X-ray and Neutron Reflectometry Investigation of Langmuir−Blodgett

Yubao Zhang,† Zin Tun,‡ and Anna M. Ritcey*,†. Department of Chemistry and ... All of the cellulose ethers form stable monolayers at the air-wat...
0 downloads 0 Views 166KB Size
Download PDF
Langmuir 2004, 20, 6187-6194

6187

X-ray and Neutron Reflectometry Investigation of Langmuir-Blodgett Films of Cellulose Ethers Yubao Zhang,† Zin Tun,‡ and Anna M. Ritcey*,† Department of Chemistry and CERSIM, Laval University, Quebec, Canada G1K 7P4, and Neutron Program for Materials Research (NRC), Chalk River Laboratories, Chalk River, Ontario, Canada K0J 1J0 Received September 30, 2003. In Final Form: February 12, 2004 Langmuir-Blodgett films of a series of cellulose ethers are investigated by X-ray and neutron reflectometry. Two types of samples are considered: simple alkyl ethers of cellulose and derivatives obtained by the alkylation of (2-hydroxypropyl)cellulose (HPC). All of the cellulose ethers form stable monolayers at the air-water interface. Significant differences are, however, found in the surface pressure-area compression isotherms. Ethers prepared from HPC typically exhibit larger limiting molecular areas and higher surface pressures than the corresponding simple cellulose ethers. The ease of monolayer transfer to hydrophobic silicon substrates differs greatly from one type of molecule to another. Successful transfer conditions are found only for ethers that form stable monolayers at sufficiently high surface pressures. Surprisingly, deuterated HPC ethers, prepared for neutron reflectivity measurements, exhibit monolayer properties significantly different from those of their hydrogenated analogues. Although essentially identical limiting molecular areas are found, the surface pressure corresponding to a characteristic plateau transition in the compression isotherm is found to decrease by about 8-10 mN m-1 upon side chain deuteration. X-ray reflectivity results show a linear increase of film thickness with the number of deposited layers, indicating a regular and reproducible transfer. Observed average layer spacings are, however, significantly smaller than the calculated length of fully extended side chains. Neutron reflectivity curves recorded for composite LB films composed of both deuterated and hydrogenated polymers exhibit regular Keissig fringes, but no Bragg peak. This result indicates that these LB films do not possess an internal periodic structure and the initial layer-by-layer organization is lost by large interlayer diffusion.

1. Introduction The role of molecular scale fabrication of functional materials in the development of new technologies is of ever increasing importance. One of the most frequently employed methods for the preparation of well-defined model molecular assemblies is the Langmuir-Blodgett (LB) technique.1 Although initially developed for the study of small-molecule amphiphiles, many researchers have applied this technique to the fabrication of ultrathin polymeric films, with the hope that the favorable mechanical properties of polymers will impart increased stability to the resulting LB films. The present paper reports our recent investigation of LB films formed by a series of cellulose ethers. These macromolecules are not typical amphiphiles but are characterized rather as “hairyrod” polymers. The concept of using rigid-rod-like polymers, substituted with flexible alkyl side chains, to construct LB films has been applied to several types of macromolecules, including cellulosics2-7 and polyglutamates.8,9 * To whom correspondence should be addressed. † Laval University. ‡ Neutron Program for Materials Research (NRC). (1) Ulman, A. An Introduction to Ultrathin Organic Films; Academic Press: New York, 1990. (2) Kawaguchi, T.; Nakahara, H.; Fukuda, K. J. Colloid Interface Sci. 1985, 104, 290. (3) Schaub, M.; Fakirov, C.; Schmidt, A.; Lieser, G.; Wenz, G.; Wegner, G.; Albouy, P.-A.; Wu, H.; Foster, M. D.; Majrkzak, C.; Satija, A. Macromolecules 1995, 28, 1221. (4) Basque, P.; de Gunzbourg, A.; Rondeau, P.; Ritcey, A. M. Langmuir 1996, 12, 5614. (5) Itoh, T.; Tsujii, Y.; Suzuki, H.; Fukuda, T.; Miyamoto, T. Polym. J. 1992, 24, 641. (6) Buchholz, V.; Adler, P.; Ba¨cker, M.; Ho¨lle, W.; Simon, A.; Wegner, G. Langmuir 1997, 13, 3206.

