Supporting Information Ethylenediamine at Air/Liquid and Air/Silica Interfaces: Protonation Versus Hydrogen Bonding Investigated by Sum Frequency Generation Spectroscopy MAN XU, DINGFANG LIU, AND HEATHER C. ALLEN * Department of Chemistry, The Ohio State University, West 8 th Ave., Columbus, OH 43 Phone (64) 9-477; Fax (64) 9-685; Email allen@chemistry.ohio-state.edu NAME OF JOURNAL: Environmental Science & Technology DATE OF SUBMISSION: 8//5 NUMBER OF PAGES: 3 NUMBER OF FIGURES: 6 NUMBER OF TABLES: S
SFG Spectra of EDA Aqueous Solutions (Polarization Combination sps) SFG Intensity (a.u.) 6 4 a Neat Water, sps SFG Intensity (a.u.) 5 b.5 m.f. EDA, sps SFG Intensity (a.u.) 5 c. m.f. EDA, sps 8 3 3 34 36 38 Incident Infrared (cm - ) Figure S. SFG spectra of a) neat water (scaled to the EDA experimental conditions); b).5 m.f. ethylenediamine at the air/liquid interface; c). m.f. ethylenediamine at the air/liquid interface with the sps polarization combination. Curve-fitting of SFG Spectra When several overlapping peaks contribute to the SFG spectral intensity, it is difficult to assign the accurate peak positions. Curve-fitting techniques can be used to determine the positions, amplitudes S
and areas of the component peaks of a spectrum. In this paper, SFG spectra were analyzed with the peak-fitting function in Igor Pro 4.5A software using a home-written program to incorporate phase. The mathematical model describing the profile of bands is the Lorentzian model. Before performing curve-fits on the SFG spectra, curve-fitting of IR and Raman spectra of EDA and protonated EDA solutions were completed as references. During fitting processes of IR and Raman spectra, the secondderivative curves of the spectra were used to assist in determining the starting values of the peaks and their positions before each fitting operation. The minimum of the second-derivation indicates a change in the slope of the original data and, therefore, the position of the potential peaks. However, direct comparison of SFG spectra to IR and Raman spectra may be misleading. The IR or Raman intensity is the summation of each vibration s intensity, and, consequently, IR and Raman are non-coherent and there is no phase. In contrast, SFG is coherent. The SFG intensity (I SFG ) is proportional to the square of the summation of resonant (χ () v ) and non-resonant (χ () NR ) sum frequency terms as shown below in equations. and. (). I SFG ( ) v ( ) ( ) χ = χ NR + χ (.) v A v χ ( ) v (.) ωv ωir iγv where χ () is the macroscopic second-order nonlinear susceptibility, which consists of resonant terms (χ () v ) and a non-resonant term (χ () NR ). A v refers to the amplitude of the transition moment, ω v is the frequency of the transition moment, and Γ v describes the line-width of the transition. As a result, the relative phase of the SFG response and different non-resonant parameters are included in the curvefitting of SFG spectra. Figure S and Table S show the curve-fitting results of SFG spectra in the CH stretching region. S3
3 a SFG Intensity SFG Intensity SFG Intensity 4.7 m.f. EDA 3 at ph 9.7 (Singly Protonated) 5 5.5 m.f. EDA.5 m.f. EDD (Doubly Protonated) b Silica with EDA (5 minutes purging) Silica with EDA (4 hours purging) x x 75 8 85 9 95 (cm - ) 8 6 4 SFG Intensity 5 4 3 SFG Intensity 7 75 8 85 9 95 Incident Infrared (cm - ) Figure S. SFG spectra in the CH stretching region of a).5 m.f. ethylenediamine at the air/liquid interface,.7 m.f. singly protonated ethylenediamine at the air/liquid interface, and.5 m.f. ethylenediamine dichloride at the air/liquid interface; b) ethylenediamine at the air/silica interface with 5 minutes purging (component peaks at 85. cm - and 98.9 cm - have been multiplied by ), and ethylenediamine at the air/silica interface with 4 hours purging. Blue circles: experimental; blue solid lines: overall fits; red dashed lines: Lorentzian component peaks. The probe peak is denoted by the red arrow. S4
