Langmuir 1991, 7, 2821-2826
2821
Direct Diffusion Measurements of Naphthacene on Chemically Derivatized Silica Geld D. W. Bjarneson and N. 0. Petersen' Department of Chemistry, The University of Western Ontario, London, Ontario N6A 5B7,Canada Receiued January 28,1991. In Final Form: June 24, 1991 Fluorescence microscopy techniques are employed to investigate the mobility of polycyclic aromatic hydrocarbons on hydrocarbon-derivatizedsilica gels. The silica gel is derivatized using chemisorption (silylation)and physisorption techniques. The mobility of naphthacene on various surfaces is examined in detail. The diffusion coefficient and mobile fraction are greater for naphthacene on these supports as compared to the mobility of naphthacene on normal-phase silica gels. All results are rationalized in terms of the fractional coverage of the surface and the interactions between the derivatized surface and the fluorescent probe.
Introduction The photophysical behavior of polycyclic aromatic hydrocarbons (PAHs) on many hydrocarbon-derivatized silica surfaces or reverse-phase chromatography supports has been The probes of choice in these investigations are generally pyrene or pyrene derivatives due to the photophysical properties of the pyrene moiety. Evidence for the mobility of adsorbates on these surfaces arises from bimolecular quenching studies and pyrene exhibiting dynamic excimer formation. Bogar et al. report the diffusion coefficient for pyrene on octadecyl-coated silica in contact with a 3:l methanol/water mixture to be 2.5 X 10-7 cm2 s-l.5 They also report that pyrene is immobile on the same surface in the absence of a wetting solvent, since no dynamic excimer formation was observed. We are interested in examining the lateral mobility of adsorbates on various silica surfaces. Recently we reported on the mobility of naphthacene on normal-phase silica gels.I2 In a previous study, we reported on the use of fluorescence microscopy techniques to measure the diffusion of naphthacene directly. On dry silica gel the diffusion coefficient (D)was found to be invariant ((2.4 f 0.2) X 10-lo cm2 8-l) with pretreatment temperature. The mobile fraction (X& however, has a pronounced dependence on pretreatment temperature, and varies from 0.33 to 0.53. On silica gel with 0.2 Fmol m-2 water added, the same diffusion coefficient was observed for all pretreatment temperatures. The mobile fraction was greater on these surfaces (0.50-0.67),and once again showed a dependence on the pretreatment temperature. In this paper, we use the same microscopy techniques (1) Supported by NSERC, Canada. (2) LochmiUler, C. H.; Colbom, A. S.; Hunnicutt, M. L.; Harris, J. M. Anal. Chem. 1983.55.1344. (3) Lochmbller,'C. H.-Colborn, A. S.; Hunnicutt, M. L.; Harris, J. M. J. Am. Chem. SOC.1984,106,4077. (4) Lochmdler, C. H.; Hunnicutt, M. L. J. Phys. Chem. 1986,90,4318. (5) Boaar. - R. G.: Thomas, J. C.: Callis. J. B. Anal. Chem. 1984.. 56.. 1080. (6) SWberg, J.; Almgren, M.; Alsins, J. Anal. Chem. 1988,60,2487. (7) Lisichkin, C.V.;Runov, V. K.; Staroverov, S. M.; Fadeev, A. Yu. Dokl. Akad. Nauk SSSR 1988,299,313. (8)Avnir, D.; Busee, R.; Ottolenghi, M.; Wellner, E. J . Phys. Chem. 1986,89,3521. (9) Bauer, R. K.;de Mayo, P.; Ware, W. R.; Wu, K. C. J . Phys. Chem. 1982.86.3781. --, (10) Bauer, R. K.;de Mayo, P.; Okada, K.; Ware, W. R.; Wu, K. C. J. Phys. Chem. 1983,87,480. (11) Bauer, R. K.; de Mayo, P.; Natarajan, L. V.; Ware, W. R. Can. J. Chem. 1984,62, 1279. (12) Bjameson, D.W.;Petersen, N. 0. J. Am. Chem. SOC.1990,112, 988.
