Eects of central metal ion (Mg, Zn) and solvent on singlet excited-state energy flow in porphyrin-based nanostructures Feirong Li,a Steve Gentemann,b William A. Kalsbeck,c Jyoti Seth,c Jonathan S. Lindsey,*a Dewey Holten*b and David F. Bocian*c aDepartment of Chemistry, North Carolina State University, Raleigh, NC 27695–8204, USA bDepartment of Chemistry, Washington University, St. L ouis, MO 63130–4899, USA cDepartment of Chemistry, University of California, Riverside, CA 92521–0403, USA Zinc porphyrins have been widely used as surrogates for chlorophyll (which contains magnesium) in photosynthetic model systems and molecular photonic devices.In order to compare the photodynamic behaviour of Mg- and Zn-porphyrins, dimeric and starshaped pentameric arrays comprised of free-base (Fb) and Mg- or Zn-porphyrins with intervening diarylethyne linkers have been prepared.A modular building block approach is used to couple ethynyl- or iodo-substituted porphyrins in defined metallation states (Fb, Mg or Zn) via a Pd-catalysed reaction in 2–6 h. The resulting arrays are purified in 45–80% overall yields by combinations of size exclusion chromatography and adsorption chromatography (95% purity).High solubility of the arrays in organic solvents facilitates chemical and spectroscopic characterization. The star-shaped Mg4Fb- and Zn4Fb-pentamers, where the Fb-porphyrin is at the core of the array, have pairwise interactions similar to those of dimeric MgFb- and ZnFb-arrays. The arrays have been investigated by static and time-resolved absorption and fluorescence spectroscopy, as well as resonance Raman spectroscopy.The major findings include the following. (1) The rate of singlet excited-state energy transfer from the Mg-porphyrin to the Fb-porphyrin [(31 ps)-1] is comparable to that from the Zn-porphyrin to the Fb-porphyrin [(26 ps)-1] in the dimeric arrays. Qualitatively similar results are obtained for the star-shaped pentamers. The similar rates of energy transfer for the Mgand Zn-containing arrays are attributed to the fact that the electronic coupling between the metalloporphyrin and Fb-porphyrin is approximately the same for Mg- vs.Zn-containing arrays. (2) The quantum yield of energy transfer is slightly higher in the Mgarrays (99.7%) than in the Zn-arrays (99.0%) due to the longer inherent lifetime of Mg-porphyrins (10 ns) compared with Znporphyrins (2.5 ns).(3) The rate of energy transfer and the magnitude of the electronic coupling are essentially independent of the solvent polarity and the coordination geometry of the metalloporphyrin (four- or five-coordinate for Zn-porphyrins, five- or sixcoordinate for Mg-porphyrins). (4) Polar solvents diminish the fluorescence yield and lifetime of the excited Fb-porphyrin in arrays containing either Mg- or Zn-porphyrins. These eects are attributed to charge-transfer quenching of the Fb-porphyrin by the adjacent metalloporphyrin rather than to changes in electronic coupling. The magnitude of the diminution is greater for the Mg-containing arrays, which is due to the greater driving force for charge separation.(5) The Zn-containing arrays are quite robust while the Mg-containing arrays are slightly labile toward demetallation and photooxidation. Taken together, these results indicate that porphyrin-based nanostructures having high energy-transfer eciencies can be constructed from either Mg- or Znporphyrins. However, Mg-containing arrays may be superior in situations where a succession of energy-transfer steps occurs (due to a slightly higher yield per step) or where charge transfer is a desirable feature.On the other hand, Zn-porphyrins are better suited when it is desirable to avoid charge transfer quenching reactions. Accordingly, the merits of constructing a device from Mgvs. Zn-containing porphyrins will be determined by the interplay of all of the above factors.The ability to construct molecular systems with well defined The natural antenna complexes absorb light and funnel the resultant energy to the reaction centres via excited-state energy three-dimensional architectures on the nanometre scale holds revolutionary potential for many disciplines, especially mate- migration processes.1 The energy migration process is extremely rapid (hopping time of ca. 0.1–1 ps per bacterio- rials chemistry where macroscopic objects can be designed and constructed with molecular-level precision. Nanostructures chlorophyll)2 and has a quantum eciency of nearly unity. Creating synthetic mimics of the natural antenna complexes designed for manipulation of optical phenomena are of particular interest for a variety of applications that are not possible has been a major objective of the field of artificial photosynthesis.More recently, such synthetic mimics have been tailored with bulk materials. Some examples include the following. (1) Light-harvesting nanostructures can be used as energy funnels to serve as molecular photonic devices in materials chemistry. The versatile optical (absorption and emission), redox, and with applications in solar energy or as energy sources to power molecular devices.(2) Molecular photonic wires and gates can photochemical properties of the porphyrins makes them ideally suited as components of nanostructures with optical features be used to transmit and manipulate signals in nanoscale information processing systems. (3) Structured composites of in the visible or near-IR spectral regions.Towards this goal, we have developed a modular building block synthesis of absorbers and emitters can serve as nanoscale optical sources or nanoscale imaging elements. All of these structures represent soluble multiporphyrin arrays comprised of metalloporphyrins or a composite of both metallo- and free-base porphyrins.3–11 a broad class of photonic devices whose performance can be controlled in the nanoscale regime.This building block approach has been used to construct a variety of molecular architectures containing from two to nine A major source of inspiration for the design and synthesis of optical nanostructures derives from the light-harvesting porphyrin constituents. The ability to construct successively more complicated, soluble molecular structures in a systematic antenna complexes of natural photosynthetic systems.The antenna complexes are comprised of a large number of pig- fashion has permitted us to investigate the mechanisms and factors controlling electronic communication in the synthetic ments that are arranged in a rigid three-dimensional matrix.J. Mater. Chem., 1997, 7(7), 1245–1262 1245multiporphyrin arrays, starting with the fundamental pairwise are structurally analogous to a series of previously studied Znporphyrin systems.4,6,12 The dimers and star-shaped pentamers interactions characteristic of the dimeric systems. Our previous work on dimeric, trimeric, and star-shaped pentameric arrays have dimensions along the porphyrin–diphenylethyne framework of ca. 4 and 6 nm, respectively. The availability of both has provided significant insights into the nature of these basic interactions.12–15 The information garnered from our studies sets of arrays enables a direct comparison of the synthesis, purification, chemical characterization, and spectroscopic has been used as a guide for constructing prototypical molecular photonic wires5 and optoelectronic gates10 that utilize the properties of multiporphyrin arrays comprised of Mg- vs.Znporphyrins. Inasmuch as Zn-porphyrins are four- or five- multiporphyrin motif. Other workers have also developed routes to large covalently linked porphyrin arrays. coordinate depending on the solvent, while Mg-porphyrins are five- or six-coordinate, the new arrays provide the opportunity Architectures prepared include star-shaped pentamers,16–18 linear pentamers,19 larger linear arrays up to nonamers,20 to investigate the eects of solvent and metal coordination state on the photodynamics of energy transfer. three-dimensional nonamers,18 polymeric arrays,21 and selfassembled pentamers.22 However, only a few of these routes enable precise specification of the metallation state of the Experimental various porphyrins in the array.Synthetic procedures A key structural element common to all of our multiporphyrin arrays is a diarylethyne linker that joins the constituent General. 1H NMR spectra (300 MHz, IBM FT-300, GN porphyrins at the meso-carbon atom of the porphyrin macro- 300), absorption spectra (HP 8451A, Cary 3), fluorescence cycles. The diarylethyne linkers are non-polar, establish a spectra (Spex Fluoromax) and electrochemical data12 were relatively fixed inter-porphyrin distance (ca. 20A° centre-to- collected routinely.Mass spectra were obtained by laser- centre) albeit with free rotation about the ethyne in fluid desorption mass spectrometry.11 Toluene (Fisher, certified solution,15 and enable weak electronic interactions among the ACS) and THF (Fisher, certified) were distilled from LiAlH4. porphyrins.12–14 These features of the molecular architecture CH2Cl2 (Fisher, certified ACS) was distilled from K2CO3.promote extremely ecient (ca. 99%) energy transfer which Triethylamine (Fluka, puriss) was distilled from CaH2. All predominantly involves a through-bond process mediated by reagents were obtained from Aldrich.TLC plates were pur- the diarylethyne linker.13 Our previous studies of dimeric chased from Baker (Baker-flex, aluminium oxide IB-F). arrays indicate that the rate of energy migration can be Column chromatography was performed using silica (Baker explicitly controlled by structural modification of the linker, flash silica), alumina (Fisher A540, 80–200 mesh) or various specifically, via alteration of the substituent groups on the aryl grades of deactivated alumina.Chromatography of porphyrins rings. We anticipate that the rates of energy transfer would was performed with shielding from ambient light. The isolated also be aected by other properties of the linker such as its yields of Mg- or Zn-porphyrins do not take into account any eective length and geometry.Another important design ligands on the metal ion. The Fb-porphyrin building blocks element for controlling the physico-chemical properties of the FbU, FbU¾, FbU-I, FbU-I4, and FbU-core have been prepared multiporphyrin arrays is the selection of the metal ion in the previously,3 as have the Zn-porphyrin building