9952
J . Phys. Chem. 1992, 96, 9952-9959
Orientatlons of Chain Axes and Transition Moments in Langmuir-Blodgett Monolayers Determined by Polarlzed FTIR-ATR Spectroscopy Dong June A h and Elias I. Frames* School of Chemical Engineering, Purdue University, West hfayette, Indiana 47907 (Received: June 25, 1992; In Final Form: September 3, 1992)
FTIR-ATR dichroic ratios with and without a rotation of the ATR plates were measured for calcium or cadmium or lead stearate LB monomolecular films on germanium, hydrophilic silicon, and hydrophobic silicon plates. Various models for uniaxialor biaxial orientations of chain axes and transition moments are presented and critically evaluated. A biaxial orientation distribution was observed for calcium stearate. The average tilt and azimuthal angles of the chain axes were found to be 36O and 42O, respectively. Those angles of the transition moments of the CHI group were 6 6 O and 4 6 O , respectively. Cadmium and lead stearate monolayers had nearly uniaxial chain and moment distributions. Molecular orientationsvaried substantially with the ion type bound to the polar groups but little with the substrate surface. The average tilt angle of the chain axes was ca. 11O higher in X-type monolayers (methyl groups on the substrate surface) than in Z-type monolayers (polar groups on the surface) of lead stearate. The dichroic ratio for the cadmium stearate monolayer decreased from 0.98 at 25 OC to 0.84 at 130 OC. Recrystallization of the monolayer upon cooling was observed to be reversible and showed no time lag.
it is not generally valid, as will be discussed later. A different uniaxial model focusing on the chain axis was applied to lipid and Organized molecular films such as Langmuir-Blodgett (LB) peptide cast films43and LB Biaxial models were also multilayers and self-assembled monolayers have been investigated adopted for thin polymeric film^.^*^' Recently, the most detailed extensively over the past decade. Recently, electronic and photonic orientation description for an organic single crystal was used for properties of these thin organic films have received much attenrigid-rod oligoimide LB films.48 In the previous studies, no tion. lv2 When molecules are highly oriented, their macroscopic systemtic criteria for discriminating between p i b l e orientation properties can be quite different from those of unoriented bulk distributions were used. materials. New criteria based on rigorous lTIR-ATRc-s theory Fourier transform infrared (FTIR) spectroscopy is an excellent are proposed in this study, to help choose proper orientation mod& tool for studying the orientation of thin organic films at the for various LB monolayers prepared on semiconductor substratea. molecular level, since it is nondestructive and highly sensitive. For Orientations of both transition moments and chain axes are demolecular orientation determination, various modes of IR have been used, such as reflectionabsorption (RA),= transmissi0n,4>~*'~ termined independently and are used for examining several p r e vious models and for obtaining novel orientation information. The attenuated total reflection (ATR)," with either unpolarized or relations proposed in the previous studies are tested both theopolarized infrared beams. Frequently, combinations of results retically and experimentally. The effects of the substrate surface, of several modes have been used. In many previous quantitative the metal ion type, and the temperature on molecular orientations analyses, absorbance intensities or integrated absorbances have in LB monolayers of salts of stearic acid are reported. been interpreted by uniaxial moment orientation models, in which it has been assumed that transition moments are uniaxially disTheory O f Orie~~tPtion E v d ~ t i oby~Pohrized FTIR-ATR tributed around the direction normal to the film Spectr~OPY The combination of RA and transmission modes of IR specGeneral "y.The orientations of transition dipole moments trosCopy'4-16s22-26 has been frequently employed, because these two and molecular chain axes are determined by" (1) transforming modes are effective in differentiating transition moments lying the vectorial transition moment (M) from a molecular (chain) parallel (or close to parallel) to the f h surface from those normal coordinate system into a laboratory frame and (2) calculating the (or close to normal) to that surface. It should be noted, however, projection (or coupling) of the moment M on (or with)the electric that sample films must be prepared on a specular metallic surface field (E)of the radiation at the surface of an IRE (here, we use for the IR-RA mode and on another nonabsorbing substrate for the 45O-cut trapezoid ATR plate). the IR transmission mode. previouS studies have ignored the effect The molecular crystallographic coordinates (a, b, and c axes) of the substrate surface on the molecular orientation or assumed and the laboratory Cartesian coordinates ( x , y , and z axes) are that it is the same. Furthermore, for monomolecular films the shown in Figure 1. The z axis is along the normal to the film signal-to-noise (S/N) ratios of the RA and transmission modes or to the ATR surface; the x axis is along the propgation direction are often too low to allow precise determination of Orientation. of the IR beam; they axis is along the direction of the s polarFTIR-ATR has much higher S/N ratios than the above ization. The p polarization lies in the 0 x 2 plane (the plane-ofmodes.27 Fundamental theories on IR spectroscopy through internal reflection elements (IRE) were developed by Harri~k,2'-~~ incidence), at 45O with each of the two axes. In this description, one presumes that a chain axis which can represent molecular Zbinden?l and others.3242 The dichroic ratio, which is the key orientations can be defined. An example is an all-trans chain in parameter used in determining the molecular orientations, is a crystalline film. In highly disordered systems, or gauche condefined as the ratio of the absorbance intensity (or integrated formations, defining a chain axis is not useful, and then one absorbance) of s-polarizedlight (normal to the planeof-incidence) to that of ppolarized light (parallel to the plane-~f-incidence).~**~~ describes orientations of transition moments directly. The polar angles are defined as follows: y is the tilt angle of the molecular Ulman et al.I2J3have applied a uniaxial moment model to selfchain axis (c axis) from the surface normal (z axis); 6 is the assembled monolayers in order to find average tilt angles of azimuthal angle, the one between the c' axis and the x axis; 8 transition moments (#) from the surface normal. Then, using is the angle of the transition moment M with the c axis; and 0 the transition moments of the CHI group stretching vibrations, is the third Eulerian angle between the Ozc'plane and M, which which are perpendicular to the molecular chain axis (for trans is the projection of M onto the plane (the circle in Figure lb) chains), they obtained an average tilt angle of the chain axes (yo) normal to the c axis, as depicted in Figure IC. The transformation by the following relation: yo = 90° - qy. This relation is, equations of the molecular coordinates of M into the laboratory however, strictly valid only for identically onented molecules, and Introduction
