Electrochemiluminescence of Ruthenium(II)Tris(bipyridine) Encapsulated in Sol-Gel Glasses

Maryanne M. Collinson,* Brian Novak, Skylar A. Martin,† and Jacob S. Taussig

Department of Chemistry, 111 Willard Hall, Kansas State University, Manhattan, Kansas 66506-3701

The electrogenerated chemiluminescence (ECL) of Ru-(bpy)3

2+ and tripropylamine, tributylamine, triethylamine,trimethylamine, or sodium oxalate encapsulated withinsol-gel-derived silica monoliths have been investigatedusing an immobilized ultramicroelectrode assembly. Themajor purpose of this study was to investigate the role ofthe reductant on the magnitude and stability of the ECLin this solid host matrix. For gel-entrapped Ru(bpy)3

2+/tertiary amines, the shape and intensity of the ECL-potential curves were highly dependent on scan rate. At10 mV/s, the ECL intensity was ca. 6-fold higher relativeto that observed at 500 mV/s. When the ECL acquired atlow scan rates was normalized by that obtained in solutionunder similar conditions, a value of 0.03-0.06 wasobtained. In direct contrast, the ECL of the Ru(bpy)3

2+-oxalate system showed little dependence on scan rate, andthe ECL was ca. 65-75% of that measured in solution.These differences can be attributed to differences inrotational and translational mobility between the reduc-tants (amines vs oxalate) trapped in this porous solid host.For both systems, the ECL was found to be stable uponcontinuous oxidation or upon drying the gels in a high-humidity environment for over 10 days.

Electrogenerated chemiluminescence (ECL), the productionof light from electrochemically generated reagents,1 has receivedconsiderable attention during the past several decades due to itsversatility, simplified optical setup, and good temporal and spatialcontrol. Applications in the general areas of chemical sensing,2-4

imaging,5-7 lasing,8 and optical studies9 and as detectors forchromatography10,11 have been widely reported. The entrapmentor covalent attachment of ECL precursors in a polymer host

structure has also been investigated as a means to reduce theconsumption of reagents, probe the nature of charge and masstransfer within polymers, and develop solid-state electrochemicaldevices for imaging applications.12-23 ECL, for example, has beenobserved from polymer films of diphenylanthracene and polyphen-ylenevinylene derivatives,17,18 electropolymerized films of ruthe-nium tris(bipyridine) derivatives,16,19-20 and ruthenium derivativestrapped in an ionic polymer or bound to a polymer framework.15,21-23

Sol-gel-derived materials prepared by the hydrolysis andcondensation of alkoxysilanes (i.e., tetramethoxysilane) are anattractive alternative to many polymeric materials for solid-stateapplications due to the ease at which they can be prepared,modified, and doped with various reagents.24-27 In their hydratedstate, sol-gel-derived materials can be quite porous, thus provid-ing a good matrix to entrap chemiluminescent reagents that mustdiffuse together to react.28 Recently we have shown the ECL canbe observed from ruthenium(II) tris(bipyridine), Ru(bpy)3

2+,trapped in a silica gel host using gel-entrapped tripropylamine(TPA) as the reductant.29,30 The Ru(bpy)3

2+-TPA system waschosen because it has been well studied in solution and shown togive rise to ca. 10-fold higher ECL compared to other commonlyused reductants such as oxalate.31,32 Relative to that obtained in

† Current address: Department of Chemistry, Truman State University,Kirksville, MO 63501.(1) Faulkner, L. R.; Bard, A. J. In Electroanalytical Chemistry, Bard, A. J., Ed.;

Marcel Dekker: New York, 1977; Vol 10, pp 1-95.(2) Knight, A. W. Trends Anal. Chem. 1999, 18, 47-62 and references therein.(3) Lee, W.-Y. Mikrochim. Acta 1997, 127, 19-39 and references therein.(4) Gerardi, R. D.; Barnett, N. W.; Lewis, S. W. Anal. Chim. Acta 1999, 378,

1-41 and references therein.(5) Wightman, R. M.; Curtis, C. L.; Flowers, P. A.; Maus, R. G.; McDonald, E.