----.
to measure the mobility of naphthacene on a number of hydrocarbon-derivatized silica gels. The silica is derivatized by either a silylation reaction, physisorption of a long-chain alcohol or fatty acid, or a combination of these two procedures. An increase in both the diffusion coefficient and the mobile fraction is observed for naphthacene on these surfaces over those measured for naphthacene on normal-phase silica gel. Both of these diffusion parameters also show a dependence on the derivatized surface employed. On surfaces modified by chemisorption of alkyl chains, the diffusion coefficient is found to be 6.6 X 10-locm2 s-1, and the mobile fraction varies from 0.50 to 0.62. On surfaces modified by physisorption of alkyl chains, the diffusion is measured to be 2.6 X lo4 cm2 s-l, and the mobile fraction is at least 0.75. All results are discussed in terms of the interactions between the surface and adsorbate interactions.
Experimental Section Surface Derivatization. All hydrocarbon-derivatizedsurfaces used were prepared in our laboratory. The derivatization A dilute solution procedure used was that employed by (0.1% v/v) of the silylating agent in 80% octane, 12% carbon tetrachloride,and 8% chloroform (allpurchased from BDH)was stirred with an appropriate mass of silica gel (Mallinckrodt, surface area 300 m* g-l, average pore diameter 150 A). Enough derivatizingagent was used to give a slight excess in monolayer coverage,provided the reaction went to completion. This is based on a projected area of 20 Az per alkyl molecule,*5and the assumptionthat the long-chain hydrocarbons are perpendicular to the surface. The reaction was typically run for 15 min, but one sample was prepared with a much longer reaction time (24 h). The silica was vacuum filtered and rinsed with aliquota of chloroform, doubly distilled deionized water, and chloroform again. Before use, the silica was air cured for at least 24 h. Three silylating agents were used to produce hydrocarbonderivatized silica gels. Each derivatizedsurface will be identified by the number of carbons in the chain, for example C18 will be used to represent octadecyl-coated silica. The three agenta employed were C1 (trimethylchlorosilane),C8 (octyldimethylchlorosilane),and C18 (octadecyltrichlorosilane),all purchased from Fluka. These reagents were used without any further purification. Myristicacid (Sigma)was physically adsorbed to normal-phase silica gel, and 1-undecanol (Aldrich) was physisorbed on C18 silica gel prepared as indicated above. Deposition of both of (13) Sagiv, J. J. Am. Chem. SOC.1980, 102, 92. (14) von Tscharner, V.;McConnell, H. M. Biophys. J . 1981,36,421. (15)Adamsom, A. Physical Chemistry ofSurjaces,5th ed.;John Wiley and Sons: New York, 1990, p 142.
0743-7463/91/2407-2821$02.50/0 0 1991 American Chemical Society
Bjarneson and Petersen
2822 Langmuir, Vol. 7, No. 11, 1991
these physisorbed species was from cyclohexane solution. In each case, the amount of hydrocarbon physisorbed was enough to give roughly one monolayer coverage. These reagents were used without further purification. Fluorescence Photobleaching Recovery (FPR). Samples used for FPR experiments were prepared in the same manner outlined previously.lz Briefly, the derivatized silica gel was put in the vacuum chamber and evacuated at room temperature overnight. Following evacuation, an appropriate volume of a degassed solution of naphthacene in cyclohexane was added to the chamber. The solvent was then removed slowly, and the silica and solid naphthacene were mixed and shaken for a short period of time. This resulted in sublimation of naphthacene onto the surface. For some samples, the naphthacene was deposited directly from cyclohexane solution. The surface concentration of naphthacene was kept constant at roughly0.1 % of a monolayeror 8 X 1Wrmol m-2. This assumes little or no change in the surface area of the silica gel upon derivatization. All measurements were done at room temperature, with the sample held in a 1-mm quartz cuvette. Parameters for FPR experiments with naphthacene were as follows: 75-100 ms/channel, 80-150-ms bleach pause, 20-s postrecovery pause, and an attenuation of 2.11-2.8 OD. The beam radius for all measurements was 1.0 pm. Results a n d Discussion Qualitative Observations. Adsorption of naphthacene via sublimation on the C18 silica surface yields a relatively homogeneous distribution as observed in the fluorescence microscope. Typically there is some microcrystalline material present, but far less than on any of the normal-phase silica gels. Adsorption from solution leads to the most homogeneous distributions, with very few microcrystals. Adsorption via sublimation onto either a C8 or C1 surface yields an inhomogeneous distribution. More crystalline material is present on C1 silica, but even on C8 silica there is a significant amount. For most samples prepared, the fluorescence intensities in regions of noncrystalline naphthacene are very low, resulting in a poor signal to noise ratio. Adsorption from solution on either C1 or C8 silica leads to a significant amount of microcrystalline material also. On C8 silica, the amount is somewhat less than