0022-3654/92/2096-9952$03.00/0
63 1992 American Chemical Society
LB Monolayers
The Journal of Physical Chemistry, Vol. 96, No. 24, 1992 9953
TABLE I: Normalization Factor (C) for Orientation Distribption Function in Eq 2
random uniaxial (chain axis) biaxial (chain axis) single crystal"
N0/(49) NO/@* sin 70)
sin y S(Y
- 70) sin 7
N4 - 40) S(8
S(r - 70)
sin Y
NO/m sin 'yo)
- 8,) 6(4 - t$o) S(y - yo) sin y
No/sin yo
"Unlike inorganic crystals, organic single crystals may allow defining a unique chain axis. z
surrounded by an ambient gas with n3 = 1, can be found elsewhere.35 The dichroic ratio D is then, in general
Here, contributions from absorbance through the entrance and exit faces of the ATR plate are ignored, since their effects are negligible or small.' '*12 If molecules are randomly oriented in the fh,the dichroic ratio is
0
DrandomI = E ; / ( E 2
,9 00°00 ,0/
+ E:)
(5)
with the distribution functions f i ( 8 ) = h(4) = h(y)= 1. The values of I are 0.826, 0.874, or 0.912 for Ge ATR plane (nl = 4.03) with n2 = 1.4, 1.5, or 1.6, respectively. For Si ATR plate (nl = 3.42), I = 0.840, 0.891, or 0.932, respectively. Uniaxial Orientation Models. When all molecular chain axes are uniaxially oriented around the surface normal (zaxis), that is, they have a preferred tilt angle yoand no preferred angles in 8 or 4, then the distribution functions are such that f3(y) = S(y - yo)andfi(8) =&(4) = 1. Then, the absorbances are derived as follows: A, = yfl,,.@E;[(2
- 3 sin2 e) sin2 yo+ 2 sin2 e]
(6)
- 3 sin2 e) x ( E 2 sin2 yo+ 2E,2cos2 yo)+ 2(E,2 + E:) sin2 e] (7) For the transition moment parallel to the c axis (e = the
A, = yfl@[(2 (c) Figure 1. Geometry of ATR experiments according to laboratory (x, y, and z) and molecular (a, b, and c) coordinates: (a) infrared beam direction and polarization; (b) definition of polar angles y, t$, and 0; (c) depiction of 8.
coordinates can be found in the literat~re.~'Generally, Mi = Mi(O,&,y,e), where i = x, y , or z. We consider a specific transition moment of which the angle 8 is fixed and known or can be estimated; e.g., 8 = 90° for the moments of the symmetric and asymmetric stretching vibrations of the CH2 groups with the fatty acid chain axis. Then, components of M in the laboratory frame are described as Mi = Mi(B,t$,y,B=fmed). The domain being considered is 0 I 8 I 2?r, 0 I 4 I 2?r,and 0 I y I u/2. The population No of the transition moments with an orientation distribution N(8,4,y) is
If the orientation angles 8,4, and y are uncorrelated, then the distribution function is separable: N(b#bT) = Cfi(8)h(4)M Y ) =
h(4)f3h)sin 7 (2)
wnere L is a normalization racror. cxpressions or L ror vanous
orientation distributions are listed in Table I. The absorbance (Ai)is a measure of the transition moment M projected on the
OO),
dichroic ratio can be reduced to
For perpendicular transition moment
(e = 90°)
E;(2 - sin2 yo) ~ ( ~ ~ ; e = 9= 00) E,2(2 - sin2 yo) + 2E: sin2 yo
If one considers the uniaxial distribution of transition moments without defining the chain axis, then the resulting expression of the dichroic ratio is identical to eq 8, with yobeing replaced by 70M.49 Since the absorbance reflects an angular distribution of chain axes (or transition moments) about the preferred tilt angle yo (or yoM),the dichroic ratio determined from experiments is actually an average tilt angle T~of the chain axes (or # of the transition moments). Figure 2 shows how the dichroic ratio for the uniaxial distribution of chain axes varies with yofor various values of 8. When yois at the magic angle of 54.7O, as indicated by point I, the dichroic ratio is the same as that of the totally randomly oriented sample. Point A is the maximum possible dichroic ratio for 8 = Oo and 90': A = E;/E2 or 1.070 for Ge. B represents the minimum for 8 = 90°: B = Ey2/(Ex2 2E:) or 0.737. For €3 = 54.7O, which also is the magic angle of 8,the dichroic ratio is independent of yo and constant: D = I. This means that transition moments having 8 54.7O are not good probes for determining the molecular chain orientation. Using transition moments having 8 = Oo is probably the best choice for estimating chain orientation of uniaxial systems. However, with
+
-
1 ns; c;nprtssiunsui
s;ieC;rIc; iieiu
ampuruue ~un~riuns I J I ~iur a rnin
film with a refractive index n2 on an ATR plate with index nl,
(9)
9954 The Journal of Physical Chemistry, Vol. 96, No. 24, 1992 A+
Ahn and Franses 1.21
*A
I
I
,
I
I
I
I
I
1
1.o 1 .o
.-e
a"
0.8
.g
+B 0.6
Q
z .o 0.4 0
r
.-
.-U
2
e
0.8 0.6 0.4
o.2 0.0
0
20
10
30 4 0
50
60
Chain Axis Tilt Angle
70
80 9 0
20
40
50
60
(10)
They obtained yoof cadmium stearate LB films by using 77's of the symmetric and asymmetric CH2 stretching vibrations (9 CH2 and Va CH2). The above relation is tested here with 70,being determined independently of TOM'S, and has been extended to biaxial systems (see below). The effects of the refractive indexes of the ATR plate (nl) and the sample fdm (n2) on the dichroic ratio are depicted in Figure 3. The dichroic ratio is sensitive to n2. For 8 = O", change of nz by 0.1 leads to a change in yo of up to 5". By contrast, for 8 = 90°, this discrepancy can be much larger, 5" for yo = 40" and 10" to 20" for yo> 70". Hence, the use of an accurate value for n2 is important in precise determination of molecular orientation. Although less si&icant than the above effect, the effect of the substrate refractive index is larger for systems where the transition moments are parallel to the substrate surface (yo = 0" for 8 = 90" and yo = 90° for 8 = OO). In the present study, the contributions of surface oxides to nl of the ATR plate have not been considered. Biaxial Chientatioa Models. In fatty acid LB films, straight chain molecules have been reported to be in a hexagonal subcell packing of all-trans chain axes which are freely rotated so that the distribution of the CH2groups is collectively isotropic around the chain axc8.2zJ7 Hence, the biaxial model is considered to be sufficient to describe the molecular orientation in these typical LB films. One needs to establish a systematic way to discriminate the orientation distributions (uniaxial or biaxial), so that one selects an appropriate model to describe the data.
