J. Phys. Chem. 1988, 92, 5575-5578 not previously been implemented. The expanded form, eq 9, has been used, as has eq 7, in recent studies of nuclear shielding tensors and atomic polar tensors at the SCF-RPA level of approximation.14*i8v19It has been foundi4.i8J9that when very large basis sets are used eq 7 and 9 give very similar results. When smaller basis sets are used, "polarized" to optimize the accuracy of eq 7 and 9, eq 9 is found to yield results of greatly inferior accuracy to those of eq 7.l49l8 Thus, the relative accuracy of eq 7 and 9 appears to parallel that of eq 4 and 14. The calculation of vibrational dipole strengths and absorption tensors. When e @ ( A )( A = p, x ) tensors intensities utilizes are employed, alternative representations of dipole strengths, D(A), result. The results obtained here demonstrate that D(p) must be more accurate than D(7). The calculation of vibrational rotational strengths and circular dichroism intensities utilizes both and M & tensor^.^^^ When P$(A) and M;@(DO,B) ( A , B = p, x ) tensors are employed, the four alternative representations of rotational strengths R(DO,A,B) result. When &(A) and M$(CO) tensors are employed, the two alternative representations, R(C0,A) result. R(DO,A,x) = R(C0,A) follows from M$(DO,x) = M$JCO). The results obtained here demonstrate that R(DO,p,p) must be the most accurate of the various alternative represeatations. In addition, it is easily shown that R(DO,A,B) is origin-
ea.
e,,
(17) Buckingham, A. D.; Fowler, P. W.; Galwas, P. A. Chem. Phys. 1987,
112, 1. (18) Lazzeretti, P.; Zanasi, R. J . Chem. Phys. 1985, 83, 1218. (19) Jalkanen, K. J.; Stephens, P. J.; Lazzeretti, P.; Zanasi, R., submitted
for publication.
5575
independent when A = B3 and origin-dependent when A # B. R(DO,p,p) thus possesses the additional benefit of origin independence. The greater accuracy of R(DO,p,p) over R(C0,p) values and the origin dependence of R(C0,p) have been demonstrated explicitly for the chiral isotopomer of NH3, NHDT,li and a range of larger molecules.is Recent comparisons of predictions of vibrational rotational strengths and vibrational circular dichroism spectra with experimental data12J3have been based on calculations of R(DO,p,p). The present work shows that calculations of and Mi,(DO,p) tensors, and of R(DO,p,p) values therefrom, are m$re efficiently impkmented via the CHF calcylation of [a$G(R)/ ax,,] [a$G(RO,A@) /aA@]A and [a$G(RO,H@)/aH@1H p o ) the last of these at one origin $;:than via the C H F calculation of [4&?)/a&]~, and [wG(&,H@)/aH@],, the latter at N origins. Future calculations from our laboratories will use this superior computational strategy. The conclusion is general for any choice of distributed origins in the distributed origin gauge and will not be affected by any future further optimization of this choice. The conclusions that is of superior accuracy to and that consequently M;@(DO,p) is more accurate than M$(CO) = M$(DO,x) have been established for "conventional" basis sets. It is clearly possible that the use of alternative types tensors of accuracy comparable of basis sets can lead to The development of such basis sets would be of great to interest. $9
e&)
e&).
e@(")
e@(")
Acknowledgment. We acknowledge the support of NSF, NIH, NATO, and the San Diego Supercomputer Center.
Electrochemical and Raman Characterization of Molecular Recognition Sites in Self-Assembled Monolayers J.-H. Kim, T. M. Cotton,* and R. A. Uphaus* Department of Chemistry, University of Nebraska-Lincoln, Lincoln, Nebraska 68588-0304 (Received: May 26, 1988)
Contour-specific sites, produced by silanized monolayers, were shown to possess a pronounced capacity for recognition of a specific molecular species if this species was also present during the original monolayer formation. The preselected species (or template compound) can be subsequently reincorporated into the skeletonized monolayer. The topospecific character of the sites was evident by the extent of reconstitution of the original systems upon reexposure to the template molecules and by the rejection of dissimilar species. The fractional surface occupied by sites was estimated by electrochemical and surface-enhanced resonance Raman measurements. Sites accepting several related porphyrins were produced on glass, tin oxide coated glass, and silver.