M. J. Phys. Chem. 1998, 102, 49, 9991-9996.(6) Engstrom, R. C.; Pharr, C. M.; Koppang, M. D. J. Electroanal. Chem. 1987,

221, 251-255.(7) Engstrom, R. C.; Johnson, K. W.; DesJarlais, S. Anal. Chem. 1987, 59, 670-

73.(8) Kozlov, V. G.; Bulovic, V.; Burrows, P. E.; Forrest, S. R. Nature 1997, 389,

362-64.(9) Fan, F.-R. F., Cliffel, D.; Bard, A. J. Anal. Chem. 1998, 70, 2941-2948.

(10) Noffsinger, J. B.; Danielson, N. D. J. Chromatogr. 1987, 387, 520-24.(11) Lee, W.-Y., Nieman, T. A. J. Chromatogr. 1994, 659, 111-118.(12) Rao, N. M.; Hool, K.; Nieman, T. A. Anal. Chim. Acta 1992, 266, 279-286.(13) Obeng, Y. S.; Bard, A. J. Langmuir 1991, 7, 195-201.(14) Miller, C. J.; McCord, P.; Bard, A. J. Langmuir 1991, 7, 2781-2787.(15) Rubinstein, I.; Bard, A. J. J. Am. Chem. Soc. 1980, 102, 6642-44.(16) Abruna, H. D.; Bard, A. J. J. Am. Chem. Soc. 1982, 104, 2641-2642.(17) Fan, F. R. F.; Mau, A.; Bard, A. J. Chem. Phys. Lett. 1985, 116, 400-404.(18) Richter, M. M.; Fan, F.-R. F.; Klavetter, F.; Heeger, A. J.; Bard. A. J. Chem.

Phys. Lett. 1994, 226, 115-120.(19) Maness, K. M.; Terrill, R. H.; Meyer, T. J.; Murray, R. W.; Wightman, R. M.

J. Am. Chem. Soc. 1996, 118, 10609-10616.(20) Elliott, C. M.; Pichot, F.; Bloom, C. J.; Rider, L. S. J. Am. Chem. Soc. 1998,

120, 6781-6784.(21) Downey, T. M.; Nieman, T. A. Anal. Chem. 1992, 64, 261-268.(22) Lyons, C. H.; Abbas, E. D.; Lee, J.-K., Rubner, M. F. J. Am. Chem. Soc. 1998,

120, 12100-12107.(23) Wu, A.; Lee, J.; Rubner, M. F. Thin Solid Films 1998, 327-9, 663-667.(24) Brinker, C. J.; Scherer, G. Sol-Gel Science; Academic Press: New York,

1989.(25) Avnir, D. Acc. Chem. Res. 1995, 28, 328-334.(26) Dave, B. C.; Dunn, B.; Valentine, J. S.; Zink, J. I. Anal. Chem. 1994, 66,

1120A-1127A.(27) Lev, O.; Tsionsky, M.; Rabinovich, L.; Glezer, V.; Sampath, S.; Pankratov,

I.; Gun, J. Anal. Chem. 1995, 67, 22A-30A.(28) Collinson, M. M.; Zambrano, P. J.; Wang, H.; Taussig, J. S. Langmuir 1999,

15, 662-668(29) Collinson, M. M.; Taussig, J.; Martin, S. A. Chem. Mater. 1999, 11, 2594-

2599.(30) Collinson, M. M.; Martin, S. A. Chem. Commun. 1999, 10, 899-900.

Anal. Chem. 2000, 72, 2914-2918

2914 Analytical Chemistry, Vol. 72, No. 13, July 1, 2000 10.1021/ac9913208 CCC: $19.00 © 2000 American Chemical SocietyPublished on Web 04/29/2000

solution, however, the ECL intensity in the solid-state host waslow. To increase the intensity of the ECL, it is first necessary tounderstand some of the factors that influence light production inthese porous solids.