blocks ZnU metalloporphyrin.The metal ion modulates the redox poten- and ZnU¾ (Fig. 1).6 tial,23 conformation,24 and excited-state lifetime25 of the metalloporphyrin constituent and hence, could aect the electronic Preparation of alumina with various activities. Deactivated communication in the arrays. alumina of various grades was prepared for use in column Studies in the field of artificial photosynthesis ultimately chromatography.To a sample of alumina (Fisher, A540, require a molecular species which plays the role of (bacterio)- 80–200 mesh, grade I) in an open beaker was added deionized chlorophyll. Although (bacterio)chlorophyll contains a central water dropwise via a pipette under vigorous mechanical stir- magnesium ion, most studies in artificial photosynthesis have ring, after which stirring was continued for 1 h to ensure employedZn- rather thanMg-complexes.26 Relativelyfewmodel homogeneity. In this manner alumina with 9% w/w water, systems containing Mg-porphyrins have been prepared, and 12% w/w water, 15% w/w water (grade V), or 20% w/w water these have involved porphyrin monomers,27 dimers,28 trimers,29 was prepared.35 and larger aggregates of non-covalently linked porphyrins.30 The dearth of artificial photosynthetic systems containing Mg- Analytical size exclusion chromatography (SEC).The complexes originates mainly in historic synthetic diculties in methods employed for analytical SEC have been described in preparing Mg-porphyrins. Although Mg- and Zn-porphyrins detail.6,9 Briefly, analytical SEC columns (styrene–divinylben- have grossly similar features, Mg-porphyrins have four- to five- zene copolymer) were purchased from Hewlett Packard and fold larger fluorescence yields (Wf=0.15),31 four- to five-fold Phenomonex.Analytical SEC was performed with a Hewlett- longer fluorescence lifetimes (t=8–10 ns),31 and 100–300 mV Packard 1090 HPLC using 500 A° (300×7.8 mm), 500 A° lower oxidation potentials.10,23,32 The photochemical conse- (300×7.5 mm) and 100 A° (300×7.5 mm) columns (5 mm) in quences of these distinctions between Mg- and Zn-porphyrins, series eluting with THF (flow rate=0.8 ml min-1; void volume particularly with regards to energy- and/or electron-transfer ca. 14.4 ml). Reaction monitoring was performed by removing reactions, remain largely unexplored. Probing these distinctions ca. 5 ml aliquots from the reaction mixture, diluting with 500 ml is essential not only for understanding artificial photosynthetic toluene (Fisher, certified ACS) and injecting 50 ml to the models but also for the rational design of molecular devices HPLC. Sample detection was achieved by absorption spec- that transcend photosynthesis, such as molecular optoelectronic troscopy using a diode array detector with quantitation at gates where a Zn-porphyrin provides for energy transfer and a 420 nm (±10 nm bandwidth, reference wavelength 475 nm), Mg-porphyrin functions as a redox switch.10 which best captures the Soret bands of the porphyrins.Recently, we developed two simple methods for the preparation of Mg-porphyrins.33,34 The removal of this synthetic Magnesium 5,10,15-trimesityl-20-{4-[2- obstacle aords an opportunity for systematically exploring (trimethylsilyl) ethynyl]phenyl}porphyrin (MgU) the physico-chemical properties of these systems.In this paper, we combine the new synthesis of Mg-porphyrins with the Magnesium insertion was accomplished using the heterogeneous method.33 To a solution of FbU (220 mg, 0.263 mmol) building block synthesis of multiporphyrin arrays.These methods are used to prepare a series of dimeric and star- in 25 ml CH2Cl2 was added N,N-diisopropylethylamine (DIEA) (1.1 ml, 5.26 mmol) and MgI2 (731 mg, 2.63 mmol). shaped pentameric arrays containing Mg-porphyrins, and an identical set comprised of Zn-porphyrins (Fig. 1). These arrays The reaction mixture was stirred at room temperature. After 1246 J. Mater. Chem., 1997, 7(7), 1245–1262Fig. 1 Monomers, dimers and pentamers for spectroscopic study. Each compound has an unhindered, diphenylethyne linker, which is designated U for consistency with our previous nomenclature for related arrays.13 The Zn-porphyrins are four- or five-coordinate, while the Mg-porphyrins are five- or six-coordinate (ligands are not shown) depending on the solvent.Although this diagram portrays the porphyrins in the arrays in coplanar geometries, in fluid solution at room temperature the porphyrins rotate freely about the ethyne, and the diphenylethyne linker bends slightly.15 Replacement of the TMS-group with H in FbU, MgU, and ZnU aords FbU¾, MgU¾, and ZnU¾. 30 min the reaction was judged to be complete by fluorescence to proceed for 60 min at room temperature. The reaction mixture was diluted with 30 ml ethyl acetate, extracted with excitation spectroscopy. The reaction mixture was diluted with 30 ml CH2Cl2, washed with 10% NaHCO3 (2×50 ml), dried 5% NaHCO3 (2×50 ml) and water (2×50 ml) and then the organic layer was dried (Na2SO4 ).Column chromatography (Na2SO4 ), filtered, concentrated and chromatographed [Fisher A540 alumina, toluene–acetone (1051), 3.8×5 cm] aording [Fisher A540 alumina, toluene–acetone (1051), 3.8×5 cm] aorded 175 mg (95%). 1H NMR (CDCl3) d 1.82 (s, 12 H, 223 mg (99% yield). 1H NMR (CDCl3) d 0.36 (s, 9 H, SiCH3), 1.81 (s, 18 H, ArCH3 ), 2.61 (s, 9 H, ArCH3 ), 7.82 (AA¾BB¾, 2 ArCH3), 1.87 (s, 6 H, ArCH3), 2.60 (s, 9 H, ArCH3), 3.32 (s, 1 H, CCH), 7.82 (AA¾BB¾, 2 H, ArH), 8.16 (AA¾BB¾, 2 H, ArH), H, ArH), 8.14 (AA¾BB¾, 2 H, ArH), 8.59–8.70 (m, 8 H, bpyrrole); C58H54MgN4Si calc.av. mass 859.5u, obs. m/z 859.4; 8.59–8.72 (m, 8 H, b-pyrrole); C55H46MgN4 calc. av. mass 787.3u, obs. m/z 787.0; labs (toluene) 406(sh), 426, 566, 606 nm. labs (toluene) 406(sh), 426, 566, 606 nm.Magnesium 5,10,15-trimesityl-20-(4-ethynylphenyl )porphyrin 4-(Magnesium 5,10,15-trimesityl-20-porphinyl )-4¾-(5,10,15- trimesityl-20-porphinyl )-diphenylacetylene (MgFbU) (MgU¾) A solution of MgU (200 mg, 0.233 mmol) in 30 ml THF was The Pd-mediated coupling reaction follows the general procedure established previously.6,9 Samples of ethynyl porphyrin treated with tetrabutylammonium fluoride (TBAF) on silica (373 mg, 1.0–1.5 mmol F g-1) and the reaction was allowed MgU¾ (32.2 mg, 41 mmol) and free-base iodoporphyrin FbUJ. Mater.Chem., 1997, 7(7), 1245–1262 1247I (29.5 mg, 34 mmol) were dissolved in 15 ml toluene–triethyl- 4-(Magnesium 5,10,15-trimesityl-20-porphinyl )-4¾-(magnesium 5,10,15-trimesityl-20-porphinyl) diphenylacetylene (Mg2U) amine (TEA) (551) in a 25 ml one-neck round-bottomed flask.The flask was heated to 35°C and was fitted with a 15 cm To a solution of Fb2U (13.7 mg, 0.0091 mmol) in 1 ml CH2Cl2 reflux condenser through which a drawn glass pipette was was addedN,N-diisopropylethylamine (32 ml, 0.182 mmol) and mounted for deaeration with argon. The reaction vessel was MgI2 (25.5 mg, 0.091 mmol).33 The reaction mixturewas stirred deaerated with a high flow rate of argon for 15 min.The tip at room temperature. After 30 min the reaction was judged to of the pipette was then immersed in the solution. The argon be complete by fluorescence excitation spectroscopy. The reac- flow rate was turned down and bubbling was continued for tion mixture was diluted with 10 ml CH2Cl2, washed with another 15 min.The condenser was then elevated, leaving the 10% NaHCO3 (2×10 ml), dried (Na2SO4), filtered, concen- pipette tip in the solution, and Pd2dba3 (4.7 mg, 5.1 mmol) and trated and passed over a column [Fisher alumina, toluene– AsPh3 (12.5 mg, 41 mmol) were added to the mixture as solids acetone (1051), 3.8×5 cm] aording 14 mg (99% yield). simultaneously.The pipette was removed from the reaction 1H NMR (CDCl3) d 1.85 (s, 24 H, ArCH3), 2.38 (s, 18 H, mixture and positioned about 2 cm above the solution. The ArCH3), 2.62 (s, 12 H, ArCH3), 7.24 (s, 12 H, ArH), 8.00 argon flow rate was turned up slightly and the reaction was (AA¾BB¾, 4 H, ArH), 8.26 (AA¾BB¾, 4 H, ArH), 8.61–8.84 (m, allowed to proceed. After 2 h the reaction mixture was analysed 16 H, b-pyrrole); C108H90Mg2N8 calc.av. mass 1548.6u, obs. by analytical SEC prior to concentration under reduced press- m/z 1547.0; labs (toluene) 406(sh), 426, 526, 566, 606 nm. ure. Analytical SEC showed a trace of material eluting at the leading edge of the product (tR<24.4 min), the product (tR Mg4FbU 25.4 min), and small amounts of monomeric porphyrin materials (tR 27.5 and 28.2 min).TLC analysis [Baker alumina, In a 50 ml reaction vessel was added tetraiodoporphyrin FbUI 4 (19.9 mg, 17.8 mmol) and 20 ml toluene–triethylamine (551). toluene–acetone(1551)] showed starting material FbU-I (Rf 0.9), a non-fluorescent component (Rf 0.6), MgFbU (Rf 0.5), Sonication (Fisher sonicating bath, FS14) aorded complete dissolution of FbU-I4, after which MgU¾ (70 mg, 89.2 mmol) MgU¾ (Rf 0.3), and some baseline materials.The reaction mixture was dissolved in toluene and passed over an alumina was added. The flask was immersed in an oil-bath at 35 °C and was equipped with a reflux condenser through which a (deactivated with 15% w/w water) column (3.8×5 cm) and eluted with toluene. AsPh3 elutes rapidly, followed by a mixture drawn glass pipette was positioned for deaeration with argon. The reaction apparatus was deaerated with a high flow rate of of mobile porphyrins, and dark material including the Pdreagents remains at the top of the column.The band consisting argon for 15 min. The solution was then deaerated with the tip of the pipette immersed in the solution with gentle bubbling of MgFbU and trace amounts of high molecular mass materials and porphyrin monomers was collected.The mixture of por- of argon for 15 min. Then the condenser was elevated and Pd2dba3 (9.8 mg, 10.7 mmol) and AsPh3 (26.2 mg, 85.6 mmol) phyrins was concentrated, dissolved in toluene and loaded onto a preparative SEC column (BioRad Bio-Beads SX-1 in were added as solids simultaneously. The condenser was repositioned and argon bubbling was continued for 5 min.The THF, 4.8×60 cm, gravity-flow, 4 ml min-1). The high molecular mass material eluted first followed by the dimer MgFbU. pipette tip was then replaced about 2 cm above the solution and the argon flow rate was turned up slightly. The reaction Analytical SEC indicated that the dimer band contained trace amounts of high molecular mass material and monomeric course was monitored by analytical SEC.Aliquots (ca. 5 ml) were taken by a drawn glass capillary tube through the porphyrin. The impure dimer was then chromatographed again on alumina as described for the first column. A faint yellowish condenser in order to minimize admission of air into the reaction flask. After 2 h the reaction mixture consisted of a band eluted quickly.The second band was collected and aorded MgFbU (38 mg) in 74% yield. 