We have previous reported the behavior of cellulose ethers spread as monolayers at the air-water interface.4,7,10 In general, the surface pressure-area compression isotherms of this class of cellulose ethers exhibit a liquid analogous phase followed by a transition region of relatively high compression at constant pressure. The characteristic surface pressure at which this transition occurs is found to vary systematically with side chain length4 and subphase temperature7 and can be attributed to the formation of bi- or multilayers. Spread monolayers of cellulose ethers have also been investigated by polarization modulated infrared reflection-absorption spectroscopy (PM-IRRAS).11,12 These studies support a model for the monolayer in which the cellulose backbone lies flat at the air-water interface and the alkyl side chains adopt a net orientation normal to the water surface. Furthermore, in the case of the hexadecyl ether of (2-hydroxypropyl)cellulose (denoted as HPC-C16), the positions of the symmetric and antisymmetric methylene stretching vibrations11 correspond to the frequencies characteristic of an alkyl chain in a near all-trans conformation.13 PM-IRRAS spectra of HPC-C16, spread as a monolayer at the air-water interface, also exhibit bands at 1472 and 1463 cm-1. These bands can be attributed to the methylene scissoring vibration, and the (7) Xiao, Y.; Ritcey, A. M. Langmuir 2000, 16, 4252. (8) Embs, F.; Funhoff, D.; Laschewsky, A.; Licht, U.; Ohst, H.; Prass, W.; Ringsdorf, H.; Wegner, G.; Wehrmann, R. Adv. Mater. 1991, 3, 25. (9) Duda, G.; Schouten, A. J.; Arndt, T.; Lieser, G.; Schmidt, G. F.; Bubeck, C.; Wegner, G. Thin Solid Films 1988, 159, 221. (10) Fischer, P.; Brooks, C. F.; Fuller, G. G.; Ritcey, A. M.; Xiao, Y.; Rahem, T. Langmuir 2000, 16, 726. (11) Xiao, Y.; Bourque, H.; Pe´zolet, M.; Ritcey, A. M. Thin Solid Films 1998, 327, 299. (12) Desbat, B.; Mao, L.; Ritcey, A. M. Langmuir 1996, 12, 4754. (13) Mendelsohn, R.; Brauner, J. W.; Gericke, A. Annu. Rev. Phys. Chem. 1995, 46, 305.

10.1021/la0303658 CCC: $27.50 © 2004 American Chemical Society Published on Web 06/25/2004

6188

Langmuir, Vol. 20, No. 15, 2004

Zhang et al.

Figure 1. Structures of the two types of cellulose ethers considered in this study: (a) simple alkyl ethers with R ) n-hexyl (Cel-C6), n-dodecyl (Cel-C12), or a mixture of methyl and n-pentyl (Cel-MP); (b) alkyl ethers of HPC with R ) n-hexyl (HPC-C6), n-dodecyl (HPC-C12), and n-octadecyl (HPC-C18). Because of the possible sequential addition of propylene oxide moieties during the synthesis of HPC, not all side chains of HPC are necessarily of the same length; this is illustrated here by a second oxypropylene moiety placed arbitrarily at position 2 of the anhydroglucose ring. Table 1. Sample Identification and Structural Characteristics of the Various Cellulose Ethers Considered in This Studya sample

backbone

alkyl chain

DS

Πplateau (mN/m)

Πtransfer (mN/m)

τ

Cel-MP (ref 4) Cel-C6 (ref 4) Cel-C12 (ref 4) HPC-C6 HPC-C6d HPC-C12 HPC-C12d HPC-C18 (ref 7)

cellulose cellulose cellulose HPC HPC HPC HPC HPC

methyl; pentyl hexyl dodecyl hexyl hexyl-d13 dodecyl dodecyl-d25 octadecyl

∼3.0 ∼3.0 ∼3.0 >2.5 >2.5 >2.5 >2.5 2.5

29 16 10 36 29 37 33

25 13 8 33 28 33 27 33

>0.9 0.2 0.2 >0.9 ∼0.9b >0.9 ∼0.9b ∼0.8

a The surface pressure at the transition plateau of the monolayer compression isotherm, Π plateau, the surface pressure at which monolayer transfer was carried out, Πtransfer, and the observed transfer ratios, τ, are also provided. b Transfer observed only upon substrate withdrawal.

observed splitting of about 10 cm-1 is indicative of a perpendicular orthorhombic subcell.14 The observation of these bands thus provides further evidence that a high degree of side chain order is achieved in the spread monolayer. Transmission FT-IR spectra7 of the same cellulose ether transferred as an LB film to ZnSe show a series of four evenly spaced bands appearing in the region between 1400 and 1250 cm-1. These bands can be assigned to the methylene wagging band progression and arise from the coupling of oscillations in adjacent CH2 groups. This coupling is observed only for hydrocarbon chains in an all-trans conformation,15 indicating that side chain order is maintained in the transferred layers. The present study extends our previous work through the characterization of LB films of cellulose ethers by the techniques of X-ray and neutron reflectivity. 2. Experimental Section 2.1. Materials. Three different types of cellulose ethers are investigated in this study. Simple alkyl ethers of the general structure shown in Figure 1a were prepared from cellulose acetate (Aldrich, Mw ) 30 000, acetyl content ) 39.8%) and alkyl bromides under basic conditions in DMSO as previously reported.4 A mixed ether, (methyl)2(pentyl)cellulose, was similarly prepared from (methyl)cellulose (Aldrich, Mw ) 40 000, acetyl content ) 39.8%) and pentyl bromide.4 Ethers of the general structure shown in Figure 1b were prepared from (2-hydroxypropyl)cellulose (HPC) (Aldrich, molecular weight ) 100 000, MS ) 4, DS ) 3.0) and alkyl bromides in the presence of potassium tertbutoxide in tetrahydrofuran.7 All starting materials and solvents were obtained from Aldrich with the exception of 1-bromododecane-d25 and 1-bromohexane-d13 which (14) Cameron, D. G.; Casal, H. L.; Mantsch, H. H. Biochemistry 1980, 19, 3665. (15) Umemura, J.; Cameron, D. G.; Mantsch, H. H. Biochim. Biophys. Acta 1980, 602, 32.