In this work, we have examined the ECL of gel-entrapped Ru-(bpy)3

2+ in the presence of reductants of different sizes and chargeto investigate the role of the reductant on the magnitude andstability of the ECL in this solid host matrix. Specifically, tertiaryamines with different alkyl chain lengths (methyl, ethyl, propyl,butyl) and sodium oxalate were used to provide insight intosurface confinement and interfacial interactions in this solid hoststructure. Such interactions have been shown to be important inregulating molecular motion in silica glasses33-35 and will inevitablyinfluence the performance and usability of these materials in solid-state electrochemical applications.

A simplified reaction mechanism for the reaction of Ru(bpy)32+

with these reductants is given below. In this reaction, oxalate isoxidized to an intermediate radical anion (e.g., CO2

•) which actsas a strong reducing agent.3 For the tertiary amines, the inter-mediate is believed to be the deprotonated amine radical.2-4

These results show that the ECL in these solid host structuresstrongly depends on the nature of the entrapped reagents andthe physicochemical properties of the host material.

EXPERIMENTAL SECTIONReagents. Tetramethoxysilane (99%), tributylamine (98.5+%),

tripropylamine (98%), triethylamine (99.5%), trimethylamine (40%in water), sodium oxalate (99.5%), and ruthenium(II) tris(bipyri-dine) (98%) were purchased from Aldrich. All reagents were usedas received. Potassium hydrogen phosphate and potassium dihy-drogen phosphate were purchased from Fisher Scientific. Deion-ized water was purified to type I with a LabConco waterpurification system.

Procedures. The microdisk electrodes (r ) 13.3 µm Pt) andthe silica sols were prepared as previously described.29,30 In brief,tetramethoxysilane (TMOS) was mixed with water and hydro-chloric acid (0.1 M) and the solution stirred for several hours.An aliquot of the silica sol was combined in a 2:1 volume ratiowith the precursor (e.g., sodium oxalate) previously dissolved inphosphate buffer (either 0.3 or 0.1 M, pH 6.2) in a silanized glassvial containing a r ) 13.3 µm Pt electrode and a silver chloridecoated silver wire reference/auxiliary electrode. The sol gelledwithin a few minutes. The pH of the gels was estimated usingindicator dyes. For those gels prepared with 0.3 M, pH 6.2phosphate buffer, the pH of the sol was ca. 6. For those gelsprepared with 0.1 M, pH 6.2 buffer, the pH was estimated to beca. 5-5.5. The gel was aged and dried for typically 2-3 days under

a 60-70% relative humidity environment. The glass vial containingthe gel was placed within 4-5 cm of the photocathode of aHamamatsu 4632 photomultiplier tube and the signal amplifiedwith a EGG VT 120A fast preamplifier and collected with an EGGOrtec T-914 multichannel scaler. A BAS 50 potentiostat modifiedto trigger the multichannel scaler or a Pine AFRDE5 bipotentiostatwas used to apply the excitation waveform.29,30

RESULTS AND DISCUSSIONVoltammetry and ECL of Gel-Encapsulated Ru(bpy)3

2+.Ru(bpy)3

2+ and/or the reductant (oxalate or amine) along with ar ) 13 µm Pt working electrode and a Ag/AgCl referenceelectrode were encapsulated into the silicate host by simply addingthe reagents and the electrodes to the sol prior to gelation. Thesols gelled quickly after addition of phosphate buffer, thusentrapping the reagents into a porous, hydrated solid.

Figure 1 shows cyclic voltammograms (CVs) of gel-encapsu-lated Ru(bpy)3

2+ and Ru(bpy)32+-oxalate at a r ) 13.3 µm Pt

electrode. In the absence of oxalate, the CVs at slow sweep rates(e.g., 10 mV/s) for gel-encapsulated Ru(bpy)3

2+ are steady statein nature and superimposed on a large rising background due tosolvent oxidation. As the scan rate is increased, the voltammo-grams become more peak shaped in appearance as expected for(31) Leland, J. K.; Powell, M. J. J. Electrochem. Soc. 1990, 137, 3127-3131.