for sublimed naphthacene. On C1 silica, most of the naphthacene is still crystalline. Consequently, very few quantitative results are obtained for naphthacene on C1 or C8 silica. The heterogeneity of these probe distributions may be explained by the interactions between the surface and the adsorbed aromatic hydrocarbon. On a normal-phase surface, the attractive interactions are hydrogen bonds, which are relatively strong.lB-l8 On a hydrocarbon-derivatized surface, the largest forces of attraction are dispersion forces between the hydrocarbon chains and the aromatic molecule. The magnitude of these forces in turn depends on the chain length and density of the coating. As the chains become shorter, the number of interactions per naphthacene molecule decreases. An extreme is reached with the one-carbon chain. This very short chain will yield the smallest number of interactions per naphthacene molecule. C8 chains are approximately as long as a naphthacene molecule, whereas the C18 chain is more than twice as long. Naphthacene will experience the greatest number of dispersion forces on the C18 surface. (16) de Mayo, P.; Natarajan, L.V.;Ware, W. R. J. Phys. Chem. 1985, 89, 3526. (17) deMayo,P.;Natarajan,L.V.;Ware, W.R.InOrganicPhototransformtione in Nonhomogeneous Media; Fox, M. A., Ed.; American Chemical Society: Washington, DC, 1985; Chapter 1. (18) Pohle, W. J. Chem. Soc., Faraday Trans. 1 1982, 78,2101.
If naphthacene adsorbs to the surface of the alkyl coating, then the distribution of naphthaceneon all coatings should be equally homogeneous. This is not observed, indicating that naphthacene is probably found within the alkyl coating, and not adsorbed on the upper surface. Adsorption from solution in all cases yields fewer microcrystals. The adsorption process here is quite different. Infrared spectroscopy data indicate that, in the absence of a wettingsolvent, the hydrocarbon chains will be laying tangled on the surface.1e With a wetting solvent present, the long hydrocarbon tails become untangled and are free to adopt many different configuration^.^ As the solvent is evaporated, the naphthacene is trapped by the hydrocarbon chains as they settle back onto the surface. This is particularly important in the C18 case. Naphthacene adsorbed on C18 silica shows no evidence of desorption when left on the vacuum line overnight. Naphthacene desorbs from normal-phasesilica pretreated at high temperatures, as shown by coloration of a cotton plug in the vacuum chamber.12 Since no desorption is detected from a C18 surface, the cumulative effect of a number of weaker interactions per naphthacene molecule must be greater than a few, individually stronger, interactions on the normal-phase silica. This indicates that naphthacene may be trapped in the hydrocarbon entanglement or at least adsorbed within the hydrocarbon layer, and not adsorbed on the hydrocarbon surface. On the other hand, an increase in the diffusion coefficient for naphthacene on C18 silica over that for naphthacene on normal-phase silica indicates the strength of the interactions binding naphthacene to normal-phase silica is greater than that binding naphthacene to a C18 surface. Thus, the activation energy for diffusion of naphthacene on alkyl-coated silica gels is less than that of naphthacene on normal-phase silica, while the activation energy for complete desorption from the alkyl coating is greater than for desorption of naphthacene from normal-phase silica. We defined the success rate of FPR experiments as the fraction of the total number of experiments that are used in further calculations of the diffusion parameters. For C18 silica the success rate is 80-85%. The efficiency of the photochemistry is greatly enhanced on the alkyl-coated surface relative to the normal-phase silica surface. With a strong bleaching pulse, quite often too many of the fluorophores are destroyed. As a result, the perturbation to the system is too large to have great confidence in the calculated diffusion parameters. Data obtained from experiments where the extent of bleaching exceeds a predetermined level are not used in further statistical analyses. Quantitative Observations. Diffusion of PAHs occurs on a variety of derivatized silica surfaces. Figure 1 shows a typical FPR recovery curve along with the fit to the data for naphthacene on C18 silica with undecanol coadsorbed. The measured diffusion parameters for naphthacene adsorbed on various alkyl-coated silica gels are presented in Table I. As shown in a previous paper of ours, naphthacene is mobile on normal-phase silica, and experiences at least two types of microenvironments, those in which naphthacene is mobile and those in which it is not.'2 Chemisorption or physisorption of long alkyl chains provides for at least one more microenvironment. For simplicity, the derivatized silica surface is modeled to have only three environments: normal-phase silica, where the naphthacene populations are grouped into two categories, a mobile one (1)and an immobile one (2), and alkyl-coated silica, where naphthacene is assumed to be completely ~
_
_
_
_
(19) Sander, L.C.; Callis, J. B.; Field, C. R. Anal. Chem. 1983,55,1068.