Introduction The aims of this study were (1) to establish the degree of molecular recognition capacity shown by a porphyrin template for a series of porphyrins in contradistinction to acceptance of some slightly larger molecules, (2) to measure the intensity of surface-enhanced Raman scattering (SERRS) from the adsorbed species as a function of the template area present, and (3) to relate the density of sites to electrochemical activity and provide correlation with SERRS and electrochemical measurements. The initial study of monolayer molecular recognition systems by S E R R S indicated that a high degree of discrimination for molecular contours can be attained by the use of skeletonized silylated monolayers.' Species reincorporated into template sites were indicated by use of SERRS; this technique is capable of detecting (1) Kim, J.-H.; Cotton, T. M.; Uphaus, R. A. Thin Solid Films 1988, 160, 389. Cf. Absrracts of Papers, Third International Conference on Langmuir-Blodgett Films, Gattirigen, FRG, July 1987.
chromophores at submonolayer concentrations (femtomolar or less).2 The present study uses the technique originally developed by S a g i ~cosorption :~ of octadecyltrichlorosilane (OTS) and a surface-active cyanine dye onto a glass surface produced a closely packed mixed monolayer which preserved the contour of the dye molecule after the dye was removed by solvent. The chemically bonded OTS thus becomes a skeletonized monolayer and is capable of resorbing the dye and reconstituting the system originally formed. Such structures showed considerable discriminative capacity, rejecting dyes of closely related structure upon attempted reconstitution. (2) Uphaus, R. A.; Cotton, T. M.; Mobius, D. M . Thin Solid Films 1985, 132, 173. (b) Cotton, T. M.; Uphaus, R. A,; Mobius, D. J. Phys. Chem. 1986, 90, 6071. (3) Sagiv, J. Isr. J. Chem. 1979, 18, 339, 346; J . Am. Chem. SOC.1980, 102, 92.
0022-3654/88/2092-5575$01.50/0 0 1988 American Chemical Society
5576 The Journal of Physical Chemistry, Vol. 92, No. 20, 1988
Letters
Subsequent to the revival of the Sagiv method by the more recent studies,' results of electrochemical studies on silylated template systems have been reported by Tabushi et al.4a and Yamamura et al.4b Silylated matrices on silica gel have been r e p ~ r t e d . ~Additionally, recent studies by Rubinstein et a1.6 have provided a comprehensive characterization of the electrochemistry of monolayers formed by the self-assembly of silane and octadecyl mercaptan on conducting surfaces. Other electrochemical studies' of alkyl mercaptans and silanes on gold surfaces indicate effective blocking of electron transfer with minute currents carried by pinhole defects and defect sites. FTIR-ATR has also been used for the characterization of self-assembled monolayers of silanes.* The Sagiv technique may be more versatile and facile than approaches that fabricate recognition sites by polymer-embedded cavities having a specific topological and functional complementarity with the desired guest c o m p o ~ n d . ~ J ~