(32) Noffsinger, J. B.; Danielson, N. D. Anal. Chem. 1987, 59, 865-868.(33) Koone, N.; Shao, Y.; Zerda, T. W. J. Phys. Chem. 1995, 99, 16976-16981.(34) Sieminska, L.; Zerda, T. W. J. Phys. Chem. 1996, 100, 4591-97.(35) Watson, J.; Zerda, T. W. Appl. Spectrosc. 1991, 45, 1360-1365.

(36) Wightman, R. M.; Wipf, D. O. Voltammetry at Ultramicroelectrodes. InElectroanalytical Chemistry; Bard, A. J., Ed.; Marcel Dekker: New York,1989; Vol. 15.

Ru(bpy)32+ f Ru(bpy)3

3+ + 1e-

Ru(bpy)33+ + reducing agent f [Ru(bpy)3

2+]*

[Ru(bpy)32+]* f Ru(bpy)3

2+ + hν (610 nm)

Figure 1. Current-potential curves for gel-encapsulated Ru(bpy)32+

(a) and gel-encapsulated Ru(bpy)32+ (10 mM) and oxalate (5 mM)

(b) acquired at 10 and 500 mV/s. Corresponding emission potentialcurves are also shown for gel-encapsulated Ru(bpy)3

2+ and oxalate.The gels were prepared with 0.1 M, pH 6.2 phosphate buffer.

Analytical Chemistry, Vol. 72, No. 13, July 1, 2000 2915

an ultramicroelectrode.36 In the presence of oxalate, the oxidationcurrent for gel-entrapped Ru(bpy)3

2+ is significantly larger thanthat observed for Ru(bpy)3

2+ alone, and the half-wave potential isshifted negatively, consistent with an electrocatalytic reactionmechanism. The corresponding ECL-potential curves are alsoshown in Figure 1. The increase in luminescence is contiguouswith the oxidation of gel-encapsulated Ru(bpy)3

2+ and oxalate. Atlow sweep rates, the ECL-potential curves are steady state inappearance analogous to the current-potential curves observedat the microdisk electrodes. At higher scan rates, the forward scanin the current-potential curve becomes more peak shaped as doesthe resultant ECL due to the onset of planar diffusion.36 In theabsence of either Ru(bpy)3

2+ or oxalate, no ECL was observed.The current-potential and ECL-potential curves for Ru-

(bpy)32+ in the presence of TMA, TEA, TPA, or TBA are

considerably different. Figure 2 shows the CVs and correspondingECL-potential plots for gel-encapsulated Ru(bpy)3

2+ and Ru-(bpy)3

2+-TEA. In contrast to that observed for the Ru(bpy)32+-

oxalate system, the current for the Ru(bpy)32+-TEA system is

only slightly higher than that observed for Ru(bpy)32+ alone even

at 5 mV/s. The shape and intensity of the ECL-potential curveare also highly dependent on scan rate. At low sweep rates (5 or10 mV/s), the ECL observed on the forward scan is higher thanthat observed on the reverse scan, and it drops at large overpo-tentials. This drop is likely due to concentration polarizationresulting from the reduced diffusion of the amines in the gel (seebelow). As the scan rate is increased, the ECL observed on theforward scan becomes significantly less than that observed onthe reverse scan until no ECL is observed at all (>0.75 V/s). Thisdrop cannot be the result of an increase in ohmic drop (iR) at

the higher scan rates because under the conditions used in thiswork (ca. 30-100 mM supporting electrolyte, 5-750 mV/s), ther ) 13 µm Pt microelectrode will be essentially immune from sucheffects.36

The observed scan rate dependence of the ECL can moreclearly be seen in Figure 3, which shows a plot of the maximumECL intensity acquired from either the forward or reverse scanvs scan rate (ν) for the Ru(bpy)3