1H NMR (CDCl3) d small amount of high molecular mass materials, a large amount of pentamer, small amounts of porphyrinic intermediates and -2.53 (br s, 2 H, NH), 1.87 (s, 24 H, ArCH3), 2.44 (s, 18 H, ArCH3), 2.64 (s, 12 H, ArCH3), 7.28 (s, 8 H, ArH), 7.29 (s, 4 unreacted MgU¾.Additional Pd2dba3 (9.8 mg, 10.7 mmol) and AsPh3 (26.2 mg, 85.6 mmol) were added. At 6.5 h, the reaction H, ArH), 8.02–8.08 (m, 4 H, ArH), 8.26–8.30 (m, 4H, ArH), 8.62–8.87 (m, 16 H, b-pyrrole); C108H92MgN8, calc. av. mass was judged to be complete. TLC analysis [Baker alumina, toluene–acetone (1051)] showed the starting material MgU¾ 1526.3u, obs. m/z 1526.6; labs (toluene) 430, 516, 564, 604, 650 nm.followed by a streak of fluorescent porphyrinic materials and some baseline materials. The crude mixture was concentrated 4-(Zinc 5,10,15-trimesityl-20-porphinyl )-4¾-(5,10,15-trimesityl- to dryness, dissolved in toluene–CHCl3 (352) and passed over 20-porphinyl )-diphenylacetylene (ZnFbU) a short column (3.8×5 cm) of alumina (deactivated with 15% w/w water) and eluted with toluene–CHCl3 (352). The chroma- Prepared previously.6 tography column was shielded with a black cloth.AsPh3 eluted first followed by an intense band consisting of porphyrins. 4-(5,10,15-Trimesityl-20-porphinyl )-4¾-(5,10,15-trimesityl-20- Some dark-coloured materials remained at the top of the porphinyl )-diphenylacetylene (Fb2U) column. The mixture of porphyrins was concentrated, dissolved in toluene and loaded onto a preparative SEC column (BioRad Samples of ethynyl porphyrin FbU¾ (19.9 mg, 26 mmol) and iodoporphyrin FbU-I (20.4 mg, 23.5 mmol) were coupled under Bio-Beads SX-1 in THF, 4.8×60 cm, gravity-flow, 4 ml min-1).The pentamer band was collected with small amounts of high similar conditions as described for MgFbU. Analytical SEC of the crude reaction mixture showed a trace amount of high molecular mass materials, tetrameric and trimeric porphyrin intermediates. The impure pentamer was rechromatographed molecular mass materials (tR ca. 24.0 min), the product (tR 25.5 min) and a small amount of monomeric porphyrin mate- by SEC in the same manner aording a mixture of the pentamer and trace amounts of higher molecular mass porphy- rials (tR 28.3 min).TLC [silica, toluene–hexanes (352)] showed starting material FbU-I (Rf 0.74), Fb2U (Rf 0.53) and some rinic materials, and tetrameric and trimeric porphyrins. This mixture was then chromatographed on alumina (deactivated slow-moving materials (Rf <0.18). No butadiyne-linked dimer9 (Rf 0.59) was observed. Flash chromatography [silica, toluene– with 15% w/w water; column=3.8×10 cm) with elution using toluene–CHCl3(552).The third band was collected, aording hexanes (352), 3.8×10 cm] aorded the product (28 mg, 79%). 1H NMR (CDCl3) d -2.56 (s, 4 H, NH), 1.85 (s, 36 H, o- 30 mg (45%) of pentamer Mg4FbU. A final passage over the same preparative SEC column, with removal of the leading ArCH3), 2.61 (s, 18 H, p-ArCH3), 7.26 (s, 12 H, ArH), 8.04 (AA¾BB¾, 4 H, ArH), 8.26 (AA¾BB¾, 4 H, ArH), 8.65,8.78 (m, 16 edge of the band (<5% of total material) resulted in sharpening of the analytical SEC peak from FWHM=0.71 to 0.65 min.H, b-pyrrole); C108H94N8 calc. av. mass 1503.9u, obs. m/z 1502.4; labs (toluene) 424, 516, 550, 594, 650 nm. 1H NMR 500 MHz (CDCl3 ) d -2.61 (s, 1 H, NH), 1.88 (s, 24 1248 J.Mater. Chem., 1997, 7(7), 1245–1262H, o-ArCH3), 1.89 (s, 48 H, o-ArCH3), 2.65 (s, 12 H, p-ArCH3), UV chip) was used as the detector. All RR experiments were conducted at ambient temperature on samples dissolved in 2.67 (s, 24 H, p- ArCH3), 7.28 (s, 8 H, ArH), 7.31 (s, 16 H, ArH), 8.10 (AA¾BB¾, 8 H, ArH), 8.17 (AA¾BB¾, 8 H, ArH), rigorously degassed HPLC grade or spectroscopic grade solvents (toluene, 2-nitrotoluene, CH2Cl2, DMF, or THF).The 8.35–8.40 (m, 16 H, ArH), 8.65 (m, 16 H, b-pyrrole peripheral porphyrins), 8.76 (d, 8H, J=4.4 Hz, b-pyrrole peripheral por- sample concentration was typically 0.05 mM. The sample solutions were contained in spinning 5 mm quartz NMR tubes. phyrins), 8.89 (d, 8H, J=4.4 Hz, b-pyrrole peripheral porphyrins), 9.07 (s, 8 H, b-pyrrole core porphyrin); C264H206Mg4N20 Spinning was found to be essential in order to prevent photodecomposition of the Mg-porphyrins.The excitation wavelengths calc. av. mass 3755.9u, obs. m/z 3759.4; labs (toluene) 430, 522, 564, 606, 648 nm. were provided by the output of an Ar ion (Coherent Innova 400–15UV) laser. The Raman shifts were calibrated by using Zn4FbU the known values of indene, fenchone and acetonitrile.The Raman shifts are accurate to ±1 cm-1 for strong and/or Samples of FbU-I4 (19.9 mg, 17.8 mmol) and ZnU¾ (73.9 mg, isolated bands. The laser power at the sample was typically 89.2 mmol) were coupled exactly as described for Mg4FbU, 5–7 mW and the spectral resolution was ca. 2 cm-1 at a and the product distribution observed by analytical SEC was Raman shift of 1600 cm-1.identical with that of Mg4FbU. The crude mixture was concentrated to dryness, dissolved in CH2Cl2 and passed over a short Time-resolved absorption and fluorescence spectroscopy. column of silica (CH2Cl2, 3.8×5 cm) shielded with a black Fluorescence lifetimes were acquired on an apparatus having cloth. AsPh3 eluted first, followed by an intense band of a time response of ca. 0.5 ns. Samples (ca. 50 mM) in 1cm porphyrins, with dark-coloured materials left at the top of the cuvettes were degassed on a vacuum line. Excitation flashes at column. The mixture of porphyrinswas concentrated, dissolved 532 nm having a duration of 30 ps were obtained by frequency in toluene and loaded onto a preparative SEC column (BioRad doubling the output of an actively/passively mode-locked Bio-Beads SX-1 in THF, 4.8×60 cm, gravity-flow, 4 ml min-1). Nd5YAG laser operating at 7 Hz.The flashes had an energy The pentamer band was collected with small amounts of high of 0.5 mJ and were focused to 3 mm at the sample. Emission molecular mass materials, tetrameric and trimeric porphyrin at 90° from the excitation path was collected by a lens, intermediates. The impure pentamer was rechromatographed transmitted through a long-pass filter (Schott OG570 for via SEC in the same manner aording a mixture of the metalloporphyrin emission or OG630 for Fb-porphyrin emis- pentamer and trace amounts of higher molecular mass porphy- sion) and focused on a pin photodiode (Newport Research rinic materials, and tetrameric and trimeric porphyrins.This 818-BB-21 PIN). The output of the photodiode was connected mixture was then chromatographed on silica (3.8×10 cm) directly to the input of a Tektronix 7912AD transient digitizer using CH2Cl2–hexanes(352). The first band aorded 38 mg that was controlled by a personal computer. Typically 64 (55%) of pentamer Zn4FbU. A final passage over the same traces were averaged to obtain a fluorescence decay profile.preparative SEC column, with removal of the leading edge of Transient absorption data were acquired as described else- the band (<5% of total material) resulted in sharpening of where.39 Samples in 2 mm pathlength cuvettes had a concen- the analytical SEC peak from FWHM=0.68 to 0.64 min. tration of ca. 10mM for measurements in the 410–560 nm 1H NMR 500 MHz (CDCl3) d -2.60 (s, 2 H, NH), 1.90 (s, 72 region and a concentration of ca. 100 mM for measurements in H, o-ArCH3), 2.62 (s, 12 H, p-ArCH3), 2.65 (s, 24 H, p-ArCH3), the 600–750 nm region. The samples were excited with flashes 7.31 (s, 8 H, ArH), 7.33 (s, 16 H, ArH), 8.13 (AA¾BB¾, 8 H, at 582 nm having a duration of 0.2 ps. The flashes had an ArH), 8.18 (AA¾BB¾, 8 H, ArH), 8.36 (AA¾BB¾, 8 H, ArH), 8.40 energy of 100 mJ and were focused to 1.5 mm at the sample.(AA¾BB¾, 8 H, ArH), 8.75 (m, 16 H, b-pyrrole peripheral The absorption changes were probed with weak white-light porphyrins), 8.85 (d, 8 H, J=4.5 Hz, b-pyrrole peripheral (400–1000 nm) pulses also having a duration of ca. 0.2 ps. porphyrins), 8.98 (d, 8 H, J=4.5 Hz, b-pyrrole peripheral Absorption changes over a 150 nm wavelength span were porphyrins), 9.08 (s, 8 H, b-pyrrole core porphyrin); acquired using a two-dimensional detection system.For each C264H206N20Zn4 calc. av. mass 3920.2u, obs. m/z 3918.5; labs(to- spectrum, data acquired with 300 flashes were averaged, giving luene) 424, 429, 519, 551, 590, 651 nm. a resolution in DA of ±0.005.Absorption changes were obtained as a function of time by sending the probe pulse Spectroscopic methods down an optical delay line, which permitted pump–probe time Absorption and fluorescence spectroscopy. Absorption spectra delays of -300 ps to 3 ns. were collected using a Varian Cary 3 with 1 nm bandwidths and 0.25 nm data intervals. Fluorescence spectra were collected Results using a Spex Fluoromax with 1 mm slit widths (4.25 nm) and 1 nm data intervals.Emission spectra were obtained with Synthesis of the building blocks and the arrays Alexc <0.1. Quantum yields were determined by ratioing inte- Porphyrin building blocks. Monofunctionalized porphyrin grated corrected emission spectra to MgTPP (0.15),31 ZnTPP building blocks were prepared via mixed aldehyde conden- (0.030)36 or TPP (0.11)36 in toluene.Fluorescence quantum sations. Lindsey and co-workers previously prepared the free- yield measurements in other solvents were corrected for refracbase trimesityl-monoethynylporphyrin (FbU) by condensing tive index dierences relative to toluene.37 Excitation spectra 4-[2-(trimethylsilyl)ethynyl]benzaldehyde and benzaldehyde were not corrected.Measurements were made at room temwith pyrrole,3 but the separation of FbU from the mixture of perature without deaeration of samples. The solvent relative six porphyrins was dicult. A more ecient separation method permittivities (e) at room temperature for various solvents are (Scheme 1) involves separation of the Zn- rather than the Fb- as follows: toluene (2.38), ethyl acetate (6.02), tetrahydrofuran porphyrins.6 In this method, the crude reaction mixture