70
80 9 0
(")
Figwe 3. Dichroic ratios for the uniaxial dietribution with n2 = 1.4, 1.5, and 1.6. Solid lines are for Ge (nl = 4.03); and dotted lines are for Si (nl = 3.42). 0
5 4
k
+ cos2 yo = 1
30
Chain Axis Tilt Angle
typical LB films, where chain axes are aligned close to or along the surface normal and perpendicular transition moments lie toward the plane of surface, it is more commonto employ moments having 8 = 90" (e.g., CH2 groups), which are easier to resolve and have strong absorptivities in FI'IR-ATR spectroscopy." In our approach, we calculate either 70 or # from independent equations, which account for uniaxial distributions of either chain ax= or moments. By contrast, Ulman et al.12'3and Song et al." have used equations which estimate 77. Then, they approximated yo from simple geometry by assuming that yo = 18 - 771, a p parently because yo = 18 - yoM( for individual molecules. Nonetheless, the two distributions, one in which the chain axes and the other in which the moments have an average angle to have an average an le # from the z axis, are not the same. Hence, to# 18 - yo 1, in general, except for the case that 8 = Oo, for which yo = 77, and exce t for certain monodisperse systems, for which yo = yo and y#= 77. For the same yo, D can be very different in the two approaches, as depicted in Figure 2. Talcenaka et al.16*36 proposed an alternative method for calculating the chain axis tilt angle by the use of the tilt angles of two transition moments. For an orthogonal system of the chain and two moments (MI and M2)'O cos2 7 3
10
(")
Figure 2. Dichroic ratios for the uniaxial distribution with various 8 values (solid lines). The dichroic ratio according to Ulman's formulat i ~ n ' z(dotted ~ ~ line) is compared to that for 8 = 90° (thick solid line). nl = 4.03 (Ge), n2 = 1.5, and n3 = 1.0 were used.
cos2 # I +
i 0
4
3
3
2
2
1
1
90
0 135
Chain Axis Azimuthal Angle (")
__
0
Chain Axis Tilt Angle
(")
180 .
4. Dichroic ratios for the biaxial distribution with 8 = Oo. Dichroic ratios larger than 5 were "cut off". Values of refractive indexes are the same as in Figure 2.
When the molecular chain axes are biaxially oriented, 80 that they have preferred angles yoand &, and no preferred angle 8, the distribution functions are thath(y) = 6(y - yo),fi(t$) = 8(t$ - do),andA(8) = 1. The absorbances are, from eq 3 A, = f/2N,@Ey2[(2 - 3 sin2 8 ) sin2 I$o sin2 yo
+ sinZ81
A, = &N,@[(2 - 3 sin2 8 ) X (E? cos2 t$o sinZyo + E: cosz yo) + (E?
+ E?)
(11)
sinZ81 (12) The dichroic ratios for 8 = 0" and 90" in biaxial systems areS1 E; sin2 t$o sin2 yo
D(do,ro;e=o") =
E 2 m2I$o sin2 yo + E:
~ ( t $ ~ , y ~ ; e = 9=0 0 )
cos2 yo
(13)
E:( 1 - sinZt$o sin2yo)
E2(1 - cos2 t$o sin2 yo)
+ E?
sin2 yo (14) As in the case of the uniaxial orientation model, eq 13 is identical to an expression for the biaxial transition moment distribution, with the chain axis not explicitly considered. The above two functions are plotted in Figures 4 and 5 against the tilt angle yo and the azimuthal angle t$o of the chain axis. These functions have a period of r both in angles yoand do. The dichroic ratio changes rapidly as yo 90" and t$o 90°, for 8 = 0". and as yo 90° and t # ~ ~ O", for 8 = 90". D(8=Oo) is steeper than D(8=9Oo) and it goes to infinity when yo = t$o = 90". If yoand &, are distributed, then the average tilt angle yo and the azimuthal angle a. of the chain axes are used in the above
-
--
-
le Journal of Physical Chemistry, Vol. 96, No. 24, 1992 9955
LB Monolayers
sin2 yo (16) If transition moments or chains are distributed, these equations and $o instead of yo and With two experare used with to iments without and with 90' rotation, two dichroic ratios per transition moment can be used to determine fo and io from eqs 14 and 16, independently of 70"and $0". Another way of determining f o and Z0 is possible by using # and $2. By analogy to cq 10, the following relation is derived for an orthogonal system consisting of MI, M2, and the c axis: sin2 -#I cos2 @ I + sin2 7 3 cos2 $3+ sin2 40 cos2 $0 = 1 (17) where sin2 f o = cos2 771 ea2fp from eq 10. Two dichroic ratio6 far M Iyield and qF from eqs 13 and 15; two additional dichroic ratios for M2 provide # and $p. Then, fo and $o are determined from eqs 10 and 17. The above two schemes can be uaed for properly discriminating orientation models. When $0 or AM is calculated at the mugic angle of 4S0, the orientation distribution is considered to be uniaxial or effectively so. The LB monolayers with cadmium, lead, and calcium stearate are evaluated in the present study with these schemes. F d y , if it is possible to rotate the ATR plate by other angles and obtain a larger number of dichroic ratios, one should be able to enhance the accuracy in determining the average orientation angles, or even use more elaborate orientation models.