2+-oxalate and Ru(bpy)32+-amine

systems. The ECL intensity for the Ru(bpy)32+-oxalate system

was nearly invariant with scan rate up to ca. 1 V/s, whereas theECL of the Ru(bpy)3

2+-amine systems showed a significant dropat scan rates greater than 100 mV/s. The slight increase in ECLwith scan rate for the Ru(bpy)3

2+-oxalate system may be due tothe slight adsorption of oxalate on the electrode surface at lowsweep rates.37

The overall rate at which this system responds to changes inelectrode potential will depend on the chemical kinetics of theECL system as well as on the rate of diffusion in the solid host.Previous work using a stopped flow method has shown the Ru-(bpy)3

2+-TPA system to be of intermediate reaction kinetics inpH 6 buffer relative to that observed for the Ru(bpy)3

2+-oxalate(fastest) and Ru(bpy)3

2+-proline systems.38,39 The diffusion ofmolecules trapped within sol-gel-derived glasses has been shownto strongly depend on the structure of the gel as well as the sizeand charge of the molecules trapped within.28,33-35 Both surfaceinteractions and surface confinement effects are important.33-35

The translational and rotational mobility of the alkylamines andoxalate entrapped within this porous host matrix will be dissimilardue to their different sizes and charges. Specific surface interac-tions between the amine and the walls of the silicate host coupledwith their larger size could significantly inhibit their diffusion inthis host matrix.

To better evaluate the differences in ECL for the differentreagents in the solid host, a parallel set of experiments wereconducted in solution with the r ) 13 µm Pt microelectrode. Inthese experiments, a freshly prepared Ru(bpy)3

2+ stock solutionin water was mixed with the reductant dissolved in phosphate

(37) Rubinstein, I.; Bard, A. J. J. Am. Chem. Soc. 1981, 103, 5007-5013.(38) Shultz, L. L.; Stoyanoff, J. S.; Nieman, T. A. Anal. Chem. 1996, 68, 349-

354.(39) Lee, W.-Y.; Nieman, T. A. Anal. Chem. 1995, 67, 1789-96.

Figure 2. Current-potential curves for gel-encapsulated Ru(bpy)32+

(a) and gel-encapsulated Ru(bpy)32+ (10 mM) and triethylamine (TEA)

(5 mM) (b) acquired at 5 and 250 mV/s. Corresponding emissionpotential curves are also shown for gel-encapsulated Ru(bpy)32+ andTEA. The gels were prepared with 0.1 M, pH 6.2 phosphate buffer.

Figure 3. Scan rate dependence of the maximum ECL intensityfor gel-encapsulated Ru(bpy)3

2+ and (b) TPA, (3) TEA, (9), TMA,and (]) oxalate. The gels were prepared with 0.1 M, pH 6.2phosphate buffer.

2916 Analytical Chemistry, Vol. 72, No. 13, July 1, 2000

buffer (pH 6.0, 0.3 M) in a 2:1 volume ratio as described in theExperimental Section. The final concentration of Ru(bpy)3

2+ was10 mM, and that of the reductant (TPA, TMA, TBA, TEA, oroxalate) was 5 mM. In solution, the ECL of the Ru(bpy)3

2+-alkylamine systems was more than an order of magnitude largerthan that observed in the gel. In addition, the ECL intensity wasessentially invariant with scan rate, indicating that the ECLreaction kinetics are not a limiting factor under these conditions.Also, no significant drop in the ECL intensity at large overpoten-tials at low scan rates was observed as noted in the solid host(see Figure 2). For the Ru(bpy)3

2+-oxalate system, the ECLobserved in solution and the gel were very similar in terms ofboth magnitude and scan rate response.