Experimental Methods Systems were made up by the techniques originally described in detail by S a g i ~ Le., ; ~ surfaces were exposed (10-30 min) to hexadecane in which was dissolved octadecyltrichlorosilane and also the compound intended as the template molecule. The mole fraction of template compound was varied in the range 0-10%. The binucleate phthalocyanine S-25 (two macrocycles joined by a bridging oxygen) was provided by Leznoff." Bacteriochlorophyll a (BChl) was prepared by literature methods.I2 Magnesium phthalocyanine (MgPC), tetraphenylporphine (TPP), zinc tetraphenylporphine (ZnTPP), and octadecyltrichlorosilane (OTS) were used as received from Aldrich or Sigma Chemical Co. All other reagents and solvents were analytical or reagent grade, used without further purification. SERRS measurements were made using the 457.9-nm line of an Ar' laser using the Raman system described previously.* A polycrystalline Ag disk electrode was used as the working electrode in the electrochemical experiments. Underpotential deposition of Pb on Ag was carried out in the conventional mannerI3 with a BAS- 100 electrochemical analyzer. Results and Discussion Recognition template systems were prepared by using ZnTPP as the chosen template molecule. Following removal of the template molecule by solvent, the monolayers were exposed to solutions of various compounds of similar or slightly larger structure (Figure 1, top). After exposure, the slides were rinsed briefly with toluene, dried under nitrogen, and coated with thin ( 5 nm) Ag films. As shown in the SERRS spectra in Figure 1, ZnTPP (Figure 1A) and its demetalated form, TPP (Figure lB), were readily reconstituted into sites based upon ZnTPP. Additionally, ZnTPP sites also accepted the structurally quite similar TPyP (Figure 1C) as well as the slightly smaller molecule PPIX (4) (a) Tabushi, I.; Kurihar, K.; Naka, K.; Yamamura, K.; Hatakeyama, H. Tetrahedron Lett. 1987,28,4299. (b) Yamamura, K.; Hatakeyarna, H.; Naka, K.; Tabushi, I.; Kurihara, K. J . Chem. Soc., Chem. Commun.1988, 79. (5) Tao, Y.-T.; Ho, Y.-H. J . Chem. SOC.,Chem. Commun. 1988, 417. (6) (a) Sabatani, E.; Rubinstein, I.; Moaz, R.; Sagiv, J. J. Electroanal. Chem. 1987, 219, 365. (b) Sabatani, E.; Rubinstein, I. J . Phys. Chem. 1987, 91, 663. (c) Rubinstein, I.; Steinberg, S.;Tar, Y.; Shanzer, A.; Sagiv, J. Nature (London) 1988, 332, 426. (7) (a) Finklea, H. 0.; Robinson, L. R.; Blackburn, A,; Richter, B.; Allara, D.; Bright, T. Langmuir 1986,2,239. (b) Finklea, H. 0.;Lynch, M.; Furtsch, T. Ibid. 1987, 3, 409. (8) (a) Maoz, R.; Sagiv, J. Lungmuir 1987, 3, 1034. (b) Maoz, R.; Sagiv, J. Ibid. 1987, 3, 1045. (9) For a comprehensive review, see: Wulff, G. ACS Symp. Ser. 1986, NO. 308, 186-230. (10) Cf. Shea, K. J.; Dougherty, T. K. J . Am. Chem. Soc. 1986,108, 1091,
and references cited therein for recent efforts using this approach. (1 1 ) Surface isotherms of this compound spread at the air/water interface indicate an apparent molecular area consistent with a folded, cofacial structure. Absorption on a surface would produce a cross section little different from the area of the simple monomer. Cf. Kim, J.-H.; Cotton, T. M.; Uphaus, R. A.; Leznoff, C. C. Thin Solid Films 1988, 159, 141. (12) Strain, H. H.; Svec, W. A. In The Chlorophylls; Vernon, L. P., Seely, G. R., Eds.; Academic: New York, 1966; p 21 ff. (13) (a) Guy, A. L.; Pemberton, J. E. Langmuir 1985, I , 518. (b) Guy, A. L.; Bergami, B.; Pemberton, J. E. Surf.Sci. 1985, 150, 226.
I
I
I600
aoo Raman Shift (cm") 1200
400
Figure 1. SERRS spectra of reconstituted template systems and controls: A, ZnTPP; B, TPP; C, TPyP; D, PPIX; E, control overcoated with Ag island film; F, control-totally silylated glass slide overcoated with Ag island film; G, BChl a; H, MgPC; I, S-25 (dimeric phthalocyanine). Experimental conditions: excitation wavelength, 457.9 nm; power, 10 mW; 10 scans, 1-s integration per scan; silver overlayer thickness, 5 nm.