were purchased from CDN Isotopes Inc. The degree of substitution (DS), defined as the average number of substituents per anhydroglucose ring, was determined as previously described.4,7 A list of the samples is provided in Table 1. 2.2. Spread Monolayers and Langmuir-Blodgett Films. Cellulose ethers were spread as dilute (∼0.5 mg mL-1) chloroform solutions on deionized filtered water (18.3 MΩ cm, Nanopure II, Barnstead) in a KSV model 3000 film balance. After allowing 15 min for solvent evaporation, the monolayer was compressed at a constant barrier rate of 10 mm min-1 (∼10 Å2 per repeat unit min-1). The surface pressure was monitored by an electrobalance with a platinum Wilhelmy plate. Monolayers were transferred to substrates by the vertical dipping method at constant surface pressure. Unless otherwise specified, deposition speeds of 10 mm min-1 were employed for both the up- and downstrokes, with no delay between dipping cycles. Typical transfer ratios (τ) were calculated from the substrate area and the reduction of the surface area of the trough with each stroke. The substrates (glass slides or silicon wafers) were cleaned in a heated bath of 70% H2SO4/30% H2O2. Si wafers were rendered hydrophobic by treatment with a 10% aqueous solution of HF for about 2 min at room temperature. Hydrophobic glass slides were prepared by placing the substrates in a 5% chloroform solution of hexamethyldisilazane at 60 °C for about 30 min. 2.3. X-ray Reflectivity. X-ray reflectivity measurements were performed with a Rigaku X-ray diffractometer. A Rigaku Rotaflex RU-200BH rotating anode operating at 55 kV/190 mA with Ni-filtered Cu KR radiation λ ) 1.542 Å was employed as the source. A θ-2θ reflection geometry was accomplished with a grazing incidence reflectometry stage constructed in our laboratory. The overall film thickness, L, was determined from the X-ray

LB Films of Cellulose Ethers

Langmuir, Vol. 20, No. 15, 2004 6189

Table 2. Scattering Length Densities (SLD) for the Alkyl Chains and HPC Backbone Employed for the Calculation of Neutron Reflectivity Curves species

M (g mol-1)

∑bi (10-5 Å)

SLD (10-6 Å-2)

HPC-C12H25 -C12D25 HPC-C12H25 HPC-C12D25 Si wafer

364 169 194 871 946

44.7 -13.75 246.5 3.45 784

0.74 -0.49 7.6 0.024 5.0 2.7

reflectivity curves according to eq 1:16

L ) (2n - 1)π/qmin

(1)

where qmin is the scattering vector at the nth order reflected intensity minimum. 2.4. Neutron Reflectivity (NR). The neutron reflectivity measurements were performed with the C5 spectrometer of the DUALSPEC facility, Chalk River. All samples were prepared at most 3 days before the start of the measurements and stored at room temperature. The neutron beam, with a wavelength of 2.3705 Å, was Bragg reflected by a pyrolitic graphite (PG) monochromator and filtered with a PG filter. An approximately constant beam footprint was maintained on the sample during θ-2θ scans by a pair of motorized slits in the incident beam. This approach leads to an effective wavelength spread and a beam collimation of ∆λ ) 0.15 Å and ∆θ ) 0.000 36° (full width at half-maximum (fwhm)), respectively. Reflectivity profiles were fitted with a nonlinear leastsquares program, MLAYER.17 The LB film was divided into a number of layers of specific thicknesses, scattering length densities (SLDs), and roughnesses. The best fit to the experimental data was obtained by refining these three parameters for each layer within the thin film. Scattering length densities of the cellulose ethers and the related alkyl chains and backbone were estimated according to eq 2 and are listed in Table 2.

SLD )

∑i biNAd M

(2)

where the sum is over all isotopic constituents of the molecule of molecular weight M, and bi is the coherent scattering length of isotope i, NA ) 6.02 × 1023 mol-1, and d is the mass density of the material. Not knowing the actual mass density of the films, the initial value for the refinement was obtained by assuming d ) 1 g cm-3. 3. Results and Discussion 3.1. Synthesis and Characterization of Cellulose Ethers. Simple alkyl ethers, of the general structure shown in Figure 1a, can be prepared by the reaction of alkyl bromides with cellulose acetate under basic conditions.4 Similarly, mixed side chain derivatives can be conveniently prepared by the reaction of alkyl halides with the free hydroxyl groups of commercially available low DS cellulose ethers, such as methyl cellulose. Unfortunately, these synthetic routes are only appropriate for the addition of alkyl side chains containing 12 carbons or fewer. In the case of longer side chains, neither the starting alkylating agents nor the resulting cellulose ethers are (16) Kjar, K.; Als-Nielsen, J.; Helm, C. A.; Tippmann-Krayer, P.; Mohwald, H. Thin Solid Films 1988, 159, 17. (17) Ankner, J. F.; Majkrzak, C. F. Neutron Opt. Devices Appl. 1992, 260, 1738.