Table S. Curve-fitting results of SFG spectra in the CH stretching region (all the component peaks are in phase). Non-resonant response Real part Imaginary part.5 m.f. EDA -.8 -.9. m.f. EDA -.9 -.95.7 m.f. singly protonated EDA.5 m.f. EDD (doubly protonated EDA) Silica with EDA (5 min purging) Silica with EDA (4 hr purging) -. -. -.8 -.6 -.9 -.8 -.8 -. Peak position (ω) Amplitude (A) CH -SS Half width (Γ) Area Peak position (ω) Amplitude (A) CH -AS Half width (Γ) Area 85.6 9.9 5.4 334.6 97.7 43.8 7.7 38.5 866. 3. 9.6 67. 98.3.4.9 59.9 849.8.8 7.7 489.7 98..8.7 9. 86. 4.4 9. 536.9 93.4 3.3 6.5 57. 849.6 7.8 7. 33.8 9.9 65.7. 64.5 87. 55..3 83. 97.6 4.9 7.9 84. 858.8 8.8 7.8 38. 94.8 4.8 3.8 35.7 88.6 47. 4. 463.8 934. 37.4 4. 87.9 85..5 7..7 98.9. 5..8 88. 64.8 3.5 53. 93.8 6.. 7.4 848.7 4.9 9.4 39. 96.8 85.9.9 835.8 87. 5..3 96. 93. 49..8 57.9 Speciation of EDA at Different ph The species of EDA vary at different ph. The predominant species can be EDA, one NH protonated (singly protonated) EDA or two NH protonated (doubly protonated) EDA. To calculate the concentrations of each species at different ph values, the following equations are used (): [ ] = BT B (3.) + [ H ] / K K + [ H + ]/ K + a a a [ HB + ] = [ H + ]/ K a B T + + K a /[ H + ] (3.) H B = + T [ ] + + + K a /[ H ] + K ak a /[ H B ] (3.3) where B refers to EDA, HB + is singly protonated EDA, and H B + refers to doubly protonated EDA. For EDA, K a = -.7, K a = -7.56. The percentages of different EDA species at different ph values are S5
shown in Figure S3. By controlling the ph, the singly protonated EDA solution was obtained at ph 9.7, where singly protonated EDA accounts for 95% of all EDA species. 8 Singly protonated EDA Doubly protonated EDA EDA 6 % 4 4 6 8 4 ph Figure S3. Percentages of EDA species at different ph values. Orientation Calculations The capability of determining molecular orientation is a significant advantage of SFG studies (3). Different components of χ () are related to the spectroscopically active components of the molecular hyperpolarizability tensor β () by the average orientation angle of the functional group (3-7). The molecular symmetry determines the nonzero elements of β () (6). In the present work the methylene groups of ethylenediamine are treated as having a C v symmetry. For the methylene group, the nonzero SFG tensor elements of the symmetric stretch are β aac, β bbc and β ccc that are related by β + β = β for methylene at the standard bond angle 9.5 (4, 6, 8). Though the bond angle in aac bbc ccc EDA is not exactly 9.5, the value of β + β β remains a small number close to zero. For the aac bbc ccc antisymmetric mode, β aca equals β caa (4, 6, 8). The abc represents the molecule-fixed coordinate system. S6