one organic adsorption, two ion-exchange columns, and a final fdter having a pore size of 0.22 pm. The resistivity of the deionized (DI) water used in the experiments was 18 Mll cm. The stearic acid salt monolayers were prepared by spreading a hexane (HPLC grade from Aldrich) solution (1 mg/ 1 mL) of stearic acid on water containing Pbz+ or Cd2+or Ca" ions. The pH was regulated from 6 to 11 with NaOH or HCI solutions. The ionic concentrations ranged from lo4 to M. Monolayer compositions were controlled by adjusting pH and ionic ~ncentrations?~ ATR plates were Ge or Si (1 X 10 X 50 mm, 45O trapezoids). The Ge ATR plates were cleanedS5by sonication at 58 'C in ethanol, trichloroethylene, ethanol, and DI water respectively for 15 min. The treatment was done twice and was followed by drying in an Ar jet. For the hydrophilic Si ATR plates, additional 15-min sonications in NH40H:H202:H20 (1: 1:s by volume) and HCI:H202:H20 (1:l:S) were done between the two cleaning sequences, with the second sequence for a shorter time (each step for 5 min). For producing a hydrophobic Si surface the plate was dipped into 1% aqueous H F solution for several minutes, then rinsed with bulk DI water, and used within 5 min. The monolayers were transferred to the substrates at 22 'C in a clean room by a standard LB depasition "dat 8 mm/min, with the Joyce Loebl IV Langmiur trough. The monolayer was maintained at a constant surface pressure Il = 19 1 mN/m for at least 30 min before and during transfer. The large face of the ATR plate (Oxy plane in Figure 1) was perpendicular to the direction of the barrier annpmsion. The p l a b were normally covered by 45 mm along the x axis (the long axis of the ATR plate). When the plate was rotated by 90°, the dipping direction was along they axis, and then the plate was almost completely covered by the monolayer. For Ztype monomolecular f h (polar groups attached to the substrate surface), the substrates were dip@ before compression of the monolayer (n = 0 mN/m). For X-type monolayers (methyl groups to the substrate), the substrates were dipped after the equilibrium surface pressure was reached. To ensure that no deposition occurs upon withdrawal of the plate out of water, the remaining monolayer was first removed from the water surface. Then, the plate was withdrawn at a rate of 50 mm/min. Two or more samples were prepared for each LB monolayer. The transfer ratio was found to be 1.0 f 0.05. A cast film of stearic acid (StH) was made on a Ge ATR plate by dropping 20 pL of 1 mg/mL of hexane solution and covering a. 4.5 cm2 on one large face of the plate. The estimated average thickness was about 0.05 pm. The infrared spectra were obtained at 25 OC with a resolution of 2 cm-I and 500 scans by using a Nicolet 800 FTIR spectrophotometer equipped with an MCT detector. A bench accessory from Connecticut Instruments Inc. was used for the ATR measurements. 5 1 reflections of the IR beam were allowed for the ATR plates used. Each background spectrum was taken right before the monolayer deposition, with the same configuration of the ATR plate to that for each sample spectrum. The monolayer was heated with a Connecticut Instruments Inc. heat source unit. The temperature was increased rapidly between annealing steps and was maintained constant within f 1 'C at least for 20 min, before the spectra were taken. The sample was allowed to recool in the IR chamber. The advancing contact angle was measured for water (pH 5.5) with the Ram€-Hart goniometer. At least five independent measurements were taken for each surface. The rasults were reproduced with several surfaces. Measured angles were the maximum contact angles of water droplets as their volumes were increased by small amounts, supplied by a square-cut syringe needle, without a substantial change in the contact areas at the interface of dense materials, as previously
Experimental Section Stearic acid (>99%) and CaC12.2H20 (>99%) were purchased from Flulra, PbCI2 (>99.999%) was from Johnson Matthey Electronics, and CdC12 (>99.999%) was from Aldrich. All were used without purification. Distilled water was purified by Millipore-Q water system consisting of one reverse-osmosis module,
Results and Diacuaalon Polarized IR spectra for a cast f h of stearic acid on a Ge ATR plate are shown in Figure 6. Band assignmenb arc given in Table 11. The spectra show that stearic acid is in its normal polycrystalline s t r u c t ~ r e . ~ From ~ * ~ ' the absorbance intensities for the two CHI group stretching vibrations, the dichroic ratio is found
f
5
4
4
3
3
2
2
1
1
90
0 0
Chain Axis Tilt Angle
Chain Axis Azimuthal Angle (")
Figure 5. Dichroic ratios for the biaxial distribution with 8 = 90'. Values of refractive indexes are the same as in Figure 2.
two equations. From a single dichroic ratio, one can determine one angle if the other is known. To determine both angles, one needs to use at least two moments with Werent values of 8. Since v' of CH2 and VS of CH2 with the same 8 of 90' are the most prominent bands available for calculating dichroic ratioe in typical LB films, a better method, if feasible, should involve two experiments, with one after rotating the ATR plate around the z axis.52*53In practice, if the rotation is not possible in a given instrumental setup, one can deposit the LB film with the ATR plate rotated, as has been done here and elsewhere.44 With the ATR plate rotated by -90°, new polar angles are such that 40' =,,$t - 9oo and y( = yo. The resulting expressions for the dichroic ratios are E: COS' 40sin2 70
w(40,yo;e=oo) =
E : sin2 4osin2 yo + E? ea2yo
~ ' ( 4 ~ , ~ ~ ; e = 9=0 0 ) E(:1
frl
(15)
E,Z(1 - cos2 4&n2 yo)
- sin2 40 sin2 yo) + E?
+
*
9956 The Journal of Physical Chemistry, Vol. 96, No. 24, 1992 0
.
5
1
Ahn and Fransee
m
0.15
0.15
0.10
0.10
0 05
0.05
0.2
0 0,
m
ein 9 0
3000
2900
2800
-
1800 1400 1200 1000
Wavenumber (cm") Figure 6. IR spectra of stearic acid cast films on Ge; ppol (solid line) and s-pol (dotted line). Band assignments are given in Table 11.
TABLE Ik Ri.ci).l LB Bands in Steuic Acid ( S a ) Cast and SteurteMonomokcdmli" peak position (hO.5 cm-') cadmium lead calcium transition StH salt salt cast film salt momentsa V. CHI 2954 2957 2958 2957 V. CH; 2917 2917 2917 2917 V' CH2 2850 2850 2850 2849 Y c-0 1702 V. C o o 1535 1540-1 517' 1573-15406 6 CH2 1472, 1464' 1467 1467 1467 V' C o o 1409 1399 1406 (T + 0 ) CHI 1360-1180 1360-1180 1360-1180 1360-1180 b OH 942 asymmetric stretching; v' symmetric stretching; 6 scissoring; T twisting; o wagging; b out-of-plane bending. *Broad bands, or doublets, not explicitly resolved. Chublet.
to be 0.88. If we assume that the molecules are randomly oriented in the cast film, we calculate n2 = 1.52 by using eq 5 with nl = 4.03 and n3 = 1.0. Since this value is close to 1.5, which is considered typical for many organic materials?* we will use the equations in the theory section with n2 = 1.5 for further calculations of orientation angles of stearic acid or stearate salt monolayers. Figure 7 shows the spectra in the CH stretching region obtained for Z-type monolayers of cadmium, lead, and calcium stearate on Ge ATR plates. The "rotation-of-ATR plate" method was used to obtain additional data for model discrimination, as detailed in the theory section. Because the surface area covered by LB monolayers was smaller for the nonrotated ATR plates than for the rotated ones, the absorbance intensities were lower in the former cases. It was also observed that absorbance intensities increased in the order of calcium, cadmium, and lead stearates. This trend is due to differences in surface densities or in molecular orientations (see eq 3). The orientation results are summarized in Table 111. The dichroic ratios for cadmium stearate and lead stearate were nearly the same for the two configurations of the ATR plate. By contrast, those for calcium stearate were different
0.00 3000
0.00
2900
2800 3000
2900
2800
Wavenumber (em.')