Figure 2. Surface pressure-area isotherms recorded upon the compression of spread monolayers of the various cellulose ethers identified in Table 1. All isotherms were recorded at 20 °C and at a constant barrier speed of 10 mm min-1.

sufficiently soluble in the polar reaction medium to yield highly substituted products. For this reason, long side chains (exceeding 12 carbon atoms) were instead grafted onto a soluble cellulose derivative, HPC.7 In the present study, this procedure is employed for the preparation of two deuterated HPC derivatives for neutron reflectivity studies. The degree of alkyl substitution of each cellulose derivative was evaluated by 1H NMR and infrared spectroscopy as previously described.4,7 The results are summarized in Table 1. The cellulose ethers are found to be fully substituted (DS ∼ 3), whereas ethers prepared from HPC typically contain some residual unsubstituted hydroxyl groups, as evidenced by the relatively weak, but present, O-H stretching absorption in the IR spectra. Importantly, the deuterated ethers prepared in this study are found to have DS values comparable to those of the corresponding hydrogenated samples. 3.2. Surface Pressure-Area Isotherms. The surface pressure-area (π-A) compression isotherms recorded for spread monolayers of HPC-C6 and HPC-C12 are shown in Figure 2. Compression isotherms for Cel-C6, Cel-C12, and HPC-C18 have been previously reported4,7 and are reproduced here for comparison. Except for HPC-C18, all of the curves exhibit a phase transition characterized by a constant pressure plateau. This transition has been previously discussed4,7,10 in some detail and can be attributed to the formation of bi- or multilayers as molecules in the first monolayer leave the water surface. The absence of a constant pressure plateau for HPC-C18 has been related to side chain crystallization as discussed elsewhere.7 The compression isotherms of the two types of samples exhibit some significant differences. Limiting repeat unit areas, as evaluated from linear extrapolation of the first increase in surface pressure observed upon compression, are found to be larger for the HPC derivatives (∼140 Å2) than for the simple cellulose ethers (75-115 Å2), despite the consistently higher DS values of the latter. Furthermore, a stronger dependence on side chain length is found

6190

Langmuir, Vol. 20, No. 15, 2004

for the limiting areas of the Cel-Cn samples. The area required by three highly ordered alkyl chains, oriented perpendicular to the surface,18 can be evaluated as 55 Å2, whereas a single anhydroglucose ring19 can be estimated to occupy about 60 Å2. These dimensions are smaller than the limiting areas observed for the simple cellulose ethers, indicating some side chain disorder. Furthermore, the variation of molecular area with side chain length implies that the limiting area is determined by side chain interactions rather than the contact of neighboring polymer backbones. In the case of the HPC-based ethers, on the other hand, the observations support a monolayer model in which all of the ether linkages, including those originating from the hydroxypropyl moieties in the parent polymer, are located within the air-water interface. The presence of additional hydrophilic ether linkages within the side chains of the HPC derivatives would thus lead to increased molecular areas with respect to the simple cellulose derivatives and allow sufficient space to accommodate even relatively poorly ordered side chains. This model predicts that the limiting areas of the HPC derivatives should be independent of side chain length, as is indeed reflected by the strikingly similar π-A isotherms observed for HPC-C6 and HPC-C12. The compression isotherms for two types of cellulose ethers also differ in the surface pressure at which the transition to bi- or multilayers occurs. The higher surface pressures found for the HPC ethers are consistent with the increased number of hydrophilic groups present at the air-water interface. The resulting increased attraction for the aqueous subphase renders the removal of molecules from the surface more difficult, and the transition plateau consequently appears at higher surface pressures. Compression π-A isotherms were also recorded for spread monolayers of the two deuterated samples, HPCC12d and HPC-C6d. Surprisingly, these isotherms are significantly different from those of the hydrogenated analogues, as illustrated in Figure 3. Although little change in limiting area is observed upon deuteration, the plateau surface pressure drops by 8-10 mN m-1 with respect to the hydrogenated samples. Deuterium substitution is frequently employed for the study of molecular organization by NMR, IR, or neutron scattering methods and is generally assumed to have little effect on the physical chemical properties of a sample. This assumption, however, is not always justified. Blaudez et al.20 have noted, for example, that LB films of deuterated cadmium arachidate present a higher degree of order than the corresponding hydrogenated samples. One of the best documented examples of the effect of deuteration on physical properties involves the well-known liquid crystalgel phase transition of lipids. Deuterated lipids typically undergo this transition at temperatures about 4 °C lower than do the corresponding hydrogenated analogues.21 This effect can be attributed to the weakening of the van der Waals attractions between deuterated alkyl chains as a result of the reduction in both the average length and the polarizability of C-D bonds with respect to C-H bonds.22 The lowering of the plateau surface pressure evident in the isotherms of the deuterated HPC ethers presented in (18) Nyburg, S. C.; Lueth, H. Acta Crystallogr., Sect. B 1972, 28, 2992. (19) Giles, C. H.; Agnihotri, V. G. Chem. Ind. 1967, 4, 1874. (20) Blaudez, D.; Buffeteau, T.; Castaings, N.; Desbat, B.; Turlet, J.-M. J. Chem. Phys. 1996, 104, 9983. (21) Guard-Friar, D.; Chen, C.-H.; Engle, A. S. J. Phys. Chem. 1985, 89, 1810. (22) Bates, F. S.; Wignall, G. D. Phys. Rev. Lett. 1986, 57, 1429.