con- (THF, 7.58), acetone (20.7), 2-nitrotoluene (27.4), acetonitrile taining six porphyrins was chromatographed to remove non- (37.5), dimethyl sulfoxide (DMSO, 46.7).38 porphyrinic materials.The mixture of six porphyrins was subjected to zinc-insertion conditions and column chromatog- Resonance Raman spectroscopy.Resonance Raman (RR) spectra were recorded with a triple spectrograph (Spex 1877) raphy of the mixture of Zn-porphyrins readily aorded ZnU. After isolation, ZnU was demetallated with trifluoroacetic acid equipped with either a 1200 or 2400 groove mm-1 holographically etched grating in the third stage. A liquid-nitrogen- (TFA) in CH2Cl2 to aord the corresponding FbU.The magnesium chelate was prepared by treating FbU with MgI2 cooled, UV-enhanced 1152X298 pixel charge coupled device (Princeton Instruments, LN/CCD equipped with an EEV1152- and N,N-diisopropylethylamine in CH2Cl2 at room tempera- J. Mater. Chem., 1997, 7(7), 1245–1262 1249procedure. The attempted separation of MgU was carried out on alumina of various activities and with eluents of various polarities.Separation of the mixture of six Mg-porphyrins was achieved on alumina TLC [Baker alumina, toluene–CHCl3 (2051)]. However, column chromatography on alumina grade I or deactivated alumina containing 9, 15 or 20% water was ineective in separating the porphyrins. Accordingly, the preferable route for forming MgU involves isolation of ZnU followed by demetallation and magnesium insertion.By this sequence, each of the metalloporphyrin building blocks (ZnU, MgU) was prepared in 200 mg quantities in ca. 2 days. Trimesitylmonoiodoporphyrin (FbU-I) was prepared via condensation of 4-iodobenzaldehyde and mesitaldehyde with pyrrole as described previously.3 In contrast to FbU, FbU-I was easily separated by column chromatography on silica.3,6 The tetraiodoporphyrin FbU-I4 was synthesized by condensation of 4-iodobenzaldehyde and pyrrole and was purified by column chromatography.3 Dimeric arrays.The key reaction for the synthesis of the arrays involves the Pd-catalysed coupling of an ethynylphenyl porphyrin and an iodophenyl porphyrin.We previously optimized this coupling reaction for the synthesis of multiporphyrin arrays containing Fb-porphyrins and Zn-porphyrins,9 and used this method to prepare ZnFbU.6 The optimization was performed in order to achieve good yields of the diarylethyne- Scheme 1 Separation of FbU from the mixed condensation reaction linked porphyrin array under mild conditions in the absence mixture of any copper cocatalysts, and to minimize formation of higher molecular mass byproducts as well as diarylbutadiyne-linked ture with stirring for 30 min.33 Column chromatography of the dimers.The optimized conditions enable the coupling to be crude reaction mixture aorded MgU in 99% yield. performed with 1.5–5 mM of each porphyrin in toluene–tri- The trimethylsilyl group of MgU or ZnU was removed by ethylamine (551) in the presence of tris(dibenzylideneacetone) treatment with TBAF on silica in THF at room temperature dipalladium(0) (Pd2dba3 ) and triphenylarsine (AsPh3) under for 60 min (Scheme 2).No demetallation of either metallopor- argon at 35°C for 2 h. The molar ratio of the components is phyrin was observed during this reaction. ZnU¾ was isolated as follows: ethyne (1.25), iodide (1), Pd2dba3 (0.15), AsPh3 by chromatography on silica.3 However, due to the slightly (1.2).To establish the compatibility of Mg-porphyrins with acidic nature of silica, chromatography of MgU¾ was performed these coupling conditions, magnesium tetraphenylporphyrin on alumina in order to avoid demetallation. Both ZnU¾ and (MgTPP) was subjected to the same coupling reaction con- MgU¾ were isolated in 95% yield.In general, we have found ditions. After 2 h, TLC and analytical SEC showed no that chromatography of the Zn-porphyrins could be performed decomposition, demetallation, or transmetallation of MgTPP. on either silica or alumina, while Mg-porphyrins generally For the synthesis of porphyrin arrays, the reaction course is require chromatography on alumina rather than silica to avoid easily monitored by analytical size exclusion chromatography demetallation. (SEC) coupled with a UV–VIS diode array detector.In pursuit of a direct route to MgU mirroring that used to Chromatograms collected periodically provide a clear indi- isolate ZnU, we converted the mixture of Fb-porphyrins into cation of product distribution over time. Mg-porphyrins using the heterogeneous magnesium insertion The reaction of ethynylporphyrin MgU¾ (2.73 mM) and iodoporphyrin FbU-I (2.27 mM) under the Pd-mediated coupling conditions (30 mol% Pd atom–iodide and 120 mol% AsPh3–iodide) was performed under argon at 35°C (Scheme 3).After 2 h, SEC analysis of the reaction mixture showed a trace amount of high molecular mass materials, dimer MgFbU, and a small amount of porphyrin monomers.TLC analysis [Baker alumina, toluene–acetone (1551)] showed FbU-I (Rf 0.9), a non-fluorescent component (Rf 0.6), MgFbU (Rf 0.5), MgU¾ (Rf 0.3), and some baseline materials. The reaction mixture was concentrated and chromatographed [toluene–acetone (1551)] on a short column of deactivated alumina (15% water), which removed the Pd-reagents and AsPh3 from the porphyrin components. Preparative size exclusion chromatography of the mixture of porphyrins aorded MgFbU and a trace amount of monomeric porphyrins.Subsequent chromatography on deactivated alumina (15% water) aorded MgFbU in 74% overall yield. Characterization of MgFbU was performed by TLC, analytical SEC, 1H NMR spectroscopy, absorption and fluorescence spectroscopy, and laser desorption mass spectrometry. In an early attempt to synthesize MgFbU, we sought to insert magnesium selectively into the core of one porphyrin unit in the all-free-base dimer, Fb2U. Fb2U was prepared via Scheme 2 Synthesis of porphyrin building blocks ZnU¾ and MgU¾ the Pd-catalysed coupling reaction of ethynyl porphyrin FbU¾ 1250 J. Mater.Chem., 1997, 7(7), 1245–1262Scheme 3 Formation of dimeric arrays MgFbU and Fb2U and iodo porphyrin FbU-I, and was isolated in 79% yield revealed that the reaction mixture consisted of pentamer, small amounts of high molecular mass material, intermediates (tetra- after one flash silica chromatography column.Upon treatment of Fb2U with MgI2 (2–8 equiv.) and DIEA (4–16 equiv.) in mer, trimer and dimer) and starting material ZnU¾.Tetraiodoporphyrin FbU-I4 was completely consumed. CH2Cl2, no metallation was observed, while exposure to 10 equiv. MgI2 and 20 equiv. DIEA aorded the completely Another portion of catalyst and ligand was added, and after an additional 3 h, the formation of Zn4FbU levelled o. The metallated Mg2U. These experiments aimed at selective metal insertion in a preformed array comprised of Fb-porphyrins product distribution during the course of reaction is shown in Fig. 2. The peaks of the dimeric, trimeric and tetrameric proved ineective in yielding the monomagnesiated dimer. In contrast, the desired MgFbU dimer is easily prepared by the materials are well separated (tR dierences>1 min). The peaks of the tetrameric, pentameric and high molecular mass mate- rational coupling of Fb-porphyrin and Mg-porphyrin building blocks.rials are clearly present but are not well resolved. By visual inspection, the pentamer-forming reaction remained homogeneous at all times. The crude reaction mixture was chromato- Star-shaped pentameric arrays. Previously we prepared two star-shaped Zn4Fb-pentamers bearing 2,6-dimethoxyphenyl graphed first on silica to remove Pd-reagents and AsPh3, then on two SEC columns which aorded the pentamer with trace units or mesityl groups at the non-linked meso-positions.This synthetic work was done prior to the investigation of optimized amounts of tetrameric and trimeric porphyrins, and finally on silica which aorded the purified Zn4FbU (55% overall yield). Pd-coupling methods and before we had refined the chromatographic separation methods for these compounds.The di- We found that the FWHM of the peak obtained by analytical SEC aorded one measure of purity. A subsequent passage of phenylethyne-linked pentamer bearing 2,6-dimethoxyphenyl groups was prepared by a coupling reaction at 100°C for 12 h Zn4FbU over the preparative SEC column led to removal of only a trace amount of the leading edge of the band, but (45.6 mg, 45% isolated yield),4,6 while a diphenylethyne-linked pentamer bearing mesityl groups was prepared by a coupling caused the peak in the analytical SEC to sharpen from 0.68 to 0.64 min.This material was used for spectroscopic studies. at 50°C for two weeks (8.8 mg, 5.5% isolated yield).40 For solubility reasons we have since focused on the mesityl- In the preparation of Mg4FbU, the coupling reaction of MgU¾ and FbU-I4 and the product distribution as determined substituted arrays.6 We now report a refined synthesis of a diphenylethyne-linked pentamer (Zn4FbU), extend this route by analytical SEC were indistinguishable from that in the synthesis of Zn4FbU.However, the purification procedure to the synthesis of aMg4FbU pentamer, and develop improved separation methods for isolating the pentamers.involved successive chromatography columns of deactivated alumina, SEC, SEC, deactivated alumina, and SEC. The devel- The precursorto the core of the pentameric arrays, tetraiodoporphyrin FbU-I4 ,has limited solubility in the couplingsolvent opment of this chromatography sequence involved examination of alumina having various degrees of deactivation, and toluene–triethylamine(551). To facilitate formation of the pentamer, we sought to keep the porphyrin concentrations as high alumina containing 12% water was found to give the best separation.The chromatography of Mg4FbU on deactivated as possible while maintaining homogeneous solutions.We found that the concentration of FbU-I4 can be raised to 0.9 mM alumina was often complicated by streaking during the prolonged elution. Nonetheless, deactivated alumina was more by dissolution with the aid of sonication. Thus, the metalloethynylporphyrin (ZnU¾ or MgU¾) and tetraiodoporphyrin eective for these arrays than sugar and other mild chromatographic media, which have traditionally been employed for FbU-I4 concentrations were kept at 4.5 mM (5 equiv.) and 0.9 mM, respectively.The Pd-mediated coupling reaction was separation of