Langmuir, Vol. 7,No. 11, 1991 2823
Direct Diffusion Measurements of Naphthacene CU0212.FPR 90-06-08 1 0 : 4 3 : 2 1
1. 0.
IY 0.
I
I
aM.
600.
CaWL
I
la4.
,
D = CJiDi with
Xii = 1
0.
-101.
at low temperatures.12 With physisorption of undecanol on C18 silica, the mobile fraction increases further. The significant increases in both D and X, show that the mobility of naphthacene in an environment very different from normal-phase silica is contributing to the recovery of the fluorescence intensity in the FPR experiment. This must be the hydrocarbon environment provided by derivatization of the surface with the various alkyl chains. Chemisorption of alkyl chains will decrease both the surface area which appears to be normal phase and the amount of naphthacene adsorbed on normal-phase silica. The fraction of molecules adsorbed in the immobile environment on normal-phase silica must also decrease, leading to an increase in the mobile fraction. Further derivatization of the surface by physisorption of undecanol yields a further increase in both the diffusion coefficient and mobile fraction. These changes are explained below in terms of the microenvironments present. The measured diffusioncoefficientis a weighted average of all diffusion coefficients for each microenvironment present that undergo fast exchange on the time scale of the experiment (eq 1):
-I
'
I
Figure 1. Typical FPR plot for naphthacene adsorbed on C18 silica further derivatized by coadsorption of undecanol. The surface concentration of naphthacene is 0.1 % . Experimental details: 75 ms/channel, 512 channels, 120-me bleach pulse, 2 0 4 postrecoverypause. The recovered diffusion coefficient is 2.7 X 104 cm2 s-1, and the mobile fraction is 0.79. Table I. Measured Diffusion Parameters for Naphthacene on Chemically Derivatized Silica Gels X, Ne N'd surface0 methodb D/10-10 cm2 5-1 2.3 f 0.d 0.37 f 0.04 107 4 silicae sub 5.03 f 1.61 0.24 f 0.06 C8 eoln 13 1 6.65 f 0 . 4 1 0.50k 0.01 214 6 C18 sub 6.62 f 0 . 5 0 0.62 f 0.01 121 3 C18 soln 0.87 f 0.01 117 4 undec 25.3 f 1.8 C18 26.5 f 6.1 0.75 f 0.05 silica ma 15 1 Type of silica surface used. C8 is silicaderivatized with an eightcarbon-chainagent,C18 withan eighteen-carbon-chainagent. Preparation: soln, adsorption of naphthacene from solution; sub, adsorptionof naphthacenevia sublimation;undec, undecanol coadsorbed on C18 silica;ma, myristic acid physisorbed on dry silica. N is the number of FPR experimenta used in the averages of D and X,.d N' is the number of samples prepared in the accumulation of the FPR data. e Results for naphthacene on dry normal-phase silica gel pretreated at 25 O C from ref 12. f All errors reported as standard error of the mean at a 97-59; confidence level.
