand as the formation of cobalt stearate occurs during the deposition process at the air/ water interface itself and the ‘twisted’ conformer is chosen by the supramolecular forces to culminate in the FM growth mode with low free energy. This is manifested in the mesoscopic scale as “perfect” LB multilayers.13,14 In contrast, for bulk reaction, CoSt with the “boat” conformer as headgroup (CoStb and CoStp) is formed. As we have shown, this is a rigid structure that remains unchanged throughout. Hence, θ is frozen, leading to the orientational rigidity of the chains and lessening the role of supramolecular forces. This is manifested mesoscopically by the fact that LB multilayers cannot be deposited on this rigid structure. On repeated dipping of the substrate, increase in coverage is observed in CoStp, but no stable multilayers are formed. This explanation answers our second question.

Conclusion and Outlook We have shown, through Fourier transform infrared (FTIR) spectroscopy that, cobalt stearate prepared in bulk and at air/ water interface have identical metal ion-carboxylate coordination, viz. the bidentate bridged configuration. However, both X-ray reflectivity (XRR) and atomic force microscopy (AFM) results show that preformed cobalt stearate film (CoStp) is deposited as a V-W type monolayer in contrast to the ones at

Mukherjee et al.

air/water interface (CoStn), the latter having multilayers with good out of plane crystallinity and near perfect defect free morphology. This difference in their mesoscopic behavior is associated with different conformations of their Co-bearing headgroups. FTIR study reveals that CoStp headgroup (and also CoStb) exist as a ‘boat’ conformer with linear O-Co-O linkage, whereas CoStn forms a lower energy “twisted” conformer with a bent O-Co-O configuration having a bond angle of 105°. The flexible headgroup structure in CoStn allows weak supramolecular interactions between methyl tops at chain-ends. These weak interactions couple the hydrocarbon chains end-to-end and order their skeletal planes, thereby reducing free energy and leading to excellent Langmuir-Blodgett multilayers. On the other hand, the headgroup structure in CoStp (or CoStb) is very rigid and inhibits supramolecular bonding. As a result, multilayer growth cannot take place. Such adverse effect of monolayer rigidity on multilayer formation has been observed elsewhere53 and our results thus support a general notion in this regard. The importance of our present work lies in demonstrating a definite role of the air-water interface in selecting bond coordination and conformation, on one hand and relating morphology in the mesoscopic length scale to bonding molecules via molecular conformations that allow or inhibit specific supramolecular forces and large amplitude motions, on the other. The first point is presumably connected to the observation of hydronium ion enrichment at water surface,54 while the second maybe utilized to tune self-organization of complex structures by the low-energy modes of large amplitude motions. Acknowledgment. The authors greatly acknowledge Dr. Sarathi Kundu for his help during the bulk sample preparation. We wish to thank Prof. Satyajit Hazra for the availability of the VXRD facility in carrying out XRR measurements. We gratefully acknowledge Prof. Debabrata Ghose and Mr. Puneet Mishra for their time and support in carrying out AFM measurements. LA8023502 (50) Outka, D. A.; Stohr, J.; Rabe, J. P.; Swalen, J. D. J. Chem. Phys. 1987, 88, 4076. (51) Bagus, P. S.; Weiss, K.; Schertel, A.; W Braun, W.; Hellwig, C.; Jung, C. Chem. Phys. Lett. 1996, 248, 129. (52) Hitchcock, A. P.; Ishii, I. J. Electron Spectrosc. Relat. Phenom. 1987, 42, 11. (53) Malcharek, S.; Hinz, A.; Hilterhaus, L.; Galla, H. J. Biophys. J. 2005, 88, 2638. (54) Petersen, P. B.; Saykally, R. J. Chem. Phys. Lett. 2008, 458, 255.