Figure 7. IR spectra in the CH stretching region for stearate LB monolayers on Ge: (a) without 90° rotation; (b) with 90° rotation of the ATR plate: (1) Z-cadmium stearate; (2) Z-lead stearate; (3) Z-calcium stearate; ppol (solid line) and s-pol (dotted line).
by ca.0.1, which impliea that the molecular orientations of calcium stearate are not uniformly distributed around the z axis. By the use of eqs 14 and 16, toand bowere determined, independently of 70"and $0". For St2Cd and St2Pb, 40'sare found to be close to 45O, the mu@ angle, within 0.7'. Those results indicate that the moleculea in St2Cd and St2Pbmonolayers are in the uniaxial (or nearly uniaxial) distribution, or that the uniaxial model is sufficiently valid. By contrast, 40is 41.6' for St2Ca monolayer, and the molecular chain axes are aligned on the average slightly along the x axis (the withdrawal direction of the plate). This is, to the authors' knowledge, the first quantitative report on the existence of the biaxial distribution in LB monomolecular systems of straight-chain surfactants.22 The average tilt angle T ~ de, termined by using the biaxial model, is ca. 23' for St2Pb; those for St2Cdand St2Ca are about 36'. For St2Ca, this value is close to what was previously observed on a Si surface by NEXAFS?9.60 The alternative method (eq 17) of using #'s and #'s of the CH2 stretching vibrations was also employed to calculate chain axis orientations. Consistent results are obtained for both methods, as shown in Table 111. It is noted that for detecting biaxial orientation distributions in very thin films,the ATR experiments with rotation-of-plate are more sensitive than the transmission experiments, which often have too low S/N ratios. Moreover, by employing the biaxial model, we use a more rigorous criterion for testing the validity of the uniaxial model for LB monolayers of cadmium and lead stearate. Various LB monolayer systems with St2Cdand St2Pbhave been prepared in order to inveatigate the effect of monolayer-substrate interactions on the molecular orientations. The spectra given in Figure 8 show the subtle differences of the substrate and the ion type on absorbance intensities, peak positions of the carboxylate group (See Table 11). and peak line shapes. As expected from a theoretical consideration, Si substrates with a lower refractive index (nl = 3.42) gave higher absorbances as compared to Ge (n,
TABLE IIk Estimdon of Avemge Cun Orientation from Rot.tiOa-of-ATR Plate Expdmenta Urhg Biaxial Modelsa monolayer Z-StZCd
ATR plate Ge
Z-St2Pb
Ge
Z-St,Ca
Ge
band V. CH2 V' CH2 $CH2 V' CH2 V. CH2 vl CH2
D without 90" rotation 0.99 0.98 1.04 1.02 1.03 1.02
D with 90° rotation 0.98 0.98 1.04 1.03 0.93 0.94
eqs 14 and 16 70(deg) (deg) 34.7 44.6 35.7 45.0 20.6 45.0 25.1 45.7 36.0 41.3 35.9 42.0
eqs 13 and 15 (deg) 66.3 45.1 65.6 45.0 75.6 45.0 72.5 44.9 65.5 45.8 65.5 45.6
# (deg)
ib (de81 (eq 17) 44.8 45.5 41.6
'Dichroic ratios have been calculated with absorbance intensities. Standard deviations of D are less than 0.006. In biaxial models, D is very sensitive to angles (see Figures 4 and 5), and their values are sometimes significant up to three digits.
The Journal of Physical Chemistry, Vol. 96, No. 24, 1992 9957
LB Monolayers
Z-StzCd
Ge
Z-StZCd
Si/Si02
Z-St2Pb
Ge
CH; v' CH2 V' CHI v' CH2 V' CH2 Z
2-StZPb
Si/Si02
X-StiPb
Si/HF-treated
.oo .oo
v' CHI vl CH2 v' CH2 vl CH2 v' CH2
55
54 36 36 31 31 20 24 26 29 31 39
0.98 0.98 1 1 1.04 1.03 1.05 1.04 1 0.99
.oo
109 102 102 102 102 96 98 98 99 102 103
66 66 65 65 16 14 12 60 65 64
36 31 22 21 38
'Standard deviations of dichroic ratios are less than 0.008. 1 .oo
2925
TABLE V Results of Contact Angle Measwewnts
surface
advancing contact angle with water (PH 5.5) (del?)
teflon Si/Si02 Si/HF-treated Z-St2Pbmonolayer on Si/Si02 X-St,Pb monolayer on Si/HF treated
105 t 2 53 19 & 2 101 3 63 t 1
2923
0.95
2921 0.90
-
-
2919
i
2917 1.00
2855
f
0
B
0.2
0.08
0.90 2849
8 5
e
4
0.06 0.1 0.04
0.95
2853 2851
2847
1
2845 I 10°C
v'CH, I
50%
I
I
90%
130°C
I
1 0.80 25OC
Annealing Step 0.02
0.0
0.00 3000
2800
1800
1600
1400
1200
1000
Wavenumber (cm") Figure 8. IR spectra (s-pol) for stearate LB monolayers: (1) Z-cadmium stearate on Ge; (2) Z-cadmium stearate on hydrophilic Si; (3) 2-lead stearate on Ge with the ATR plate rotated by 90'; (4) X-lead stearate on hydrophobic Si; ( 5 ) Z-calcium stearate on Ge with the ATR plate rotated.