Zhang et al.

Figure 3. Surface pressure-area isotherms recorded upon the compression of spread monolayers of HPC-C12, HPC-C12d, HPC-C6, and HPC-C6d. All isotherms were recorded at 20 °C and at a constant barrier speed of 10 mm min-1.

Figure 4. Typical transfer ratios as a function of layer number for the various cellulose ethers. Monolayers were transferred to hydrophobic silicon wafers at 20 °C and with a dipping speed of 10 mm/min for both the up- and downstrokes. Transfer surface pressures are as indicated in Table 1.

Figure 3 can therefore be attributed to the weaker van der Waals attractions between neighboring side chains. This observation, in turn, implies that lateral side chain interactions contribute significantly to monolayer stability. 3.3. Monolayer Transfer to Solid Substrates. Below the plateau surface pressure, all of the monolayers are typically able to sustain constant surface pressures with reductions in area of less than 2% during a 30 min stabilization period. The formation of a stable monolayer at the air-water interface is not, unfortunately, a sufficient criterion to ensure subsequent successful monolayer transfer to solid substrates. Figure 4 shows typical transfer ratios (τ) for the six hydrogenated samples transferred to

LB Films of Cellulose Ethers

hydrophobic silicon. High-quality Y-type transfer (τ consistently greater than 0.9) is observed for all of the HPC ethers and the mixed ether, Cel-MP. LangmuirBlodgett films containing 80 monolayers were prepared from HPC-C6 and HPC-C12 with consistent transfer ratios near 1. Thicker films can probably be prepared; the upper limit was not established. Less satisfactory results were obtained for HPC-C18, with transfer ratios around 0.85 being typically found. In the case of Cel-MP, LB films could be prepared with transfer ratios consistently surpassing 0.90 up to at least 900 layers. Monolayer transfer was much less successful for the simple cellulose ethers (Cel-C6 and Cel-C12). Transfer ratios were found to be typically below 0.6 and decreased with increasing numbers of layers deposited. Attempts to improve transfer ratios by varying the temperature (5-40 °C), the stabilization pressure, or the dipping speed were unsuccessful. Although all of the samples share a similar structure and all form stable monolayers at the air-water interface, the transfer behavior, and consequently the quality of the resulting LB films, differs greatly from one derivative to another. For the limited number of samples considered here, transfer behavior appears to depend primarily on surface pressure. The three samples (HPC-C6, HPC-C12, and Cel-MP) which consistently exhibited transfer ratios near unity are also characterized by relatively high plateau surface pressures. The significantly lower plateau pressure of the other samples (Cel-C6 and Cel-C12) limits the surface pressures at which transfer can be carried out to 13 and 9 mN m-1, respectively. As noted above, poor transfer ratios were obtained in these cases. At present, the reasons why monolayer transfer is successful for one polymer and not for another are not well understood. The importance of good adhesion between the first transferred layer and the substrate, as well as sufficient cohesion between successively deposited monolayers, has been evoked by several authors.18,23 The importance of monolayer elasticity has also been noted.23 In the present case, we can expect that the stronger hydrophilic nature of the backbone in polymers derived from HPC and methylcellulose would lead to stronger interlayer cohesion at the backbone-backbone interface. This is the interface created during substrate withdrawal. The transfer characteristics plotted in Figure 4, however, indicate that poor transfer ratios are observed for Cel-C6 and Cel-C12 on both the upstroke and the downstroke, including the deposition of the first layer on the hydrophobic substrate. The differences in transfer behavior between HPC-C12 and Cel-C12 and between HPC-C6 and Cel-C6 observed during the downstroke are difficult to correlate with polymer structure, since the hydrophobic side chains are identical for the two families of cellulose derivatives. In fact, the denser monolayer observed for the simple cellulose ethers would suggest that transfer during the downstroke would be more favorable than for the corresponding HPC derivatives. Through the careful measurement of the interfacial forces applied to the moving substrate during the transfer process, Egusa et al.24 have been able to establish a relationship between the transfer ratio and the changes in surface energies caused by the vertical displacement of the substrate. These authors have demonstrated that for a given substrate hydrophilicity, there exists a welldefined critical surface pressure at which the transfer ratio undergoes a transition from zero to unity. The (23) Hagting, J. G.; de Vos, R. E. T. P.; Kinkovics, K.; Vorenkamp, E. J.; Schouten, A. J. Macromolecules 1999, 32, 3939. (24) Egusa, S.; Gemma, N.; Azuma, M. J. Phys. Chem. 1990, 94, 2512.