The components of χ for the symmetric stretch (SS) are given by (6), χ is described in the surface-fixed (xyz) coordinate system: χxxz _ SS( θ ) = Ns ( β + β + β ) cos( θ ) + ( β + β β ) cos 3 aac bbc ccc Ns aac bbc ccc ( θ ) (4.) 4 4 χxzx _ SS( θ ) = Ns ( β + β β ) [cos( θ ) cos 3 aac bbc ccc ( θ )] (4.) 4 χzzz _ SS( θ ) = Ns ( β + β ) cos( θ ) ( β + β β ) cos 3 aac bbc Ns aac bbc ccc ( θ ) (4.3) For the antisymmetric stretch (AS), the components of χ are given by (5): χxxz _ AS( θ ) = Ns β [cos( θ ) cos 3 aca ( θ )] (4.4) χxzx _ AS( θ ) = Ns β cos 3 aca ( θ ) (4.5) χzzz _ AS( θ ) Ns β [cos( θ ) cos 3 ( θ )] (4.6) = aca where θ is the tilt angle from the surface normal. χxxz _ SS( θ ) and χxxz _ AS( θ ) are related to the SFG spectra using the ssp polarization combination. According to previous work (4), the relations between β ccc, β aac and β caa are β = and β caa =. 53 β aac for CH with C v symmetry. Combined with the relationships aac β ccc β aac + β bbc = β ccc and aca β caa β = shown above, and assuming that Ns β ccc =.ArbitraryUnit( a. u.), equation 4. and 4.4 can be greatly simplified: χxxz _ SS( θ ) = cos( θ ) (4.7) χxxz _ AS( θ ) =.6 [cos( θ ) cos 3 ( θ )] (4.8) Here, absolute values of χ are used for the antisymmetric stretch. χxxz _ AS( θ ) =.6 [cos( θ ) cos 3 ( θ )] (4.9) S7
Figure S4 shows the dependence of second order susceptibilities on the tilt angle. The calculated tilt angle θ from the SFG intensity ratio of the symmetric and antisymmetric stretch are listed in Table S. χ ssp.5..5 a ssp, SS ssp, AS..5 χ ssp SS ssp, AS, / χ. 8 6 b χ ssp SS ssp, AS, / χ 4 4 6 8 θ (deg ree) Figure S4. SFG intensities for the symmetric (SS) and antisymmetric (AS) stretch of methylene (a) and the ratio between them (b) as a function of tilt angle. S8
Table S. Orientation calculation results of SFG spectra. (Error in θ was determined by the error in the reproducible fits.) A(SS) A(AS) A(SS)/A(AS) χss/χas Cos (θ) θ (degree).5 m.f. EDA 5 368 4.8..778 43±5. m.f. EDA 543 57.64.6.6445 5±3 Silica with EDA (5min) 56 7.96.7.6689 48±6 Singly protonated EDA 35 49.99.953.6 86±4 EDD (doubly protonated EDA) 6 639.94.97.473 8± Silica with EDA (4hr) 9 354.953.976.647 8±7 A(SS): the area of CH -SS peaks; A(AS): the area of CH -AS peaks; χss: the second order macroscopic nonlinear susceptibility of CH -SS; χas: the second order macroscopic nonlinear susceptibility of CH -AS. Surface Mole Fraction The adsorption of a solute at the liquid/air interface, hence the surface mole fraction, is characterized by the relative adsorption (Γ, ) (9, ). The relative adsorption is related to the surface tension (γ) through the Gibbs equation (9, ): Γ, γ x( ) T, P = x ln f RT ( + x ) x = Γ Γ x x (5.) where x refers to the bulk mole fraction of ethylenediamine, f is the activity coefficient of EDA in the solution, R is the universal gas constant 8.34 J/(Kmol), T refers to the temperature, Γ and Γ refer to the surface excess concentrations of water and EDA, with the unit µmol/m. In the present paper, the surface tension values of ethylenediamine aqueous solutions were obtained by using a DeltaPi Tensiometer (Kibron Inc.) at ~97 K. According to the linear regression result of the experimental surface tension data (Figure S5a), γ ( ) x T, P is constant which equals 3.. The activity coefficient f for S9
EDA in the binary system of EDA and water is predicted by using the UNIFAC group-contribution method (, 3), as shown in Figure S5b. The relative adsorption Γ, versus bulk mole fraction of EDA is plotted in Figure S5c. Assume the surface molecular areas of water (a ) and EDA (a ) are concentration independent, the following relationship exists based on the monolayer model. 6 N AV ( Γ a + Γa ) = (5.) where Γ and Γ are expressed in µmol/m, a and a are expressed in Å, and N AV is Avogadro s number. a is 8 Å for water according to literature (). a is estimated to be 4 Å for EDA using the relative molecular surface area values provided by Gmehllng et al. (3), which agrees with the experimental value (4). The surface mole fraction (m.f. s ) can be obtained by: Γ m. f. s = (5.3) Γ + Γ Figure S5d shows the surface mole fraction of EDA versus the bulk mole fraction of EDA. S