= 4.03). Dichroic ratios and orientations are shown in Table IV. Table V summarizes the results of the contact angle measurements, which confirm the type of the LB monolayer (Z or X). For Z-type StzCd monolayers, the polar group interacts similarly with Ge and Si (hydrophilic) substrates. yo is found to be about 36",6l by using the uniaxial models, one being eq 9 and the other being eq 8 combined with eq 10. When the polar groups are associated with lead instead of cadmium, there probably is a minor effect: 22" on the Ge surface and 27" on the hydrophilic Si surface. By contrast, when methyl groups are attached on the hydrophobic Si surface, as in X-type monolayers, the monolayersubstrate interaction is weaker and 70 was about 38". The alignment of molecular chain axes in the X-type monolayers could be disturbed during the withdrawal of the ATR plate out of water, because of a stronger interaction of water with polar groups. When one uses eq 9 for determining to from the dichroic ratios of the CH2 stretching vibrations (0 = 90°), a dichroic ratio change of 0.01 leads to an 11" change in T~ for yo 0" or 90° and to a change of less than 2" for qo 45". By contrast, if one uses eq 8 for estimating #, the sensitivities are 4" and 0 9 , respectively. Despite the sensitivity difference between eqs 8 and 9, both yield consistent results for yo, as summarized in Table
-
Figure 9. Shift in peak positions of symmetric and asymmetric CHI
stretching vibrations and their dichroic ratios with temperature. IV. The sums of yo, by eq 9, and yr, by eq 8, range from 96" (Z-StzPb) to 109' (cast film of StH), clearly different from 90". The deviation of 70 7: from 90" becomes larger when the chain axes are tilted closer to the substrate surface, as expected from Figure 2. Hence, it is more appropriate to determine yo and y y independently and not assume that their sum is 90°.12J3,20*21 The order+rder transition in LB films have been previously investigated by using IR-RA6z4 and Ramads spectroscopy. In the present study, we have examined the effect of the orderdisorder transition on the dichroic ratios via polarized FTIR-ATR spectroscopy. The Z-type St2Cdmonolayer on Ge has been heated from 25 to 130 "C, which is higher than the melting point (1 10 "C) reported for 7 LB layers of cadmium arachidate on silver.62 As shown in Figure 9, peak positions of both CH2 stretching vibrations increase as the temperature increases. This trend is consistent with previous report~.6~8~~.~' The shift to higher wavenumbers indicates that the monolayer becomes more liquidlike. The dichroic ratio has been observed to decrease from 0.98 at 25 "C to 0.84 at 130 OC. According to the report by senak et al.,6' the proportion of gauche conformations increases with temperature. Hence, defining yo is not useful, since trans chains are not predominant. At 130 'C,the dichroic ratio corresponds to # = 52", when one uses the uniaxial orientation model (eq 8) for transition moments. If we assume that StzCd molecules are randomly oriented at this temperature, then n2 = 1.43 is obtained from eq 5. The use of thisvalue instead of n2 = 1.5 gives ca. 3" lower and 2" higher tilt angles of chain axes and transition moments, respectively, for the systems studied. It was reported65 that the refractive index was increased to 1.55 by the incapration of cadmium ions into arachidate films.68 The hypothesis of the random orientation yields, however, a lower value of 1.43.
+
9958
The Journal of Physical Chemistry, Vol. 96, No. 24, 1992
Therefore, we can conclude that S t Q molecules are not randomly oriented at 130 OC and their average tilt angle # has increased because of the conformational change of the chains. Moreover, the peak positions and the dichroic ratios return to their original values, upon cooling to 25 OC. Such reversibility was not reported for 7-layer films of cadmium arachidate on silver.62 Ulman2 argued that monolayers are apparently in a supercooled state, because of kinetic rather than thermodynamic reasons. Hence, we infer that recrystallization may accur more rapidly than in multilayers. This rapid recrystallization of the monolayer could be enhanced by a stronger attraction of polar groups to the semiconductor surface than to the silver surface. The present results show that the dichroic ratio is a useful quantity for determining the order4isorder transition in LB films.
Conclllsions More rigorous criteria for FTIR-ATR model discrimination of oriented monolayers have been propused in the present study. Whereas cadmium and lead stearate LB monolayers are found to have uniaxial chain and moment orientation distributions, calcium stearate monolayers on Ge have biaxial orientation distributions. Different counterionsassociated with polar groups can lead to different molecular orientations. However, the substrate effects are minor for the systems studied. When methyl groups are adhered to the substrate surface, the average tilt angle of chain axes is higher by ca. 11O than the angle with the polar groups next to the surface. Unlike multilayers, the recrystallization of the cadmium stearate monolayer is found to occur immediately upon cooling. Acknowledgment. This work was supported by two equipment grants by the National Science Foundation (CTS Nos.8604904 and 9007147) and by a Purdue Research Foundation David Ross fellowship to D.J. Ahn. We appreciate helpful discussions with T. L. Marshbanks. Appendix: Notation absorbance (i = x , y , or z) absorbances for s polarization and p polarization, respectively molecular coordinates normalization factor projection of c axis on Oxy plane dichroic ratio electric field electric field intensities angular distribution functions dichroic ratio for random orientation ID,^"^^^) transition moment projection of M on a plane normal to c-axis magnitude of M in laboratory coordinates total number of a certain transition moment being probed by IR orientation distribution function refractive indicts of substrate, film, and ambient gas, respectively polarization perpendicular to the plane-of-incidence polarization parallel to the plane-of-incidence laboratory coordinates polar angles described in Figure 1 preferred tilt and azimuthal angles of an individual chain axis, respectively average tilt and azimuthal angles of chain axes, respectively preferred tilt and azimuthal angles of an individual transition moment, respectively average tilt and azimuthal angles of transition moments, respectively
Ahn and Franses
a(-)
n
delta function surface pressure R q k y No. Gc,7440-56-4; Si, 7440-21-3; St2Cd, 2223-93-0; StzPb,
1072-35-1; StzCa, 1592-23-0.