(Figure 1D). In all cases, the spectra of the porphyrins were quite similar to their R R spectra in solution, with the major differences involving relative band intensities. Exposure of the ZnTPP template systems to solutions of three slightly larger molecules, BChl a (Figure lG), MgPC (Figure lH), and the binuclear S-25 phthalocyanine (Figure lI), failed to produce evidence that these molecules were capable of reconstitution into sites with the use of an expanded (2X) intensity scale. The large, broad peak near 1106 cm-l is characteristic of Agcoated glass slides and may be due to an Si-0-Si vibration.' Thus, these spectra were indistinguishable from those obtained from Ag-coated clean glass (Figure 1E) or silylated glass (Figure 1F). A series of ZnTPP template systems were prepared on both glass and silver surfaces, using increasing mole ratios of ZnTPP/OTS during monolayer formation. A completely satu-
Letters
The Journal of Physical Chemistry, Vol. 92, No. 20, 1988 5577 1.ot
/
,
0
2
4
6
8
10
Potential (V vs
SCE)
Potential (V vs
SCE)
I
100
ZnTPP mole fraction (%)
Figure 2. Relative SERRS intensity of ZnTPP in reconstituted templates as a function of the ZnTPP/OTS mole percent used in original template preparation. The solid line is for a glass substrate, and the dotted line is for a silver electrode. Experimental conditions were the same as in Figure 1. A 1 X lo4 M solution of ZnTPP in dry toluene was used to reconstitute the templates. The ratios were calculated relative to the signal intensity observed for adsorbed ZnTPP on glass or Ag (Le., in the absence of OTS).
rated surface was defined as that which is produced by adsorption on a clean glass surface or Ag electrode exposed to a ZnTPP solution. The relative SERRS intensity for the silylated surfaces was then determined by comparing the signal intensity for the template system to that obtained for pure ZnTPP on glass (solid line) or on the Ag electrode (dashed line), as shown in Figure 2. Both substrates show a rapid rise in SERRS intensity in the 0-3 mol % region, with a slower rise a t higher concentrations of template. ID the case of the glass surface, however, the overall SERRS intensity is much higher in the region between 4 and 10 mol %. This must reflect the stronger interaction between ZnTPP and the glass surface, as compared to Ag. With respect to the actual template area, it should be noted that the relative molecular areas of ZnTPP and OTS differ by about an order of magnitude, SO that even at a mole ratio of L0:l OTS/ZnTPP the sites are rather closely spaced, as will be evident from the electrochemical data. Electrochemical methods provide a more quantitative estimate of the area of the template sites relative to that of the OTS monolayer. Underpotential deposition (UPD) of Pb was used for this purpose.I4 Figure 3A illustrates results obtained for bare Ag (dashed line) and completely silylated Ag (solid line) electrodes. The integrated charge under the anodic peak was taken as a measure of the number of Pb atoms deposited. Using this value and the radius of a Pb atom, we calculated the exposed surface area. In the case of the completely silylated surface, between 5 and 10% of the electrode area was exposed. This is undoubtedly due to pinholes that are apparently smaller than the area of the template molecule, since no SERRS could be detected for completely silanized surfaces. This value could be reduced to less than 3% if octadecyl mercaptan is used following the OTS treatment, as has been noted by Rubinstein et aL6 Figure 3B illustrates results obtained for the bare Ag (dashed line) and the silylated surface, using 5 mol % ZnTPP/OTS solution (solid line) in the template preparation. It can be seen that the reduction peak is shifted for the template. This may reflect the difficulty of transporting Pb2+ atoms through the hydrophobic monolayer to the electrode surface. At this mole ratio, the exposed area (as (14) The authors appreciate the suggestion by I. Rubinstein, who pointed out the utility of such measurements.
:;
Figure 3. Cyclic voltammetry of Pb UPD: (A) Ag electrode following OTS treatment (solid line) and in the absence of OTS (dashed line). (B) Ag electrode containing skeletonized template prepared with 5 mol % ZnTPP/OTS (solid line) and in the absence of template (dashed line). Experimental conditions: 1 mM Pb(NO& solution; scan rate, 50 mV/s; supporting electrolyte, 0.1 M NaNO,; initial potential, 0.0 V vs SCE; scan direction, cathodic.