Langmuir, Vol. 20, No. 15, 2004 6191

Figure 5. Transfer ratios for a multilayer LB film prepared by alternate deposition of HPC-C12 and HPC-C12d. Odd and even layer numbers correspond to the down- and upstroke, respectively. Monolayers were transferred to hydrophobic silicon wafers at 7 °C and with a dipping speed of 5 mm/min. Transfer surface pressures are as indicated in Table 1.

transfer mechanism described by Egusa et al.24 thus predicts a correlation between surface pressure and the quality of monolayer transfer, as is observed in the current work. The correlation between transfer surface pressure and the quality of monolayer transfer is even more striking in the case of the deuterated HPC ethers. Figure 5 illustrates the transfer characteristics observed during the preparation of a multilayer LB film, denoted as Si/(2H/1D)11, prepared by the alternate deposition of HPC-C12 and HPC-C12d. To prepare this sample, the substrate was alternately passed through a monolayer of HPC-C12 at a surface pressure of 33 mN m-1 and a monolayer of HPCC12d at a surface pressure of 28 mN m-1. Figure 5 indicates good monolayer transfer of the hydrogenated material (filled circles) during both the immersion and emersion of the substrate, as expected for a typical Y-type LB film. The deuterated material, on the other hand, is only transferred during the downstroke (open circles), as is the case for rarer X-type films. These two materials differ only in the isotopic substitution of the hydrogen atoms in the dodecyl side chains. It is difficult to link the observed transfer behavior directly to this structural difference. The transfer characteristics of the two materials differ for the substrate upstroke, for which one would predict the polymer backbone to be the structural unit primarily responsible for the quality of interlayer cohesion. The two materials, however, possess identical polymer backbones. This observation further supports the argument that the surface pressure at transfer, rather than the cohesion between layers, is responsible for the difference in the transfer behavior. This conclusion is convincingly supported by the data presented in Figure 6. This plot illustrates the transfer ratios observed for HPC-C12 during transfer to hydrophobic silicon at two different surface pressures: 28 and 33 mN m-1. This relatively small difference in surface pressure has a striking effect on the transfer behavior, in close agreement with the existence of a critical surface pressure as reported by Egusa et al.24 3.4. X-ray Reflectometry. X-ray reflectivity curves were recorded for LB films transferred to hydrophobic silicon wafers. In general, reflectivity curves showing clear Kiessig fringes were obtained for LB films prepared from samples exhibiting transfer ratios near unity over a sufficient number of deposition cycles (HPC-C6, HPC-

6192

Langmuir, Vol. 20, No. 15, 2004

Zhang et al.

Figure 6. Transfer ratio as a function of layer number for HPC-C12 at two different surface pressures. Odd and even layer numbers correspond to the down- and upstroke, respectively. Figure 8. Overall film thickness, as evaluated from X-ray reflectometry, as a function of the number of deposited layers for Cel-MP, HPC-C6, HPC-C12, and HPC-C18. From the slope of linear fitting, average single layer thicknesses of 12, 18, and 26 Å are found for HPC-C6, HPC-C12, and HPC-C18, respectively.

Figure 7. X-ray reflectivity profiles of LB films of various thicknesses, prepared from HPC-C12 transferred to silicon.

C12, HPC-C18, and Cel-MP). On the other hand, LB films prepared from Cel-C6 and Cel-C12 show irregular X-ray reflectivity profiles, most likely reflecting the irregular, rough films that result from poor monolayer transfer. X-ray reflectivity profiles obtained for LB films prepared from HPC-C12 are presented in Figure 7. Regular Kiessig fringes are clearly evident, with the fringe spacing found to decrease with increasing number of deposited layers. Up to 20 Kiessig fringes appear in the curve recorded for a 30 layer sample, indicating the smooth surface and the uniform structure of the film. Despite the presence of welldefined Kiessig fringes, no Bragg peak is observed in the reflectivity profiles. Similar X-ray reflectivity profiles, that is, clear Kiessig fringes but no Bragg peaks, were observed for LB films of HPC-C6, HPC-C18, and Cel-MP. The absence of a Bragg peak does not necessarily indicate that the LB films do not possess an internal periodic layered structure. Schaub et al.3 have demonstrated that in the case of LB films prepared from highly substituted (isopentyl)cellulose, the absence of a Bragg peak in X-ray reflectivity curves can be attributed to the low electron density contrast between the side chains and the polymer backbone. The overall film thicknesses, as evaluated from the fringe spacings of the X-ray reflectivity data according to eq 1, are plotted in Figure 8 as a function of the number of deposited layers for LB films of Cel-MP and the three HPC derivatives. The observed linear relationships, which essentially pass through the origin, indicate that uniform and reproducible monolayer transfer is achieved upon each

Figure 9. Plot of the average monolayer thickness obtained from X-ray measurements as a function of side chain length for the HPC-C6, HPC-C12, and HPC-C18.