chlorophylls.41 The pentamer can be purified to a considerable extent solely by chromatography on alumina performed similarly as for the dimer syntheses (Scheme 4). In the synthesis of pentamer Zn4FbU, SEC analysis at 3 h columns, but the use of successive columns with dierent J.Mater. Chem., 1997, 7(7), 1245–1262 1251the putative hexamer peak intensity was 5% of that of the molecule ion peak. These higher mass peaks were observed in both laser desorption mass spectrometry (neat samples) and in matrix-assisted laser desorption mass spectrometry.11 We believe these peaks to be synthesis byproducts, not mass spectrometric artifacts, although such impurities were not detected by analytical SEC.Based on these mass spectral peak intensities, and the amount of residual fluorescence emanating from the metalloporphyrins in the arrays, we estimate the purity of each pentameric array to be 95%. The dimers are estimated to be 97% pure. NMR features. The porphyrin building blocks and arrays were readily characterized by 1H NMR spectroscopy at 300 or 500 MHz at room temperature in CDCl3 at ca. 5mM (monomers and dimers) or ca. 1mM (pentamers) concentration. Upon formation of the arrays, the resonances from the bpyrrole protons and from the protons flanking the ethynylunit exhibit characteristic features. For example, upon coupling of MgU¾ and FbU-I to form MgFbU, the observed splitting pattern of the b-pyrrole protons is the sum of the splitting pattern of each of the component parts.The signals from the aryl protons flanking the ethyne linkage in the dimers shift downfield by ca. 0.2 ppm compared with the monomers. The changes in splitting pattern and chemical shift are similar to those previously reported for ZnFbU.6 The pentameric arrays (Mg4FbU and Zn4FbU) exhibited similar spectral features in comparison with their respective monomeric precursors.A key diagnostic in the star-shaped pentamers is the singlet at d ca. 9 originating from the b-pyrrole protons of the core Fbporphyrin, which has four-fold symmetry (assuming rapid NMH tautomerism). The chemical shifts of the respective protons in Mg4FbU and Zn4FbU dier by <0.2 ppm.However, the peaks of Mg4FbU, particularly in the aromatic region, are slightly broader than those of Zn4FbU. The line broadening observed with Mg4FbU may be caused by the Fig. 2 Size exclusion chromatograms of the reaction forming Zn4FbU various accessible coordination states and ligands of the mag- pentamer. Top, starting materials (ZnU¾ and FbU-I4) before the nesium in the peripheral porphyrins.At higher concentration catalyst was added. Middle, crude reaction mixture after 6 h. Bottom, purified Zn4FbU pentamer. Identical chromatograms were observed (ca. 5mM), samples of Mg4FbU and Zn4FbU exhibit severe for Mg4FbU. line broadening in the spectra, a sign of aggregation. Solubility. For easy purification and characterization, high separation modalities provides the most eective purification solubility of the arrays in various solvents is essential.The procedure. The purified pentamer Mg4FbU was isolated in operational solubilities we have observed during the course of 45% overall yield. In analogy with the Zn4FbU pentamer, a handling these compounds are listed in Table 1. These are not subsequent passage of Mg4FbU over the preparative SEC necessarily upper solubility limits.In addition, the dimers and column led to removal of only a trace amount of the leading pentamers are soluble in dilute solution in a wide range of edge of the band, but caused the peak in the analytical SEC solvents. For example, analytical SEC is performed in THF to sharpen from 0.71 to 0.65 min. This material was used for (10-5–10-4 M), and absorption and fluorescence spectroscopy spectroscopic studies. has been performed at 20 mM in solvents such as ethyl acetate, acetone, acetonitrile and DMSO.Chemical characterization and physical properties of the arrays Purity. Each array (MgFbU, ZnFbU, Mg4FbU, Zn4FbU) Chemical stability. In our routine handling of Mg-porphyrinbased arrays, we did not observe any decomposition or demet- was characterized by analytical SEC, 1H NMR spectroscopy, laser desorption mass spectrometry, and absorption and fluo- allation of solid samples stored at room temperature in the dark over a period of 1–2 weeks. The Mg-porphyrin containing rescence spectroscopy.Mass spectrometry indicates that there is no demetallation or transmetallation during the conversion arrays remain intact at -5 °C for ca. 3 months. The Mgporphyrin monomers could be stored at -5 °C for longer of the porphyrin building blocks into the arrays. Discerning the presence of any impurities having lower molecular mass periods, indicating their greater stability compared with the corresponding arrays. TLC analysis provided an eective assay than the molecule ion is dicult due to fragmentation of the molecule ion.However, higher molecular mass impurities are for small amounts of decomposition of the Mg-porphyrin arrays. In particular, a fast-moving greyish-blue decomposition readily observed. In the mass spectrum of Mg4FbU a strong molecule ion was observed at m/z 3759.4, and in addition a component was observed on alumina TLC.The solution absorption spectra of such samples also exhibited slight much weaker peak was observed at m/z 4549.6. The latter is consistent with a hexamer comprised of five magnesium por- changes in the relative intensities of the Q bands. Samples of Mg-porphyrin arrays that exhibited any signs of deterioration phyrins and one free-base porphyrin. Similarly in the spectrum of the Zn4FbU pentamer, a strong molecule ion was observed were passed over a short column of deactivated alumina, which readily removed the mobile greyish-blue decomposition at m/z 3918.5 and a much weaker peak was observed at m/z 4750.0.The latter is consistent with a hexamer comprised of product(s). During synthesis and characterization, all Mgporphyrins were protected against unnecessary exposure to five zinc porphyrins and one free-base porphyrin.In each case 1252 J. Mater. Chem., 1997, 7(7), 1245–1262Scheme 4 Building block approach in the synthesis of pentamers Mg4FbU and Zn4FbU air and light. In contrast to the Mg-porphyrin arrays, the Zn- (MgTMP),33 and magnesium tetrakis(pentafluorophenyl)- porphyrin (MgPFPP)34 in CH2Cl2 with acetic acid (0.3 M) at porphyrin arrays appear to be stable indefinitely when stored in solid form at -5 °C.room temperature. After 1 h, MgTPP and MgTMP were demetallated quantitatively, while MgPFPP exhibited no Although the Mg-porphyrins and Mg-porphyrin-based arrays were suciently stable for routine handling and spectro- demetallation after 24 h. This experiment illustrates the design principle that incorporation of electron-withdrawing groups scopic characterization, thoughts about future, more robust, arrays containing Mg-porphyrins prompted us to follow up provides enhanced stability of Mg-porphyrins toward demetallation.an earlier observation concerning electronic eects on stability of Mg-porphyrinic compounds. We observed that Mgphthalocyanines are more stable toward demetallation than Mg-porphyrins,34 which can be attributed to electron with- Electrochemical properties.The redox potentials of the Znand Fb-porphyrins have been previously reported.12,14 For drawal by the four meso-nitrogens in the former compounds. To investigate whether this phenomenon carried over to Mg- these arrays the redox potentials of the constituent porphyrins are identical to those of the monomers.This is also the case porphyrins bearing electron-withdrawing groups, we treated 2 mM solutions of MgTPP, magnesium tetramesitylporphyrin for the components of MgFbU and Mg4FbU. It is noteworthy Table 1 Operational solubilities (mM) of arrays array toluene–TEA (551)a chromatography solventb toluenec CDCl3d Fb2Ue 5.8 5–8f 8 MgFbU 2.3 5–7g 10 8 Mg4FbU 1 2.5h 2.5 3 Zn4FbU 1 2.5i 2.5 5 aSolvent for Pd-mediated coupling reactions. bSolvent for adsorption chromatography (silica or deactivated alumina).cSEC loading solvent. dNMR solvent. eThe solvent for magnesium insertion is CH2Cl2 in which the solubility is 9.1 mM. fToluene–hexanes (352). gToluene-acetone (1551).hToluene–CHCl3 (351). iCH2Cl2–hexanes (352). J. Mater.Chem., 1997, 7(7), 1245–1262 1253Table 2 Absorption maxima in various solvents (298 K) ethyl toluene (FWHM) acetate THF (FWHM) acetone acetonitrile DMSO monomers TPP 419 (12.0) 415 417 (15.0) 414 413 419 548 545 546 546 547 550 MgTPP 426 (12.5) 422 429 (11.0) 422 425 426 563 562 570 562 565 564 ZnTPP 423 (11.2) 421 423 (11.2) 422 422 427 550 552 555 554 555 560 FbU 420 (12.7) 420 418 (13.5) 415 414 419 548 548 547 545 546 548 MgU 428 (12.5) 425 433 (12.7) 425 425 428 565 564 573 565 565 565 ZnU 423 (12.9) 424 425 (10.0) 424 424 430 550 554 556 555 558 562 FbU-core 424 (15.0) 420 (14.0) 555 550 dimers MgFbU 430 (19.5) 428 427 427 431 565 564 564 564 565 ZnFbU 426 (18.9) 426 426 (16.7) 426 426 431 550 553 556 554 556 562 pentamers Mg4FbU 431 (16.0) 433 (12.7) 522, 565, 604, 650 518, 574, 615, 648 Zn4FbU 424, 429 (20.0) 429 (15.0) 519, 551, 590, 651 516, 558, 597, 648 that the Mg-porphyrin is ca. 300 mV easier to oxidize than shifts observed for the Zn-porphyrins are attributed to solvent ligation which converts the zinc ion from a four- to a five- the Zn-porphyrin.10 coordinate geometry. Regardless of the solvent, the spectrum in the Q-band region of a given array closely resembles the Spectroscopic and photochemical properties of the arrays sum of the spectra of the component parts in the same solvent.Absorption spectra. The absorption spectra of various porphyrins and the arrays were measured in toluene at room temperature (Table 2). A slight bathochromic shift in the Soret Fluorescence spectra and quantum yields.The fluorescence emission spectra of the dimers were measured in toluene. band (2 nm) and Q bands (2 nm) is observed with building block MgU compared to MgTPP. In MgFbU, the Soret band Illumination of MgFbU at 648 nm, where the Fb-porphyrin absorbs about 20 times more strongly than the Mg-porphyrin, shows no splitting but the absorption band is slightly red shifted (lmax 430 nm, shoulder at 420 nm at ca. 75% height) results in typical Fb-porphyrin emission with quantum yield (Wf=0.13) nearly identical with the monomeric Fb-porphyrins, and broadened (FWHM=19.5 nm) compared with