To better compare the ECL in solution versus the gel, the ECLobtained at slow sweep rates in the gel was normalized by thatobtained in solution under similar conditions. In these experi-ments, the gels were prepared with 0.3 M, pH 6.2 phosphate bufferto increase the buffer capacity so that the pH of the gel bettermatched that in solution (estimated to be ca. 6 with indicatordyes). Figure 4 shows a plot of normalized ECL versus reductant.The ECL of the Ru(bpy)3

2+-oxalate systems is ca. 65-75% of thatmeasured in solution, whereas the ECL of Ru(bpy)3

2+-alkyl-amines is ca. 3-6%. It is apparent that ECL reaction of Ru(bpy)3

2+

with the alkylamines is more hindered in the gel compared tothat observed for oxalate. These large differences can be attributedto differences in the mobilities of the gel-entrapped reagents. Thefact that the normalized ECL for the different alkylamines is nearlythe same regardless of their sizes suggests that the surfaceinteractions with the matrix play a more important role than sizeeffects.

At pH 6, the silicate matrix will have a net negative charge(the pKa of silica is ca. 240) and a large population of unreactedsilanols. The oxalate anion being negatively charged will likelynot adsorb or otherwise interact with the pore walls and will diffusemore freely in the matrix. In previous work we have shown thatthe diffusion coefficient of a negatively charged analyte (Fe(CN)6

3-)in an aged gel is nearly the same as it is in the sol state. In contrast,ferrocenemethanol, a molecule that will likely interact with thewalls of the matrix, exhibits a diffusion coefficient that is 4-10times smaller when entrapped in an identical matrix.28 The tertiary

amines used in the current study would be able to hydrogen bondor otherwise interact with the silicate framework, reducing itstranslational and rotational mobility. Such effects have beenobserved for pyridine and acetonitrile incorporated in sol-gelglasses.41,42 In chromatography, tertiary amines are known tostrongly interact with residual silanol groups on silica-basedreversed-phase columns to cause broadened and skewed peaks.It is also possible that the amine groups could function as localbase-type catalysts for the condensation reactions between adja-cent silanol groups in the gel, thus diminishing its porosity andreducing the mobility of the entrapped reagents.43

ECL Stability. Two viable concerns in the development ofpractical ECL display materials based on sol-gel-derived solidsare the long-term stability of the ECL upon repetitive oxidationand the reusability (long-term stability) of the solid-state host. Asa silica monolith dries, the solid shrinks in size and mass andbecomes more cross-linked and less porous, potentially influencingthe reactivity of the entrapped reagents.25-27 In this work, thesetwo issues were addressed by (1) examining the ECL uponapplication of an oxidizing potential and monitoring the ECLintensity as a function of time and (2) evaluating the ECL as afunction of the length of time the gel was dried.

Figure 5A shows the ECL of gel-encapsulated Ru(bpy)32+-

oxalate after the microelectrode potential was stepped from 0.6to 1.2 V. The ECL increased rapidly (less than 1 s) after application

(40) Iler, R. K. The Chemistry of Silica; Wiley: New York, 1979.

(41) Nikiel, L.; Hopkins, B.; Zerda, T. W. J. Phys. Chem. 1990, 94, 7558-7464.(42) Nikiel, L.; Zerda, T. W. J. Phys. Chem. 1991, 95, 4063-69.(43) Harris, T. M.; Knobbe, E. T. J. Mater. Sci. Lett. 1996, 15, 132-133.

Figure 4. Normalized ECL versus reductant. The ECL intensityobtained at low sweep rates in the gel was normalized with thatobtained in solution under similar conditions. The gels and solutionscontain 10 mM Ru(bpy)3

2+ and 5 mM reductant, pH 6. Error bars arestandard deviations from 3-5 experiments.

Figure 5. (A) Intensity-time curve following a potential step to 1.2V at t ) 30 s for gel-encapsulated Ru(bpy)3

2+ and oxalate. (B) ECLintensity as a function of drying time for gels that were reused duringthe two-week period (b) and gels that were used only once and thendiscarded ([). Error bars are standard deviations from three gels.Gels contain 10 mM Ru(bpy)3

2+ and 5 mM oxalate and were preparedwith 0.1 M, pH 6 phosphate buffer.