mobile (3). Changing the relative proportions of these three environments has a direct effect on the measured diffusion parameters. All measured values for the diffusion coefficientof naphthacene on purely chemisorbed derivatized silica (C8 and Cl8) in Table I are at least twice the diffusion coefficient of naphthacene on dry normal-phase silica. A further increase in the diffusion coefficient is seen when undecanol is physisorbed to the C18 surface. The diffusion coefficient of naphthacene on these surfaces is 1order of magnitude larger than the diffusion coefficient of naphthacene on dry silica gel. For C18 silica,the mobile fraction is larger than that of naphthacene on dry silica pretreated
fi represents the fraction of time a molecule resides in a particular environment in which the diffusion coefficient is Di.20*21Any microenvironment in which naphthacene is immobile does not belong in this category. Not only will the diffusion coefficient in these regions be much smaller than in the mobile regions, but naphthacene in these microenvironmentswould recover independentlyand should be seen as a distinct population. This would be manifest as a slow component relative to the recovery of the other components in the recovery profile. To implement eq 1, we need only to consider two of the microenvironments stated above. These are the two in which naphthacene is mobile: that on normal-phase silica (microenvironment l)and that on alkylated silica (microenvironment 3). An estimate of the value of f1 (or f 3 ) is derived from the values of the mobile fraction for each region. On the basis of the measured diffusion coefficients for naphthacene on C18 silica and normal-phase silica, an upper limit for the diffusion coefficient of naphthacene in a completely derivatized environment (D3) is estimated to be 3.1 X 104 cm2 s-1. The lower limit for D3 is given by the measured value for naphthacene on either C18 silica with undecanol coadsorbed or myristic acid derivatized silica (2.6 X 10-9cm28-l). Further analysis using the upper limit of D3 allows for an estimate of the fraction of naphthacene molecules in the alkyl phase (f3) when adsorbed to C18silica. This value is calculated to be 14% ,indicating that the coating is incompleteon the chemisorbed samples. On C18 silica with undecanol coadsorbed, we estimate the coating to be at least 84% complete. The silylation reaction with model compounds in solution has been examined by Newman and Feher.22They have found the reaction is only 1 ?6 complete after 15 min when using similar conditions used for the derivatization of silica gel. The reaction does not go to completion for model compounds either in solution or on the silica gel surface. This is validated by the experimental differences (20) Elson, E. L.; Reidler, J. A. J. Supramol. Struct. 1979, 12, 481. (21) ONeill,L. J.; Miller, J. G.; Petersen,N. 0.Biochemistry 1986,25, 177. (22) Newman, F. J.; Feher, D. A. J.Am. Chem. SOC.1990,112,1931.
Bjarneson and Petersen
2824 Langmuir, Vol. 7, No. 11, 1991 in the values of D for naphthacene on C18 silica versus C18 silica with undecanol coadsorbed. Roughly 14% of the silica surface area is coated, indicating the reaction may be more efficienton the surface than in dilute solution. The differences observed in the diffusionparameters for C18 silicaand myristic acid coated silica (Table I) indicate that it is 14% of the total accessible surface area which is derivatized. Since the two derivatizing agents are roughly the same size, each agent will have access to the same initial surface area (Le., no effects due to differences in accessibility to pores will occur). Derivatization by physisorption of myristic acid leads to a more uniform and more complete coating than chemisorption of octadecyltrichlorosilane, as reflected in the larger diffusion parameters. A sample of silica is derivatized under the same conditions, with a reaction time of 24 h. Diffusion measurements performed on this surface gave the same results for D and X, as those of naphthacene adsorbed on silica derivatized for 15min. The reaction apparently proceeds quickly a t the start, but is then retarded, perhaps due to steric hindrance on the surface or due to other geometric factors of the surface. The increase in D for naphthacene on C18-coated silica relative to that on normal-phase silica may be explained by the mechanism of binding of naphthacene to the two surfaces. On normal-phase silica, the binding is achieved by a few, relatively strong, interactions between silanol groups and the ?r electrons of the naphthacene molecule. On alkyl-coated silica, the individual binding interactions are weaker, but larger in number. Further increases in D with physisorption of undecanol on C18 silica are explained in terms of the microviscosity of the system (eq 2):
0.0
5.0
10.0
15.0
20.0
0.75
1
so ti h C 0
5 F L G.
30
G 0
u 3,
0 0
n, 0.25
io
Xm
Figure 2. Frequency of occurrence of diffusion Coefficients (A) and mobile fractions (B)within 5 % (D)and 10% (X,)of their range. The histograms show distributions of D and X, for naphthacene on C18 silica gel. Adsorption of naphthacene via sublimation is indicated by open bars, and adsorption of naphthacene from cyclohexanesolution is indicated by diagonal bars.