MgU (428 nm) and FbU (420 nm). However, the visible absorption FbU or TPP. Illumination of MgFbU at 565 nm, where the Mg-porphyrin absorbs about 11 times as intensely as the Fb- bands are nearly the sum of the spectra of the Fb- and Mgporphyrin components.Similarly for Mg4FbU, a broadened porphyrin, results in emission predominantly from the Fbporphyrin. The Mg-porphyrin emission yield (Wf=0.009) is and red-shifted Soret band (431 nm, FWHM=16 nm) is observed compared with the model porphyrins MgU (428 nm) diminished 17-fold compared with MgU. The fluorescence yield measurements of MgFbU are summarized in Table 3.and FbU-core (424 nm), while the visible bands of Mg4FbU (522, 565, 604 and 650 nm) are nearly a superposition of those ZnFbU, which we have characterized previously,13 has nearly identical features to MgFbU, including 20-fold quenching of the building blocks. The spectra of ZnFbU in various solvents have been described previously.6,13 The pentamer of the Zn-porphyrin compared with ZnU and dominant emission from the Fb-porphyrin upon illumination at 550 nm, Zn4FbU exhibits nearly identical absorption spectral features as Mg4FbU, though the former exhibits a very slightly split where the Zn-porphyrin absorbs 80% of the light.Measurement of the small amounts of residual metalloporphy- Soret band (424, 429 nm).The absorption spectra of selected compounds were also rin fluorescence is not a reliable means of placing a bound on the energy-transfer eciency, due to diculties in quantitation collected in several more polar solvents. For MgTPP, MgU and MgFbU, only a slight shift (1–3 nm) in the Soret band arising from spectral overlap, and the presence of fluorescent impurities at the few percent level that become significant in and the Q bands is observed in ethyl acetate, acetone, acetonitrile or DMSO compared with toluene (Table 2).In THF, comparison to the strongly quenched metalloporphyrin in the arrays.13 In particular, the small but quantifiable metallopor- however, Mg-porphyrins exhibit bathochromic shifts of ca. 8 nm and a two-fold increase in intensity of the Q(1,0) band, phyrin emission in MgFbU (in contrast to the negligible metalloporphyrin emission in ZnFbU) is likely due to impurit- giving green solutions.The absorption spectral changes of Mgporphyrins in THF are attributed to the binding of two axial ies at the 3% level. Fluorescence excitation spectra provide a better overall view of the yield of energy transfer in donor– THF molecules, yielding a six-coordinate geometry, consistent with the Raman data reported below.Mg-porphyrins are five- acceptor systems.42,43 Close matching of the fluorescence excitation spectrum and absorption spectrum through the Q bands coordinate in non-coordinating solvents, where water presumably serves as the fifth ligand. In all solvents examined here (lem=720 nm) was observed for each array (MgFbU, ZnFbU, Mg4FbU, Zn4FbU) in toluene.These results indicate a high with the exception of THF, the Mg-porphyrins are predominantly if not exclusively five-coordinate. The small spectral yield of energy transfer, as absorption by the metalloporphyrin contributes fully to the observed emission of the Fb-porphyrin. shift of the Mg-porphyrins in acetone, acetonitrile or DMSO is in contrast to their zinc counterparts, which exhibit batho- The fluorescence properties of MgFbU were examined in several more polar solvents. Again, close matching of absorp- chromic shifts of up to 10 nm in these polar solvents.The 1254 J. Mater. Chem., 1997, 7(7), 1245–1262Table 3 Fluorescence yields in various solvents (298 K) obtained with ZnFbU, as noted earlier.13 However, the magnitude of the decline in Fb-porphyrin fluorescence with increased ethyl solvent polarity was less than that with MgFbU (Fig. 3). Thus, toluene acetate acetone acetonitrile DMSO MgFbU and ZnFbU exhibit high yields of energy transfer in all solvents but the emission from the Fb-porphyrin is monomers FbU 0.12 0.13 0.13 0.13 0.15 quenched as the solvent polarity increases. The quenching of MgU 0.16 0.20 0.20 0.19 —a the excited-state Fb-porphyrin is attributed to charge transfer ZnU 0.034 0.041 0.041 0.063 0.051 with the neighbouring ground-state metalloporphyrin follow- FbU-core 0.15 ing energy transfer. arrays Matched solutions of Mg4FbU and of Zn4FbU in toluene MgFbU were illuminated at 565 nm, a wavelength where the two Fb(em)b 0.13 0.13 0.069 0.050 0.017 Mg(em)c 0.009 0.013 0.009 0.009 samples exhibit equal absorbance.The fluorescence emission ZnFbU spectrum of each sample was comprised predominantly of Fb- Fb(em)b 0.13 0.11 0.14 0.072 0.065 porphyrin emission. However, the Fb-porphyrin emission Zn(em)c 0.002 0.002 0.004 0.004 0.003 (measured in the range 625–800 nm) from Mg4FbU was 13% Mg4FbU less than that from Zn4FbU.In addition, illumination of the Fb(em)b 0.11 Fb-porphyrin at 648 nm in Mg4FbU and Zn4FbU yielded Mg(em)c 0.010 Zn4FbU Wf=0.11 and 0.14, respectively (Table 3). These results indicate Fb(em)b 0.14 a slight amount of quenching of the Fb-porphyrin emission in Zn(em)c 0.004 Mg4FbU compared with that of Zn4FbU in toluene. aAggregation prevented measurement.bThe emission from the Fb- Resonance Raman spectra. The high-frequency regions of porphyrin (lexc=648 nm) was measured in the range 660–800 nm. The the B-state excitation (lexc=457.9 nm) RR spectra of ZnU, total emission from the Fb-porphyrin (620–800 nm) was then inferred assuming the Fb-porphyrin in the array has the same emission spectral MgU, ZnFbU, and MgFbU in toluene are shown in Fig. 4 profile as the Fb-porphyrin monomer, and these values are reported (left panel). The spectra obtained in 2-nitrotoluene are also in this Table.6 cThe emission from the metalloporphyrin (lexc= shown in Fig. 4 (right panel). The key spectral features shown 550–555 nm for Zn-porphyrins and lexc=562–565 nm for Mg- in the figure are the ethyne stretching mode, nCOC,12 which is porphyrins) was measured in the range 570–800 nm.The total emission observed for the monomers at ca. 2156 cm-1 and for the from the metalloporphyrin was inferred by measurement of the dimers at ca. 2213 cm-1, and the porphyrin skeletal mode, n2, emission intensity in the 570–620 nm region and assuming the metalloporphyrin in the array has the same emission spectral profile which is observed in the region 1543–1551 cm-1 for all the as the metalloporphyrin monomer.6 compounds.Inspection of the RR data reveals that the frequencies of the tion and excitation spectra was observed in all solvents, consistent with a high yield of energy transfer. However, illumination at 648 nm gave typical Fb-porphyrin emission in all solventsbut the quantum yield of the Fb-porphyrin emission decreased steadily with increasing solvent polarity (Fig. 3). In contrast, the fluorescence yields of the Fb-porphyrin monomers (or the Mg-containing monomers) changed only slightly as a function of solvent polarity. Illumination of the Mg-porphyrin (563–565 nm) in MgFbU yields a constant high degree of quenching of the Mg-porphyrin emission in all solvents, though the relative amount of Fb-porphyrin emission declined as the solvent polarity increased.Qualitatively similar results were Fig. 4 The high-frequency regions of the B-state excitation (lexc= Fig. 3 Fluorescence quantum yield of TPP (&), and the Fb-porphyrins in MgFbU ($) and ZnFbU (+) measured by illumination at 648 nm 457.9 nm) RR spectra of ZnU, MgU, ZnFbU and MgFbU in toluene (A) and 2-nitrotoluene (B) obtained at 295 K.The bands marked by as a function of solvent relative permittivity (corrected for solvent refractive index) at 298 K asterisks are due to solvent. J. Mater. Chem., 1997, 7(7), 1245–1262 1255nCOC modes of MgU and ZnU are essentially identical. This is Time-resolved absorption spectra. The energy-transfer also the case for MgFbU and ZnFbU.On the other hand, the dynamics from the excited metalloporphyrin to the groundfrequencies of the nCOC modes of the dimers are somewhat state Fb-porphyrin in the ZnFbU and MgFbU dimers was higher than those of the monomers. These dierences are not probed using femtosecond transient absorption spectroscopy. due, however, to eects of linking the porphyrins in the array, Representative data for MgFbU and ZnFbU are shown in but rather to the fact that the ethyne substituent of the Fig. 5. The use of 582 nm pump flashes results in absorption monomers (FbU, ZnU, MgU) contains a terminal trimethylsi- by both the metalloporphyrin and Fb-porphyrin. Therefore, lyl group rather than an appended aryl ring. In monomeric the transient absorption dierence spectra immediately after metalloporphyrins containing a diarylethyne substituent excitation contain mixtures in which either the metalloporphy- (MgU, ZnU), nCOC is at ca. 2217 cm-1,12 nearly the same rin of the dimer is excited or the Fb-porphyrin is excited. This frequency observed for MgFbU and ZnFbU. The frequencies fact is revealed by the 1ps spectrum for MgFbU in toluene of the nCOC modes of the various Mg- and Zn-porphyrins are shown in Fig. 5A. Although there are dierences in peak also identical in toluene and 2-nitrotoluene [and CH2Cl2, positions and intensities, the singlet excited states of both the DMF, and THF (not shown)]. Together, these results indicate Mg- and Fb-porphyrins exhibit strong absorption between that the ground-electronic-state structure of the linkers is 430 and 500 nm corresponding to the Soret band of the excited essentially identical in the Mg- and Zn-porphyrins and is not state.44 The dip near 515 nm in the 1ps spectrum is due to influenced by the properties of the solvent.bleaching of the shortest wavelength of the four ground-state The RR spectra show that the frequencies of the n2 modes Q bands of the Fb-porphyrin, namely the Qy(1,0) band.The of both MgU and MgFbU are lower than those of the Zn- dip near 560 nm is mostly due to bleaching of the Q(0,0) containing analogues. These frequency dierences are attri- ground-state absorption band of the Mg-porphyrin, along with buted to the fact that Mg- and Zn-porphyrins have slightly some Fb-porphyrin bleaching. The trough at 610 nm is com- dierent structures and core geometries.24 On the other hand, prised of Q(0,0) bleaching and stimulated emission from the the frequencies of the n2 modes of MgFbU and ZnFbU are Mg-porphyrin along with some bleaching of the