Refereaces and Notes (1) Langmuit-Blodgett Films; Roberta, G., Ed.; Plenum Prese: New York, 1990. (2) Ulman, A. An Introduction to Ultrathin Organtc Films: From Lungmuir-Blodgett to Se.!fAssembly; Academic Press: New York, 1991. (3) Greenler, R. G. J. Chem. Phys. 1966,44, 310. (4) H a m n , W. N. J . Opt. Soc. Am. 1968.58, 380. (5) Allara, D. L.; Swalen, J. D. J. Phys. Chem. 1982,86, 2700. (6) Allara, D. L.; Nuw,R. G. hngmuir 1985, 1, 52. (7) Nuao, R. G.; FUBCO, F. A.; Allara, D. L. J. Am. Chem. Soc. 1987,109, 2358. (8) Porter, M.D.; Bright, T. B.; Allara, D. L.; Chidsey, C. E.D. J. Am. Chem. Soc. 1987,109, 3559. (9) Chollet, P.-A.; Messier, J.; Rosilio, C. J. Chem. Phys. 1976,64, 1042. (101 Chollet. P.-A. Thin Solid Films 1978. 52. 343. (1 1) Maoz, R.; Sagiv, J. J. Colloid Interface Sci. 1984, 100, 465. (12) Tillman, N.; Ulman, A.; Schildkraut, J. S.;Penner, T. L. J. Am. Chem. Soc. 1988,110,6136. (13) Ulman, A. ACS Symp. Ser. 1991, 447, 144. (14) Kamata, T.; Umemura, J.; Takenaka, T.; Takehara, K.;Isomura, K.; Taniguchi, H. Thin Solid Films 1989, 178, 427. (15) Umemura, J.; Hishiro, Y.;Kawai, T.; Takenaka, T.; Gotoh, Y.;Fujihira, M. Thin Solid Films 1989, 178, 281. (16) Umemura, J.; Kamata,T.; Kawai, T.; Takenaka, T. J . Phys. Chem. 1990,94,62. (17) Kawai, T.; Umemura, J.; Takenaka, T. Langmuir 1989, 5, 1378. (18) Kawai, T.; Umemura, J.; Takcnaka, T. Langmuir 1990, 6, 672. (19) Cropek, D. M.;Bohn, P. W. J. Phys. Chem. 1990,94,6452. (20) Song, Y. P.; Petty, M.C.; Yanvood, J.; Feast, W.J.; Tsibouklis, J.; Mukherjee, S.Langmuir 1992,8, 257. (21) Song, Y. P.;Yanvood, J.; Tsibouklis, J.; Feast, W. J.; Cresswell, J.; Petty, M.C. Lungmuir 1992,8, 262. (22) Rabolt, J. F.; Burns, F. C.; Schlotter, N. E.;Swalen, J. D. J. Chem. Phys. 1983, 78, 946. (23) Rab,J. P.; Rabolt, J. F.; Brown, C. A.; Swalen, J. D. J. Chem. Phys.
1986,84,4096. (24) Stroeve, P.;Saperstein, D. D.; Rabolt, J. F. J. Chem. Phys. 1990,92, 6958. (25) Geddes, N. J.; Jurich, M.C.; Swalen, J. D.; Twieg, R.; Rabolt, J. F. J . Chem. Phys. 1991, 94, 1603. (26) Su, W.-F. A.; Kurata, T.; Nobutoki, H.; KoezuLa, H. Langmuir 1992, 8, 915. (27) Kimura, F.; Umemura, J.; Takenaka, T. Lungmuir 1986, 2,96. (28) Hamck, N. J. Internal Refection Spectroscopy; Wiley: New York, 1967. (29) Hamck, N. J. J . Opt. Soc. Am. 1%5, 55,851. (30) Harrick, N. J.; du PrC, F. K.Appl. Opt. 1966,5, 1739. (3 1 ) Zbinden, R. Infrared Spectroscopy of High Polymers; Academic Press: New York, 1964. (32) Fraaer, R. D. B. J . Chem. Phys. 1953, 21, 1511. (33) Floumoy, P. A.; Schaffers, W . J. Spectrochim. Acta 1966, 22, 5 . (34) Floumoy, P. A. Spectrochim. Acta 1966, 22, 15. (35) Haller, G. L.; Rice, R. W . J. Phys. Chem. 1970, 74, 4368. (36) Higashiyama, T.; Takenaka, T. J . Phys. Chem. 1974, 78,941. (37) Stupp, S. I.; Carr, S.H. J. Polym. Sci.; Polym. Phys. Ed. 1978,16, 13. (38) Landreth, B. M.;Stupp, S.I. Appl. Spectrosc. 1986, 40, 1032. (39) J a w , B.; Koenig, J. L. J. Mucromol. Sci.-Rcv. Mucromol. Chem. 1979, C17,61. (40) Mirabella, Jr., F. M.J . Polym. Sci.: Polym. Phys. Ed. 1984, 22, 1283. (41) Mirabella, Jr., F. M.J . Polym. Sci.: Polym. Phys. Ed. 1984, 22, 1293. (42) Mirabella Jr., F. M.Appl. Spectrosc. Rev. 1985, 21, 45. (43) Brauner, J. W.; Mendelsohn, R.; Prendergast, F. G. Biochemistry 1987. 26, 8151. (44)Takenaka, T.; Nogami, K.;Gotoh, H.; Gotoh, R. J . Colloid Interface Sci. 1971, 35, 395. (45) Okamura, E.;Umemura, J.; Takenaka, T. Biochim. Biophys. Acta 1985,812, 139. (46) Takcnaka, T.; Harada, K.;Matsumoto, M. J . Colloid Interfocc Sd. 1980, 73, 569. (47) Takcda, F.; Matsumoto, M.; Takenaka, T.; Fujiyaahi, Y. J . Colloid Interface Sci. 1981, 84, 220. (48) Cammorata, V.; Atanasoska, L.; Miller, L. L.; Kolaskie, C. J.; Stallman, B. J. Langmuir 1992, 8, 876. (49) Equations in refs 2 and 13 should be corrected to those in ref 12. (SO) The equation used in ref 10 is wentially the same as eq 10 in thia paper. (51) In ref 46,contributions of electrical field intensitits were not taken into account in the dichroic ratio. (52) Sung, C. S.P. Macromolecules 1981, 14, 591. (53) Hobbs.. J. P.:. Sunn. - C. S.P.: Krishnan. K.: Hill. S.Macromolecules 1!M’3, j6, 193.