D = kT/f (2) k is Boltzmann's constant, Tis the absolute temperature, and f is the frictional coefficient which is a function of the viscosity of the system and the molecular size and shape. The addition of undecanol provides for a more uniform and less viscous coating, leading to an increase in D. In each system investigated a single value of D or X, is not obtained. In fact, broad distributions of values are recovered in all cases (Figures 2 and 3). This indicates that there is heterogeneity in the way naphthacene is experiencing the microenvironments a t the micrometer level of resolution. This supports the concept that chemisorption of these reagents is heterogeneous, as suggested by Lochmuller.24 Even with coadsorption of undecanol, a broad distribution is seen. The surface still does not mimic a homogeneous medium. The widths of these distributions reflect real surface heterogeneity. Diffusion measurements of both naphthacene and rubrene in isotropic and therefore homogeneous media (poly(propy1ene glycol) at 25 "C) and anisotropic but still homogeneous media dimyristoylphosphatidylcholine vesicles at 29 OC) yield, reproducibly, a narrow range for D and X,.23 The values of D are quite different for each system, while X m tends to unity in these cases. Due to inherent uncertainties in the data, distributions of diffusion coefficients and mobile fractions are expected. For simulated data with a high level of noise (10% Gaussian noise), T D (and therefore D) can be determined with a precision of 21 5% ,and X m to a precision
of 7%.24 These should represent upper limits to the standard deviations for the diffusion parameters. Typically, the standard deviations obtained in this study for D are 40-4576, and 10-30% for X,. The observed standard deviations are larger than the standard deviations obtained for individual measurements. The widths of the distributions are due to something more than just random noise alone, and are attributed to heterogeneities in both the surface structure and probe concentration. The microenvironments experienced by naphthacene are present in domains smaller than the beam width. This allows for sampling of many environments simultaneously, and leads to continuous distributions of D and Xm. If the domains were larger than the beam width, only one environment would be probed by the beam at one time. For the case considered earlier with only two microenvironments (normal-phase silica and completely coated alkyl silica), two discrete distributions for D would be expected. One distribution would have a mean value of 2.4 X 10-10 cm2s-l,the average diffusion coefficient for naphthacene on normal-phase silica. The other distribution would have cm2 s-l, the calculated value for a mean near 3.1 X naphthacene diffusing in a completely hydrocarbon coated layer. Two discrete distributions of X, would also be recovered. In all cases, single distributions in both D and X, are recovered, indicating that the microenvironment domains are present in regions smaller than the beam width. Changes in X m are observed on changing the method of preparation. Adsorption from solution onto C18 silica
(23) Balcom, B. J. Ph.D. Dissertation, The University of Western Ontario, London, Ontario, Canada, 1990.
(24)Petersen, N. 0.; Felder, S.; Elson, E. L. In Handbook of Experimental Immunology; Weir, D. M., Ed.; Blackwell Scientific Publications: Oxford, 1986; Vol. 3, Chapter 24.
~
~
~~~
Langmuir, Vol. 7, No. 11, 1991 2826
Direct Diffusion Measurements of Naphthacene
0.0
12.5
25.0
37.5
50.0
0.00
0.25
0.50
0.75
1 .oo
Xm
Figure 3. Frequency of occurrence of diffusion coefficients (A) and mobile fractions (B) within 10%of their range for adsorption of naphthacenefrom cyclohexanesolution. The histogramsshow distributionsof D and X, for naphthaceneon C18 silicagel further derivatized by coadsorption of undecanol.
yields a much larger Xm than for a sample prepared by sublimation. This is seen even though D remains constant (Figure 2). The deposition process is very different in the two cases. By allowing the surface to come into contact with a wetting solvent,the hydrocarbon chains can become untangled, and more orderede2As the solvent is removed, the chains will eventually settle back to the surface. As the chains settle, more of the probe molecules are trapped in the hydrocarbon layer than if the probe were deposited by sublimation. Molecules in the hydrocarbon layer are mobile, and hence an increase in X, is observed. As well, the formation of immobile microcrystalsis more prevalent when adsorption via sublimation is employed. Some measurements with rubrene as the probe have been performed using C18 silica and C18 silica with undecanol coadsorbed. Frequently, the fit to the recovery for rubrene on C18 silica was of poor quality. As well, the extent of recovery was typically low. Qualitatively, the mobility of rubrene on C18 silica is much lower than the mobility of naphthacene, and the mobile fraction for rubrene on C18 silica is estimated to be 0.2. On C18 silica with coadsorbed undecanol, rubrene is far more mobile than on the C l 8 surface. From two samples with rubrene on C18 silica with undecanol coadsorbed, the diffusion coefficient is estimated to be (4.9 & 0.6) X 10-10 cm2 s-l, and the mobile fraction is (0.61 0.09). The mobility is lower than that of naphthacene on a similar surface. This is an obvious effect of the four phenyl appendages which rubrene possesses over naphthacene. Rubrene is not planar, as naphthacene is, and also has a greater surface area. These factors increase the frictional coefficient,and decrease the diffusion coefficient (eq 2). Amobile fraction of 0.61 is explained by the combined effects of microcrystalline rubrene on the surface and a heterogeneous alkyl coating.