Qx (1,0) band approximately the same as those of MgU and ZnU, respect- of the Fb-porphyrin.The dip near 650 nm also has overlapping ively, indicative of the fact that array formation does not aect contributions, namely Qx(0,0) bleaching and stimulated emis- the structure of the porphyrin ring(s).12 The general appearance sion from the Fb-porphyrin and Q(0,1) stimulated emission of the n2 spectral features of MgFbU are, however, dierent from the Mg-porphyrin.Finally, the dip near 720 nm is due from those of ZnFbU, whereas those of MgU and ZnU are quite similar.In particular, the n2 feature of ZnFbU is relatively narrow and comparable in width to that of ZnU and MgU. On the other hand, the n2 feature of MgFbU is somewhat broader and exhibits a shoulder on the high-frequency side. The narrowness of the n2 feature of ZnFbU arises because the frequencies of the n2 modes of the Zn- and Fb-components of the dimer are nearly coincident at ca. 1550 cm-1. This was confirmed by RR data obtained for the free-base monomer (not shown). The slightly downshifted frequency of the n2 mode of the Mg-component of MgFbU reveals the n2 band of the Fb-component which remains at ca. 1550 cm-1. The frequencies of the n2 modes of ZnU and ZnFbU are nearly the same in toluene and 2-nitrotoluene [and CH2Cl2 , THF and DMF (not shown)].This is also the case for MgU and MgFbU in all solvents with the exception of THF. Accordingly, the ground-electronic-state structures of the porphyrin ring(s) are relatively insensitive to the dielectric properties of the solvent. In the case of the Mg-porphyrins in THF, the n2 modes are downshifted to ca. 1536 cm-1 (not shown). This frequency shift is attributed to the fact that the Mg-porphyrin is hexacoordinate in THF, commensurate with the dierent absorption spectra observed in this solvent (vide supra).Further inspection of the RR data reveals that the relative intensities of the nCOC vs. n2 modes are approximately the same for the Mg- and Zn-porphyrins. In addition, these relative intensities remain the same in all of the solvents investigated (toluene, 2-nitrotoluene, CH2Cl2, THF and DMF). The increased relative intensities of the nCOC vs.n2 modes of the dimers vs. monomers again arises because of the presence of terminal aryl vs. trimethylsilyl groups in the two systems Fig. 5 Room-temperature transient dierence spectra acquired follow- (rather than being an eect of covalent linkage in the array).12 ing excitation of the dimers in toluene with a 0.2 ps flash at 582 nm.The apparent increased relative intensity of the nCOC vs. n2 The spectra for MgFbU in (A) were acquired at time delays of 1 ps mode of MgFbU vs. ZnFbU is solely due to the fact that the (solid) and 100 ps (dashed). The spectra for ZnFbU in (B) were n2 feature of the former dimer is broader than that of the latter.acquired at time delays of 1 ps (solid) and 100 ps (dashed). Note that Integration of the RR band contours for MgFbU and ZnFbU the data in the red region were acquired with a sample of concentration reveals that the relative intensities of the nCOC vs. n2 modes of ca. ten times that used for the blue region so that quantitative comparison of the absorption changes in the two regions should not the two dierent arrays are in fact identical.Collectively, these be made. The insets show representative kinetic traces at 510–515 nm. results indicate that the excited-state electronic coupling Only the first 200 ps are shown, because from 200 ps to 3 ns there is between the ethyne group and the p system of the porphyrin no further change due to the long lifetime (12–13 ns) of the excited ring (which dictates the relative intensities of the nCOC vs.n2 Fb-porphyrin under these conditions. The curves through the kinetic modes12,14 ) is similar in the Mg- and Zn-porphyrins and is data are fits to a single exponential function with time constants of not strongly influenced by the coordination of the metal ion 31±3 ps (A) and 26±3 ps (B) that represent the lifetime of the excited metalloporphyrin component of the dimer.or the dielectric properties of the solvent. 1256 J. Mater. Chem., 1997, 7(7), 1245–1262to Qx (0,1) stimulated emission from the Fb-porphyrin. The in toluene. In 2-nitrotoluene, MgFbU and ZnFbU each exhibited lifetimes of ca. 8 ps; however, in this solvent, MgU and stimulated (by the white-light probe pulse) emission features are characteristic of the excited singlet states of the porphyrins ZnU also exhibited dramatically shortened lifetimes (14 and 52 ps, respectively). The shortened lifetime of each metallopor- and occur near the wavelengths of the features observed in the spontaneous emission (fluorescence) spectrum, as is observed phyrin (MgU, ZnU) indicates that the results observed in the dimers are attributable to solvent–porphyrin interactions for the dimers under investigation here.Fig. 5A shows that at 100 ps the Qy(1,0) bleaching near rather than to enhanced energy-transfer rates or competitive electron-transfer pathways inherent in the dimers. 515 nm due to the Fb-porphyrin in MgFbU has grown in magnitude while bleaching of the Q(0,0) band near 560 nm of the Mg-porphyrin has decreased.Changes are also observed Time-resolved fluorescence. The lifetimes of the Fbporphyrins in the arrays, and the metalloporphyrins in the in the other regions of the spectrum. The dierences between the spectra at 1 and 100 ps clearly reflect disappearance of the MgU and ZnU monomers, were measured by time-resolved fluorescence spectroscopy.The results are summarized in Mg-porphyrin excited state (Mg*) and the formation of the Fb-porphyrin excited state (Fb*) in that fraction of the dimers Table 4. Also included in Table 4 are lifetimes obtained previously in toluene and DMSO for the ZnFbU dimer and the in which the Mg-porphyrin had been excited by the pump flash. Note, however, that the fraction in which the Fb* excited ZnU and FbU control complexes.13 These values and those obtained here for these systems are in excellent agreement. For state was initially produced does not change over the ca. 3 ns timescale of the measurements because the Fb* lifetime in example, the fluorescence lifetime of 13.1±0.6 ns obtained here for the Fb-porphyrin emission in the ZnFbU dimer in toluene toluene is 12–13 ns, as determined by fluorescence methods (vide inf ra).is the same within experimental error as the value of 12.5±0.2 ns obtained previously from time-correlated single A representative kinetic trace is shown in the inset to Fig. 5A along with a fit to a single exponential function with a time photon counting measurements. The same lifetime is found for FbU in toluene (13.3 vs. 12.5 ns). The lifetime of the Fb- constant of 31 ps. The same value (31±4 ps) is found from analysis of the data at all wavelengths where suciently large porphyrin in MgFbU in toluene is the same again (13.1 ns). The finding of the same lifetime (12–13 ns) of the excited absorption changes are observed. We assign this time constant as the Mg* lifetime for MgFbU in toluene. singlet state of the Fb-porphyrin in the monomer and dimers is in full accord with the observation that the fluorescence The ZnFbU dimerintoluene showssimilar results as observed for MgFbU.The transient absorption spectra and kinetic data yield (Wf#0.12) is basically the same in these molecules (Table 3 and ref. 13). Collectively, these results demonstrate in Fig. 5B are in excellent agreement with the data presented for ZnFbU previously.13 The observed lifetime of the Zn- that neither the Zn- nor Mg-porphyrin in these dimeric arrays in non-polar media introduces any new decay channels that porphyrinexcitedstate of 26±3 ps isthe same within experimental error as the value of 22±2 ps obtained previously.In compete eectively with the inherent decay routes (fluorescence, internal conversion, intersystem crossing) of the Fb- analogy with the above results on MgFbU, this time constant reflects the energy-transfer process Zn*Fb�ZnFb*.The life- porphyrin. The lifetime observed for the Fb-porphyrin in each of the pentamers (Mg4FbU and Zn4FbU) is shorter than that times of the pentameric arrays and representative monomers were also collected in toluene (Table 4).Each pentamer of the FbU-core porphyrin and of the Fb-porphyrin in the dimers (Table 3). The shortened values could be due to a (Mg4FbU, Zn4FbU) shows a metalloporphyrin lifetime that is slightly shorter than that observed in the respective dimer. quenching process or an altered structure of the core Fbporphyrin due to the presence of the four appended metal- The excited-state lifetimes of the metalloporphyrin (M*) in the MgFbU and ZnFbU dimers were examined in polar loporphyrins.As the polarity of the solvent increases, the lifetime of the solvents (Table 4). In acetone and DMSO, MgFbU and ZnFbU each exhibit essentially the same lifetime as observed Fb-porphyrin in the dimers becomes shorter. For example, the Table 4 Excited-state lifetimes of metallo- (M) and free-base (Fb) porphyrins (296 K)a toluene EAb THF acetone 2-nitrotoluene acetonitrile DMSO compound M Fb Fb M Fb M Fb M Fb M Fb M Fb MgFbU 31 ps 13.1 ns 13.8 ns 34 ps 12.8 ns 37 psc 6.5 ns 9 ps 2.2 ns 3.8 ns 31 ps 1.3 ns 2.0 nsd 1.8 nsd ZnFbU 26 ps 13.1 ns 13.5 ns 13.3 ns 30 psc 11.7 ns 8 ps 4.4 ns 7.3 ns 4.3 ns 22 pse 12.5 nse 4.5 nsd 23 pse 4.8 nse Mg4FbU 23 ps 10.7 ns Zn4FbU 17 ps 10.6 ns MgU 10.0 ns 9.7 ns 9.8 ns 14 ps 9.6 ns 8.4 ns ZnU 2.5 ns 2.6 ns 2.7 ns 52 ps 2.4 ns 2.4 ns 2.4 nse 2.3 nse FbU 13.3 ns 13.8 ns 13.7 ns 13.6 ns 12.3 ns 13.5 ns 13.0 ns 12.6 nse FbU-core 12.5 ns MgTPP 8.3 ns ZnTPP 2.2 ns 2.1 ns 2.1 ns aThe lifetimes of the excited metalloporphyrin in the dimers and pentamers were determined using transient absorption spectroscopy. The lifetimes for the excited Fb-porphyrin monomers, metalloporphyrin monomers, and the Fb-porphyrins in the dimers were obtained by timeresolved fluorescence spectroscopy.The error limits for the lifetimes obtained by time-resolved fluorescence spectroscopy are ±5% and those from time-resolved absorption spectroscopy are ±10%. bData reported only for the Fb-porphyrin in ethyl acetate. cA slightly longer value (ca. 60 ps) is measured in the range 450–500 nm, whereas the reported value was obtained at 515 nm. In all other solvents, the same lifetime was observed at alwavelengths between 450 and 550 