J. Phys. Chem. 1992,96,9959-9964 (54) Ahn, D. J.; Franses, E. I. J . Chem. Phys. 1991.95, 8486. (55) Wahlgren, M.;Amebrant, T. J. Colloid Interface Sci. 1990,136,259. (56) Evans, S. D.; Sharma, R.; Ulman, A. hngmuir 1991, 7, 156. (57) Snyder, R. G. J. Mol. Specrrosc. 1960,4,411. (58) Allara, D. L.; Nuuo, R. 0. hngmuir 1985, 1, 45. (59) Outka, D. A.; St8hr, J.; Rabe, J. P.; Swalen, J. D.; Rotermund, H. H.Phys. Rev. Lrtt. 1987,59, 1321. (60) Rabe, J. P.; Swalen. J. D.; Outka, D. A.; Stijhr, J. Thin Solid Films 1988,159,275. (61) In ref 59, the average tilt angle of the chain axes of cadmium arachidate monomolecularf h was reported to be less than 15O as measured by
NEXAFS. The reason for the difference between their result and the present one is unclear at preamt. Possible r e a s o ~may be the following: (i) difference in surface roughness of the solid substratesu s 4 (ii) different experimental
9959
conditions. for example, usc of buffer solution;or (iii) different techniques (NEXAFSvs FTIR-ATR). (62) Naselli, C.; Rabolt, J. F.; Swalen, J. D. J . Chem. Phys. 1985, 82, 2136. (63) N d , C.; Rate, J. P.; Rabolt, J. F.; Swalen, J. D. Thin Solid Films 1985,134, 173. (64) Cohen, S. R.; Naaman, R.; Sagiv, J. J. Phys. Chem. 1986,90,3054. (65) Rabe, J. P.; Novotny, V.; Swalen, J. D.; Rabolt, J. F. Thin Solid Films 1988, 159, 359. (66) Schcrcr, J. R.; Snyder, R. G. J . Chem. Phys. 1980, 72,5798. (67) Scnak, L.; Moon, D.; Mmdebhn, R. 1. Phys. Chem. 1992, %, 2749. (68) If one u ~ e sn2 = 1.55 instead of 1.5, the average tilt angles of chain axes and transition momenta are ca. 3O higher and 2O lower, respectively, for
the systems studied.
The Inhlbtting Effects of Tetraalkylammonium Cations on Simple Heterogeneous Electron Transfer Reactions in Polar Aprotic Solvents W. Ronald Fawcett,* Milan Fedurco, and Marcin Opallo Department of Chemistry, University of California, Davis, California 95616 (Received: June 26, 1992; In Final Form: August 28, 1992)
Kinetic data for the electroreductionof nitromaitylene at mercury in propylene carbonate have been determined in the presence of tetraalkylammonium salts of varying chain length, both as a function of electrode potential and temperature. Although the standard rate constant decreases with increase in cation size, the experimental activation enthalpy is independent of the cation for variation in this ion from tetraethylammonium to tetraoctylammonium. These results indicate that the tetraalkylammonium ions are adsorbed on the electrode forming a blocking layer whose thickness incrmcs with alkyl chain length. Electron transfer to the redox center taka place by tunneling through this fh. Theoretical atimates of the activation enthalpy support this conclusion and suggest that the redox reaction occurs in a region where solvent reorganization determines the largest portion of its magnitude.
IntrodPCtion Studies of electrode kinetics are almost always carried out in the presence of an inert electrolyte in order to keep solution conductivity reasonably high and prevent migration as a means of mass transfer. In nonaqueous media, the salts used for this purpose often involve tetraalkylammonium cations (TAA+) because of their good solubility and weak tendency to form ion associates with anions in the system. However, it has been found that the kinetics of simple electron in a number of transfer reactions involving organic molecules depend on the nature of the TAA' ion, the usual trend being that the rate constant decreases with increase in the size of the alkyl group.'-' An interesting exception occurs when the TAA' cation is not symmetrical with respect to the alkyl groups.2 On the basis of this work, a general picture of the effect of TAA+ ions on the kinetics of simple electron transfer reactions is beginning to emerge. In general, it seems ~ l e a r ' . that ~ ? ~ the magnitude and direction of the double layer effect resulting from the presence of these ions cannot be predicted by the Frumkin model according to which the location of the reaction site is coincident with the outer Helmholtz plane (OHP).~Petersen and Evans6 showed that the inhibiting effect of the tetraheptylammonium (THpA+) cation with respect to that of the tetraethylammonium (TEA') cation became more pronounced for reactions occurring at more negative potentials. This suggests that TAA+ ions adsorb more strongly at the electrode/solution interface at more negative potentials, and thus have a greater inhibitory effect. Another important source of experimental information is the activation enthalpy obtained from the temperature dependence of the standard rate constant. This quantity is often independent of the nature of the TAA+ cation within experimental however, it does depend on cation nature when the orientation of the reactant in the double layer results in variation in the effect of the TAA+ cation on the transition state with variation in its size.' Petersen and Evans6 suggested on the basis
of their results that the probability of electron transfer through a T U +cation monolayer at the interface decreases with increase in cation size. The purpose of the present study was to extend previous work in a systematic way by obtaining kinetic data, especially, the activation enthalpy, for a simple reaction with a wide range of tetraalkylammonium electrolytesfor which the number of carbon atoms in the alkyl group varies from two to eight. The reaction chosen is the electroreduction of nitromesitylene (NM) at a mercury electrode, a system which has been investigated under a wide variety of conditions.4-"' The solvent chosen is propylene carbonate (PC) which has a high dielectric constant (64.9 at 25 "C12) so that ion pairing effects are expected to be relatively unimportant. Moreaver, in this system the reaction occurs at fairly negative charge densities on the electrode where the double layer capacity does not change significantly with p0tentia1.l~ This is important when one wants to obtain precise data with the ac admittance technique used for the present study. ExperimentalSection
MateMs. Nitromesitylene ( l-nitro-2,4,6-trimethylbenzene) (Aldrich) and cobalticinium hexafluorophosphate (Strem) were used as received. All salts, tetraethylammonium (TEAP), tetrapropylammonium (TPAP), tetrabutylammonium (TBAP), tetrahexylammonium (THxAP), and tetraoctylammonium (TOAP) perchlorates (Fluka), were twice recrystallized from a water-methanol mixture and dried under reduced pressure at 70 "Cfor 24 h. AgC104(Strem) was used as received. The solvents, propylene carbonate (Burdick & Jackson) and acetonitrile (AN) (HPLC grade, Aldrich), were purified as previously dc~cribed.'~3 The nitrogen used for electrochemical experiments and distillations was 99.998% pure. Apparatus. The apparatus for phase sensitive ac admittance voltammetry was similar to that described previo~sly.'~J~ The Solartron 1250 frequency response analyzer served as a source
0022-3654/92/2096-9959$03.00/08 1992 American Chemical Society