stroke (both up and down) of the dipping cycle. In the case of Cel-MP, linear fitting of the data yields L ) 0.47 + 9.6n, where n is the layer number, and thus indicates an average layer thickness of 9.6 Å. This value compares very well with the layer thickness obtained for LB films of (isopentyl)cellulose.3 The slopes of the linear fitting for the HPC samples yield average single-layer thicknesses of 12 ( 1, 18 ( 1, and 26 ( 1 Å, for HPC-C6, HPC-C12, and HPC-C18, respectively. This average thickness is, in turn, plotted in Figure 9 as a function of the number of carbon atoms in the alkyl side chains, nC. Although only three data points are available and they do not fall perfectly on a straight line, extrapolation to nC ) 0 yields a reasonable value25 of 6 Å for the estimated thickness of the HPC backbone. It is of interest to compare the average layer thicknesses with the calculated lengths of the corresponding fully extended alkyl chains. This comparison reveals that the experimental chain lengths are 10-20% shorter than those predicted for an extended all-trans conformation. Similarly, the slope of 1.1 Å per additional methylene group (25) Tibirna, C.-M. Development of New Supramolecular Polymers of Interest for Non-linear Optical Applications. Ph.D. Thesis, Laval University, Que´bec, Canada, 2003.

LB Films of Cellulose Ethers

Figure 10. Experimental neutron reflectometry profile (open circles) for an LB film prepared by the alternate deposition of HPC-C12 and HPC-C12d, as described in Figure 5. The solid line reflects the best fit for the data and is obtained for the scattering length density profile plotted in the inset.

found for the plot presented in Figure 9 is about 20% less than the 1.26 Å predicted for a fully extended all-trans alkyl chain. The observed layer spacings therefore suggest the presence of gauche isomers in the side chains. This conclusion is, however, contradicted by the infrared spectra previously reported for both spread monolayers11 and LB films7 of HPC-C16. These spectra exhibit both the methylene wagging band progression and methylene stretching frequencies characteristic of an all-trans conformation. The observed reduction in layer thickness with respect to the length of fully extended alkyl chains could also be explained by partial side chain interdigitation. Side chain interdigitation may be favored in the case of the HPC ethers because of the large average areas per repeat unit (∼140 Å2) found for spread monolayers. As discussed above, this area exceeds that required for close packing of three alkyl chains, oriented perpendicular to the airwater interface. 3.5. Neutron Reflectometry. The X-ray reflectivity measurements presented above indicate that uniform, smooth LB films are obtained from the cellulose ethers that exhibit high transfer ratios. The layer-by-layer construction of these films implies that they should possess an internal periodic structure and therefore give rise to a Bragg peak in the reflection data. As noted above, the absence of a Bragg peak in the X-ray data does not preclude the presence of a periodic structure because of the very small differences in electron density between the cellulose backbone and the alkyl side chains. Neutron reflectivity measurements were therefore undertaken. The possibility of deuteration permits the preparation of alternating structures with high contrast in scattering length densities between deuterated and hydrogenated layers. Figure 10 shows the neutron reflectivity profile obtained for a multilayer LB film, denoted Si/(2H/1D)11, prepared by the alternate deposition of HPC-C12 and HPC-C12d as described in Figure 5. Although clear Kiessig fringes indicate the uniform nature of the alternatively deposited LB film, no Bragg peak is found. Attempts to model the neutron reflectivity data with an alternating layer structure failed to capture the reflectivity behavior. Satisfactory fitting was obtained only by assuming the homogeneous film profile illustrated in the insert of Figure 10. The fitted SLD of 1.81 × 10-6 Å-2 is a reasonable average of the predicted values for deuterated and hydrogenated HPCC12, as given in Table 2. The total thickness of 595 Å

Langmuir, Vol. 20, No. 15, 2004 6193

Figure 11. X-ray reflectivity curves recorded for an LB film composed of 12 layers of HPC-C6d (open triangles) and the composite film (open circles) obtained by the subsequent deposition of 12 layers of HPC-C6 on top of the deuterated layer.