*
The results presented have a number of implications. First, we see that the largest diffusion coefficientobtained for naphthacene on chemically modified silica gels is 2 orders of magnitude smaller than the calculated value for pyrene on reverse-phase supports in the presence of a wetting solvent.6 This difference is due to the absence of a solvent in our system. It is clear, however, that naphthacene as well as the more geometrically intricate rubrene are mobile on these supports in the absence of any solvent. Bogar et al.5 report that pyrene is immobile on similar supports in the absence of solvent due to the lack of evidence for dynamic pyrene excimer formation. In fact, pyrene is probably mobile, but does not exhibit excimer formation for one of two reasons. Either the rate of excimer formation is too slow to be detected on a nanosecond time scale, or the interactions between pyrene molecules are not face to face which is required for excimer formation. It must also be noted that the FPR measurements deal with micrometer distances or over many colloid particles which make up a silica bead, while excimer formation proceeds over distances on the order of 10 A or within one microenvironment. For this reason alone the diffusion coefficient recovered from FPR should be smaller than that estimated from photophysical investigations. The fluorescence microscopy techniques employed here are far more suitable for measuring the mobility of adsorbed species than are photophysical investigations. We calculate the silylation reaction to be only 14% complete for C18 silica. This is an approximation, but it shows that the surface is not completely coated. The experiments performed here are done in vacuo and in the absence of any solvent. Hence, the results cannot be directly related to any form of liquid chromatography. However, the results show that PAHs adsorbed on reversephase supports are mobile even though the alkyl coating is incomplete and no solvent is present. Summary Naphthacene is mobile on reverse-phase chromatography supports, even in the absence of a solvent. This proceas has been previously discounted for pyrene on derivatized silica in the absence of a s01vent.~The mobility of naphthacene on alkylated silica gels is greater than that of naphthacene on normal-phase silica by up to 1 order of magnitude. An increase in the mobile fraction is also observed. Variations are observed in the diffusion parameters with the extent of derivatization and method of adsorption. As the degree of derivatization increases, both the diffusion coefficient and the mobile fraction increase. This is due to the surface coating being more homogeneous and also more fluid. Independent of deposition method, the same value for the diffusion coefficient is recovered for naphthacene on C18 silica. However, a larger value for the mobile fraction is observed for naphthacene adsorbed from solution. A more homogeneous distribution of naphthacene is also observed for this method. With fewer microcrystals present, the mobile fraction is expected to be larger. The largest values of D and Xm are observed for naphthacene adsorbed on a surface with at least a physisorbed derivatizing agent. This is a result of the greater homogeneity of the alkyl coating. Physisorption of long-chain alcohols or fatty acids provides the most homogeneously coated surface, and the largest values of D and X,. However, this is not to say that the coating produces a homogeneous surface. If this were so, one may expect the
2826 Langmuir, Vol. 7,No.11,1991 distributions obtained to be as narrow as those obtained for PAHs in isotropic media. Since broad distributions are obtained for the diffusion parameters, even for these highly derivatized surfaces, the surface is still heterogeneous particularly at the micrometer level. Numerous microenvironments are experienced by naphthacene on these surfaces also. The larger values of D and X, reflect a decrease in the microviscosity and a more homogeneous coating of the surface relative to all other surfaces. With shorter hydrocarbon chains (C1and C8),a very heterogeneous distribution of naphthacene is achieved. The shorter chains do not allow for many attractive forces between the chains and naphthacene. Some results for a
Bjarneson and Petersen
C8 sample reveal a diffusion coefficient for naphthacene comparable to that on C18silica,but a much smaller mobile fraction. The mobile fraction is small due to the large number of immobile microcrystals present. This mobile fraction is smaller than that for naphthacene on normalphase silica. Significant mobility of rubrene is found on'highly derivatized silica gels. The diffusion coefficient is found to be 5 times smaller than the diffusion coefficient for naphthacene on the same surface. The mobile fraction for rubrene is also somewhat smaller than the mobile fraction of naphthacene. The lower mobility parameters are a result of the more complex geometric structure of rubrene.