Oa41/ 0.2
P
0
I
2
t
I
4
6
I
8
10”
100
ZnTPP more fractlon (%I Figure 4. UPD charge ratio vs ZnTPP/OTS mole fraction used in the preparation of the template. The charge was determined by integrating
the oxidation peak, following background subtraction. determined from the anodic peak) is approximately 60% of the total area. Values obtained for similar UPD measurements on slides prepared froin various mole ratios of template of OTS are depicted in Figure 4. The area occupied by the template molecules rises monotonically and reaches ca. 90% of the total area at 10 mol % of template coverage. A comparison of the results shown in Figure 2 (dashed curve) with those in Figure 4 shows that the SERRS intensity change with increasing template concentration does not follow the increase in surface area, especially at low template concentrations. Rather, the SERRS intensity increases rapidly between 0 and 3 mol % template. At higher mole ratios of template, the intensity increases more slowly with increasing template concentration. These latter results are similar to data obtained with Langmuir-Blodgett monolayers from which it was determined that the relationship between SERRS intensity and dye coverage was nonlinear and the magnitude of the enhancement decreased at dye/lipid ratios greater than ca. 10 mol %.I5 In general, the relative SERRS intensities are lower at all concentrations than the relative surface ( 1 5 ) It has been shown that the intensity of the SERRS signal is proportional to the mole fraction of the chromophore to a concentration of ca. 10 mol % (Kim, J.-H.; Cotton, T. M.; Mobius, D.; Uphaus, R. A., unpublished results).
J. Phys. Chem. 1988, 92, 5578-5580
5578
areas as determined from UPD. This may reflect the presence of pinholes which add to the exposed surface but are not sufficiently large to permit ZnTPP incorporation. This possibility may be tested by two different methods. The first is an independent measurement of the amount of adsorbate in a reconstituted system using an electroactive adsorbate. The second involves the use of n-octadecyl mercaptan (OM) to block pinhole sites. In conclusion, the results presented herein demonstrate the considerable selectivity provided by self-assembled monolayers. A series of porphyrins of similar size and shape were readily reconstituted in sites prepared from ZnTPP. On the other hand, MgPC and BChl, which differ only in the substituents on the periphery of the macrocycle, were not reincorporated. The
quantitative estimates of ZnTPP by SERRS were found to be somewhat lower than the surface area measurements provided by UPD. The difference is probably due to the presence of pinholes in the OTS/template system. The total area of the OTS matrix and template system appears measurably below the ideal 100% value; for the purposes of construction of molecular-recognizing template sites, it would appear that the small void volume always present is not an impediment to the efficiency of the formation of systems having a high degree of species specificity.
Acknowledgment. The financial support of the U S . Department of Energy, Chemical Sciences Division (DE-FGO284ER13261), is gratefully acknowledged.
Photochemistry of OS/HNCO Mixtures A. P. Ongstad, X. Liu, and R. D. Coombe* Department of Chemistry, University of Denver, Denver, Colorado 80208 (Received: June 6, 1988: In Final Form: July 25, 1988)
- -
Pulsed photolysis of gaseous 0 3 / H N C 0 mixtures at 249 nm produces emission from the N H A311 X'Z- transition near 336 nm. The data suggest that the excited N H is produced by the reaction O(ID) + HNCO N H + COz. The branching fraction to the alternate products OH NCO is less than 0.1. This mechanism is quite different from that of the analogous O('D) HN,, "('A) + HN,, and "('A) + HNCO reactions. From the time profile of the N H emission, the rate constant for O(lD) + HNCO is determined to be (4.6 0.4) X lo-" cm3 s-l.