deduced from the fringe spacing of the neutron reflectivity curve is very close to that obtained by X-ray reflectivity. Similar results, that is, regular Kiessig fringes but no Bragg peak, were found for an alternately deposited HPCC6/HPC-C6d LB film. The absence of Bragg peaks in the neutron reflectivity data unambiguously indicates that the LB films do not possess an internal periodic structure. The homogeneous films could be the result of rough interlayer interfaces that essentially lead to interlayer mixing. The partial transfer of HPC-C12d during substrate withdrawal, as illustrated in Figure 5, would in fact be predicted to lead to rough interlayer boundaries. Fitting of the neutron reflectivity data provides a roughness of about 4-5 Å for the Si-LB film interface. This value is close to the surface roughness of the Si substrate. The LB film-air interface is more diffuse, with surface roughness values of about 15 Å being typically obtained. These values are comparable to the thickness of a single layer. Another possible explanation for the absence of internal film periodicity is high interlayer diffusion. To determine the extent of interlayer diffusion, an LB film, denoted as Si/12D/12H, was prepared by the initial deposition of 12 layers of HPC-C6d on a silicon substrate and the subsequent transfer of 12 layers of HPC-C6 on top of the deuterated film. The X-ray reflectivity curves for the first 12 layers of HPC-C6d and the final composite film are shown in Figure 11. These data reveal a thickness of the deuterated layer of 150 Å and a total film thickness of 300 Å. The thickness of the 12 layer hydrogenated film can thus be deduced to be 150 Å. The corresponding neutron reflectivity curve, shown in Figure 12, exhibits regular interference fringes with a spacing corresponding to a film thickness of ∼300 Å. This result provides strong evidence for interlayer diffusion. Since the SLD of the hydrogenated sample (HPC-C6) is very small (0.19 × 10-6 Å-2) and close to that of air, the top layer of the film should be invisible to the neutron beam. If the film had remained as prepared, that is, composed of a discrete hydrogenated layer deposited on top of a deuterated film, the fringe spacing of the neutron reflectivity should reflect the thickness of the highly scattering HPC-C6d layer (150 Å). Satisfactory fitting (illustrated by the solid line) of the experimental data can

6194

Langmuir, Vol. 20, No. 15, 2004

Zhang et al.

contradict the high degree of order suggested by previously reported IR spectra7 of LB films prepared from HPC-C16. Preliminary IR measurements obtained for LB films of HPC-C6 and HPC-C12 indicate that side chain ordering does, however, decrease with decreasing side chain length. Small-molecule LB films have been shown to exhibit high interlayer diffusion27,28 without disruption of their ordered periodic structure. This can be thought of as analogous to liquid crystals, where molecular motion and molecular order are simultaneously present. 4. Conclusions

Figure 12. Experimental neutron reflectivity profile (open circles) obtained for an LB film composed of 12 layers of HPCC6 deposited on top of 12 layers of HPC-C6d (sample denoted as Si/12D/12H). The solid line represents the best fit to the data, obtained for the scattering length density profile plotted in the insert. The dashed line illustrates the calculated profile assuming that a sharp interface is maintained between the deuterated and hydrogenated layers.

be generated by the scattering length density profile provided in the inset. This profile clearly corresponds to a homogeneous film, resulting from the complete intermixing of the hydrogenated and deuterated layers. The fitted SLD of 1.68 × 10-6 Å-2 is very close to the weighted average of the SLD values calculated for HPC-C6d and HPC-C6, based on their chemical structures. The dashed line of Figure 12 shows the neutron reflectivity curve that would have been obtained had the hydrogenated and deuterated layers of the sample remained separated. Clearly, it is inconsistent with the observed data. The neutron reflectivity results unambiguously demonstrate that interlayer diffusion occurs in the LB films HPC-C6 and HPC-C12. This behavior is strikingly different from that observed for LB films of other hairy-rod polymers. For example, X-ray and neutron reflectometry studies26 of LB films fabricated from poly[(γ-methyl L-glutamate)-co-(γ-n-octadecyl L-glutamate)] indicated no interlayer diffusion, even at temperatures exceeding 100 °C. The conclusion that rapid interlayer mixing occurs in LB films of the HPC ethers furthermore appears to (26) Schmidt, A.; Mathauer, K.; Reiter, G.; Foster, M. D.; Stamm, M.; Wegner, G.; Knoll, W. Langmuir 1994, 10, 3820.

Langmuir-Blodgett films can be successful prepared from alkyl ethers of HPC. When spread as monolayers at the air-water interface, HPC ethers typically exhibit larger limiting molecular areas and higher surface pressures than the corresponding simple cellulose ethers. The ease of monolayer transfer to hydrophobic silicon substrates also differs greatly from one type of molecule to another, and successful transfer can be correlated with the formation of stable monolayers at sufficiently high surface pressures. Film thicknesses, determined by X-ray reflectometry, are found to increase linearly with the number of deposited layers for LB films of all of the HPC ethers considered. The observed average layer spacings are, however, significantly smaller than the calculated length of fully extended side chains. This observation suggests side chain disorder or significant side chain interdigitation between neighboring layers. In the case of hexyl and dodecyl HPC ethers (HPC-C6 and HPC-C12), neutron reflectometry curves recorded for composite LB films composed of both deuterated and hydrogenated polymers exhibit regular Keissig fringes but no Bragg peak. This result indicates that these LB films do not possess an internal periodic structure and the initial layerby-layer organization is lost by large interlayer diffusion. Similar interlayer diffusion may not occur in derivatives containing longer side chains. Acknowledgment. The financial support of the NSERC (Canada) and FCAR (Que´bec) is gratefully acknowledged. We gratefully acknowledge Michael Wang for stimulating discussions and Y. Li for assistance with the neutron reflectivity measurements. LA0303658 (27) Shimomura, M.; Song, K.; Rabolt, J. F. Langmuir 1992, 8, 887. (28) Feigin, L.; Konovalov, O.; Wiesler, D. G.; Majkrzak, C. F.; Berzina, T.; Troitsky, V. Physica B 1996, 221, 185.