Surface Tension (mn/m) 8 7 6 5 4 3.5 a b y=-9.964x+7.8 R =.9659. f.5..5 Surface Mole Fraction of EDA Γ, (µmol/m ). c 8 6 4 d.8.6.4.....4.6.8. Bulk Mole Fraction of EDA Figure S5. (a) Surface tension values versus bulk mole fraction of ethylenediamine. Data points: experimental data with error bars; dashed line: linear regression. (b) Activity coefficient (f) of ethylenediamine in aqueous solutions. (c) Relative adsorption ( Γ, ) of ethylenediamine in aqueous solutions. (d) Surface mole fraction of EDA versus bulk mole fraction of EDA. S
SFG Spectra of Reference Systems (Polarization Combination ssp) SFG Intensity (a.u.) 4 3 a.5 m.f. EDA SFG Intensity(a. u.) SFG Intensity (a.u.) 6 4 3 b.7 m.f. EDA at ph 9.7 (Singly Protonated EDA) c.5 m.f. EDD (Doubly Protonated EDA) 8 3 3 34 36 38 Incident Infrared (cm-) Figure S6. SFG spectra of three reference systems: a).5 m.f. ethylenediamine at the air/liquid interface; b).7 m.f. singly protonated ethylenediamine at the air/liquid interface; c).5 m.f. ethylenediamine dichloride at the air/liquid interface. S
References:. Liu, D.; Ma, G.; Xu, M.; Allen, H. C. Adsorption of Ethylene Glycol Vapor on α-al O 3 () and Amorphous SiO Surfaces: Observation of Molecular Orientation and Surface Hydroxyl Groups as Sorption Sites. Environ. Sci. Technol., 5, 39, 6.. Stumm, W.; Morgan, J. J. Aquatic Chemistry: Chemical Equilibria and Rates in Natural Waters, 3rd ed.; John Wiley & Sons: New York, 996. 3. Wang, C.; Groenzin, H.; Shultz, M. J. Surface Characterization of Nanoscale TiO Film by Sum Frequency Generation Using Methanol as a Molecular Probe. J. Phys. Chem. B, 4, 8, 65. 4. Hirose, C.; Akamatsu, N.; Domen, K. Formulas for the Analysis of Surface Sum-Frequency Generation Spectrum by CH Stretching Modes of Methyl and Methylene Groups. J. Chem. Phys., 99, 96, 997. 5. Brown, M. G.; Raymond, E. A.; Allen, H. C.; Scatena, L. F.; Richmond, G. L. The Analysis of Interference Effects in the Sum Frequency Spectra of Water Interfaces. J. Phys. Chem. A,, 4,. 6. Lu, R.; Gan, W.; Wu, B.-H.; Chen, H.; Wang, H.-F. Vibrational Polarization Spectroscopy of CH Stretching Modes of the Methylene Group at the Vapor/Liquid Interfaces with Sum Frequency Generation. J. Phy. Chem. B, 4, 8, 797. 7. Wang, J.; Chen, C.; Buck, S. M.; Chen, Z. Molecular Chemical Structure on Poly(methyl methacrylate) (PMMA) Surface Studies by Sum Frequency Generation (SFG) Vibrational Spectroscopy. J. Phys. Chem. B,, 5, 8. 8. Wei, X.; Hong, S.; Zhuang, X.; Goto, T.; Shen, Y. R. Nonlinear Optical Studies of Liquid Crystal Alignment on a Rubbed Poly(vinyl alcohol) Surface. Physical Review E: Statistical Physics, Plasmas, Fluids, and Related Interdisciplinary Topics,, 6, 56. 9. Jonsson, B.; B., L.; Holmberg, K.; Kronberg, B. Surfactants and Polymers in Aqueous Solution: John Wiley & Sons Ltd 998.. Tyrode, E.; Johnson, C. M.; Baldelli, S.; Leygraf, C.; Rutland, M. W. A Vibrational Sum Frequency Spectroscopy Study of the Liquid-Gas Interface of Acetic Acid-Water Mixtures:. Orientation Analysis. J. Phy. Chem. B, 5, 9, 39.. Defay, R.; Prigogine, I.; Bellemans, A. Surface Tension and Adsorption. New York, NY: John Wiley and Sons, Inc. 966.. Smith, J. M.; Van Ness, H. C. Introduction to Chemical Engineering. New York, NY: McGraw- Hill 987. 3. Gmehllng, J.; Rasmussen, P.; Fredenslund, A. Vapor-Liquid Equlibria by UNIFAC Group Contribution. Revision and Extension.. Ind. Eng. Chem. Process Des. Dev., 98,, 8. 4. Gaines Jr, G. L. Langmuir-Blodgett Films of Long-Chain Amines. Nature, 98, 98, 544. S3