Analytical Chemistry, Vol. 72, No. 13, July 1, 2000 2917

of 1.2 V, dropped slightly, and reached a near-steady-state valuefor over 2 h. As reported previously, the ECL was also found tobe stable for the Ru(bpy)3

2+-TPA system with less than a 10-20% drop in intensity observed for over 24 h.29 The enhancedstability of the ECL can be attributed to the small size of theultramicroelectrode, which results in a decreased consumptionof reductant, and also to the steady-state flux of reagents to theelectrode surface.29,36 The initial drop in ECL for the Ru(bpy)3

2+-oxalate system observed immediately after application of theelectrode potential may reflect adsorption of oxalate on theplatinum electrode surface.37 This drop has not been observedfor the Ru(bpy)3

2+-amine systems. The time it takes to reachthe maximum ECL intensity is much faster for the Ru(bpy)3

2+-oxalate system than for the Ru(bpy)3

2+-amine system, consistentwith that observed in the scan rate dependence data shown earlier.

The ECL was also examined as the gels were dried for differentlengths of time under ca. 60-70% relative humidity to evaluatewhat effect changes in the structure of the gel will have on theECL intensity. Over the two-week period, the gels lose ca. 10-12% in mass. The results obtained from the Ru(bpy)3

2+-oxalatesystem are shown in Figure 5B. In this figure, the diamondsymbols represent the average ECL acquired from three differentgels run only once and then discarded, whereas the circlesrepresent the average ECL acquired from three gels continuouslyexamined over a two-week period. For either experiment, nosignificant difference in the ECL was observed when the gels weredried for 1 day versus 9 days. After ca. 10 days, the gels that werehandled on a regular basis showed a significant drop in the ECLintensity. This drop may be due to a disruption in the gel-electrode interface. Because the inorganic gel is macroscopicallyrigid, the gel can crack and/or pull away from the electrodesurface during drying (and especially during handling), thuscreating an air-solution gap at the electrode-gel interface thatcan support voltammetry.28 For the gels that were not handled,no significant change in the ECL was noted until the gels wereapproximately 15 days old, when two of the three electrodesexamined gave an ECL intensity lower than average and consistentwith the gels that were constantly handled. Similar results were

also observed for the Ru(bpy)32+-TPA system. It should be

possible to extend the stability of the ECL for a longer length oftime by sealing the vials completely so the gel does not dry out.

CONCLUSIONSChemiluminescent precursors can be trapped in sol-gel-

derived solids and electrochemically excited using an immobilizedelectrode assembly. The ECL thus produced is very stable andcan be used to probe diffusion in constrained environments andassess surface interactions between the entrapped reagents andthe walls of the silicate host. When Ru(bpy)3

2+ is trapped into thesilicate host with alkylamines such as tripropylamine or trimethyl-amine, the ECL intensity is considerably lower than that observedin solution. The ECL is also highly scan rate dependent. Incontrast, when sodium oxalate is encapsulated with Ru(bpy)3

2+

in the silica gel host, the ECL is similar to that measured insolution and no scan rate dependence is observed. These differ-ences can be attributed to differences in rotational and translationalmobility between the reductants trapped in this porous solid host.Surface interactions between the alkylamines and the surface ofthe silicates and/or local changes in the degree of condensationin the gel significantly reduce the mobility of gel-entrapped aminesrelative to oxalate. These results suggest that it may be possibleto increase the ECL intensity and improve the response time ofthe device via modification of the chemical and physical microen-vironment of the silicate host. Future studies will be directed inthis manner.

ACKNOWLEDGMENTWe gratefully acknowledge support of this work by the

National Science Foundation through the NSF CAREER program(CHE) and Research Experiences for Undergraduates program(Grant CHE-9732103) and by the Office of Naval Research.

Received for review November 17, 1999. Accepted March8, 2000.

AC9913208

2918 Analytical Chemistry, Vol. 72, No. 13, July 1, 2000