+
+
*
Introduction In a recent study' of the photolysis of 03/"3 mixtures, it was shown that the reaction of O(ID) atoms with H N 3 is rapid and produces OH N3. This behavior is similar to that of the isoelectronic NH(alA) reaction with HN3? which produces NH2 NS. The reaction of O(lD) atoms with N 3 radicals occurs subsequent to O(lD) H N 3 and produces chemiluminescence from the A28+ X211 transition in NO. The temporal profile of this emission was modeled to obtain rate constants for the O(lD) H N 3 and O(lD) + N 3 elementary reactions. In this paper, we describe similar observations of the reaction between O(lD) and HNCO. The analogous NH(a'A) + H N C O reaction has been studied from HNCO p h o t ~ l y s i sand ~ * ~is thought to produce N H 2 NCO. The rate constant for NH(a'A) H N C O has been reported to be 1.45 X cm3 s-l by Drozdoski and c o - ~ o r k e r s .The ~ results of the present experiments indicate that O(lD) + H N C O proceeds by a route entirely different from that found for either the O(lD) HN3 reaction or the NH(alA) reactions noted above.
+
+
-
+
+
+
+
+
Experimental Section The methods employed were similar to those described in ref 1. O3 (about 65% purity, the remainder being 0,) was prepared with a commercial ozonizer and stored in a Pyrex flask. H N C O was generated by heating mixtures of KOCN and excess stearic acid to 373 K and collecting the gaseous product in a Pyrex flask. The gaseous H N C O was diluted to a 10% mixture with helium. FTIR analysis indicated no apparent impurities in the gas mixture. Slow flows of O3and HNCO were mixed upstream of a photolysis cell, which was 20 cm in length and fabricated from stainless steel. (1) Ongstad, A. P.; Cmmbe, R. D.; Neumann, D. K.; Stech, D. J. J. Phys. Chem., in press. ( 2 ) McDonald, J. R.; Miller, R. G.; Baronavski, A. P. Chem. Phys. 1978, 30, 133. ( 3 ) Drozdoski, W. S.; Baronavski, A. P.; McDonald, J. R. Chem. Phys. Lett. 1979. 64. 421. (4) Spiglanin, T. A,; Perry, R. A,; Chandler, D. W. J . Phys. Chem. 1986, 90, 6184.
0022-3654/88/2092-5578$01.50/0
The total pressure in the cell was measured with a capacitance manometer. The partial pressure of O3in the cell was measured by optical absorption at 253.7 nm. There was no evidence of prereaction between the flowing O3and HNCO. The gas mixtures were photolyzed with the 249-nm output of a pulsed excimer laser. Chemiluminescence produced by the photolysis was detected with a 0.25-m monochromator and cooled GaAs photomultiplier tube. The presence of nonemitting species was probed by recording laser-induced fluorescence (LIF) excitation spectra. A nitrogen laser pumped dye laser (PRA LN107) was used as a probe in the LIF experiments. Spectra were recorded with a gated boxcar integrator interfaced to a microcomputer. Time profiles of the emission were recorded with a Nicolet 1270 pulsed data collection system which also was interfaced to a microcomputer. Analysis of the data was performed with an RS/1 statistical package on a VAX 780 mainframe computer.
Results and Discussion Since the absorption cross section of O3is very large5 and that of H N C 0 6 is very small at 249 nm, photolysis at this wavelength selectively dissociates the 03.Excited O(lD) and 02(a1A)are produced in 90% yield.' To the extent that this system is similar to O3/"3, the O(*D)would react with H N C O to produce OH NCO and then with NCO to produce excited N O and CO. The latter reaction (O'D + NCO) is exothermic by 148 kcal mol-', easily sufficient to produce NO(A2Z+). Figure 1 shows a portion of the UV chemihminescence recorded subsequent to photolysis at 249 nm. There is no detectable emission from the A X transition in NO. Instead, the spectrum is dominated by emission from the A311 X3Z- transition in N H near 336 nm. Very much less intense emission from the A2Z+ Xzll transition in OH was found when the maximum amplifier gain of the detection system was employed. Although these transitions in N H and OH were also observed' from photolysis of 03/"3, they were more than
+
-
-
-
( 5 ) Braun, W.; et al. J. Phys. Chem. Ref. Data 1973, 2, 267.
(6) Rabalais, J. W.; McDonald, J. R.; McGlynn, S . P. J . Chem. Phys. 1969, 51, 5103. (7) Brock, J. C.; Watson, R. T. Chem. Phys. Lett. 1980, 71, 371.
0 1988 American Chemical Society