nm. Because the same ca. 30 ps value is measured in the 515 nm region in all the solvents, we believe that this value reflects the excited metalloporphyrin lifetime in all of the media, and the slower kinetics measured in acetone in the blue region involve some other processes that do not aect the excited metalloporphyrin lifetime. dThese values, also obtained from transient absorption spectroscopy, are only rough estimates because the kinetic data did not span a sucient time span for an accurate determination.eThese values are reproduced from a previous study13 and are in good agreement with the values obtained here.J. Mater. Chem., 1997, 7(7), 1245–1262 1257lifetime of the Fb-porphyrin in MgFbU in acetonitrile is 3.8 ns, depopulating the excited metalloporphyin in the arrays other than the intrinsic processes (intersystem crossing, internal which is a factor of 3.4 shorter than the value in toluene.This result parallels the behaviour observed for the fluorescence conversion, radiative decay) also present in the benchmark monomer. The close matching of the excitation spectra and yields (Table 3, Fig. 3). Qualitatively similar results are obtained with ZnFbU, though the extent of quenching is less absorption spectra for the arrays in diverse solvents indicates a high yield of energy transfer and supports this assumption.than with MgFbU. We attribute the yield and lifetime reductions in polar media to charge-transfer quenching of the However, within experimental uncertainty we cannot exclude the possibility of a small amount (10%) of electron transfer Fb-porphyrin excited state. from the M* excited state. The lifetimes observed for the metalloporphyrins in the Discussion arrays (Table 4) were used to compute the rates and yields of energy transfer in toluene (Table 5).For MgFbU, the lifetime Zinc has been widely employed as a surrogate for magnesium of the Mg* (tMg*=31 ps) gives kEnT ca. (31 ps)-1 due to the in the preparation of porphyrin-based synthetic models of inherent lifetime (t0Mg*) of 10.0 ns observed for Mg* in the chlorophylls.26 Our new synthetic methods for preparing Mg- MgU control compound.Similarly for ZnFbU, the lifetime of porphyrins and Mg-containing porphyrin arrays obviate the the Zn* (tZn*=26 ps) gives kEnT ca. (26 ps)-1 due to the reliance on Zn-porphyrins and shift the focus in light-har- inherent lifetime (t0Zn*) of 2.5 ns observed for Zn* in the ZnU vesting or molecular photonics applications from the question control compound. Thus, the rate of energy migration is nearly of ‘What is synthetically feasible?’ to the design issue of ‘Which identical in MgFbU and ZnFbU.metal exhibits more desirable photochemical and materials Although the rates of energy transfer are nearly identical for properties?’ We have explored the latter issue by examining the Mg- and Zn-containing arrays, the yields dier slightly as the pairwise interactions between metalloporphyrins (Mg, Zn) these reflect the inherent lifetime of the metalloporphyrin and Fb-porphyrins in dimers and star-shaped pentamers.These excited state [eqn. (2)]. For MgFbU, WEnT is ca. 99.7% (31 ps studies serve as a prelude to the design and preparation of lifetime in MgFbU vs. 10 ns lifetime in MgU) while for ZnFbU, larger multiporphyrin arrays. The major photochemical results WEnT is ca. 99.0% (26 ps lifetime in ZnFbU vs. 2.5 ns lifetime from this comparative study are as follows. (1) The choice of in ZnU). Thus the yield of energy transfer is greater for the metal ion (MgII vs. ZnII ) does not appreciably aect the rate Mg-containing array in spite of the marginally slower rate.of energy transfer in the arrays. The similarity in rates of Similar results are observed for the pentamers Zn4FbU and energy transfer for the Mg- and Zn-containing arrays is Mg4FbU, where a marginally faster rate [(17 ps)-1 vs. (23 attributed to the fact that the electronic coupling between the ps)-1, respectively] is observed for the former but the latter metalloporphyrin and Fb-porphyrin is approximately the same gives the higher yield (99.3 vs. 99.8%, respectively). Thus, the for Mg- vs. Zn-containing arrays. (2) The quantum yields of general features observed for Mg- and Zn-containing arrays energy transfer are 99% for arrays containing either metal are quite similar. ion. However, the yield of energy transfer is slightly higher in The similarity in rates of energy transfer observed for the the Mg- vs.Zn-containing arrays owing to the longer intrinsic Mg- and Zn-containing arrays can be attributed to the fact lifetime of the former metalloporphyrin. (3) Solvent polarity that the electronic coupling between the metalloporphyrin and and changes in coordination geometry of the Mg- or Zn- Fb-porphyrin is approximately the same for the Mg- vs.Zn- porphyrin have very little eect on the rates and yields of containing arrays. In this connection, our prior studies of a energy transfer. (4) Polar solvents diminish the fluorescence variety of ZnFb-dimers have shown that the predominant yield and lifetime of the excited Fb-porphyrin in arrays contain- pathway for energy transfer involves a through-bond rather ing either Mg- or Zn-porphyrins.The magnitude of the dimin- than a through-space mechanism.13 The through-space energy- ution is greater for the Mg-containing arrays. These eects are transfer rate was calculated to be (720 ps)-1 for the ZnFb- attributed to charge-transfer quenching of the excited Fb- dimeric arrangement (which is substantially slower than the porphyrin by the adjacent metalloporphyrin.The enhanced observed rates), and the MgFb-dimeric structure is expected quenching in the Mg-containing arrays is a result of the greater to have the same through-space rate. The through-bond driving force for charge separation. In the following section, energy-transfer process is explicitly mediated by the nature of we discuss each of these points in more detail. Next, we the conformational energy surface of the diarylethyne linker, compare the energy-transfer characteristics of our MgFb- or which dictates the extent of electronic communication between ZnFb-containing arrays with qualitatively similar phenomena the p systems of linker and porphyrin ring(s).The Raman exhibited by selected arrays made by other workers.Finally, intensity of the nCOC mode of the arylethyne linker (relative to we comment on the merits of Mg- vs. Zn-porphyrins for the n2 mode of the porphyrin) was shown to be a convenient materials applications in the context of their photochemical static spectroscopic signature of the extent of this electronic properties and their diering stabilities. interaction.14 In particular, the intensity of the nCOC mode was shown to parallel the rate of energy transfer in a series of Photochemical characteristics of Mg- vs.Zn-containing arrays ZnFb dimers containing diering degrees of torsional con- Energy transfer rates and yields. The static fluorescence yield straint.13,14 In the case of MgFbU vs. ZnFbU, the similarity and fluorescence excitation spectral measurements indicate in the relative intensities of the nCOC modes again parallels the that the yield of energy transfer is essentially quantitative in similarity in the rates of energy transfer.Collectively, these both the Mg- and Zn-containing arrays. In this regime of high eciency, a precise determination of the yield is best obtained Table 5 Calculated energy transfer rate constants and yields in via time-resolved measurements. From the measured lifetime toluene (298 K)a of the metalloporphyrin in an array (tM*) and the lifetime of a benchmark monomeric porphyrin (t0M*), the rate constant kEnT WEnT (%) for energy transfer (kEnT) from M* to Fb and the yield of energy transfer (WEnT) can be calculated as shown in eqn.(1) MgFbU (31 ps)-1 99.7 ZnFbU (26 ps)-1 99.0 and (2).Mg4FbU (23 ps)-1 99.8 kEnT=(tM*)-1-(t0M*)-1 (1) Zn4FbU (17 ps)-1 99.3 WEnT=kEnTtM*=1-tM*/t0M* (2) aCalculated from the excited-state lifetimes in Table 4 using eqn. (1)–(4). These equations assume there are no other pathways for 1258 J. Mater. Chem., 1997, 7(7), 1245–1262Table 6 Calculated charge-transfer (MFb*�MV+FbV-) rate constants and yields (298 K)a compound process toluene ethyl acetate THF acetone 2-nitrotoluene acetonitrile DMSO MgFbU kFb*CT NAb NA (195 ns)-1 (12.5 ns)-1 (2.7 ns)-1 (5.3 ns)-1 (1.4 ns)-1 WFb*CT (%) NA NA 7 52 82 72 90 ZnFbU kFb*CT NA NA (456 ns)-1 (84 ns)-1 (6.9 ns)-1 (15.3 ns)-1 (6.4 ns)-1 WFb*CT (%) NA NA 3 14 64 46 67 aCalculated from the excited-state lifetimes in Table 4 using eqn.(1)–(4).bNot applicable. The lifetime is the same as in the FbU control compound, indicating no charge transfer occurs in this solvent. trends lead to the assessment that the extent of electronic the data obtained for MgFbU allow a more in-depth analysis of the charge-transfer process. coupling is similar in the Mg- and Zn-containing arrays. The lifetime of the Fb-porphyrin emission in the dimeric arrays (tFb*) and the fluorescence lifetime of the FbU control Eects of solvent on photodynamics of energy transfer.The solubility of the arrays provides the opportunity for examining complex (t0Fb*) can be used to determine the rate constant for charge-transfer quenching (kFb*CT ) and the yield for the the eects of dierent media on the excited state photodynamics.Our previous study of ZnFbU showed that the energy- quenching process (WFb*CT) via eqn. (3) and (4) (Table 6). transfer rate changed by a factor of 2.5-fold upon changes kFb*CT=(tFb*)-1-(t0Fb*)-1 (3) in viscosity (fluid medium to rigid glass), temperature (298–150 K), or polarity (solvent static relative permittivity WFb*CT=kFb*CT tFb*=1-tFb*/t0Fb* (4) e=2.38–46.7).13 In MgFbU, the fluorescence yield and lifetime of the Mg-porphyrin are basically unchanged in going from For example, MgFbU in acetonitrile has kFb*CT ca.(5.3 ns)-1 toluene to acetone to DMSO, indicating that the rate of energy and WFb*CT ca. 72% under these conditions (the competing transfer is unaected by this dramatic change in solvent processes being fluorescence, internal conversion, and inter- polarity.Similar results were observed for ZnFbU. system crossing), while ZnFbU in acetonitrile has kFb*CT ca. Although the rates of energy transfer do not change with (15.3 ns)-1 and WFb*CT ca. 46%. The most polar solvent increased solvent polarity, the reasonably polar solvent 2- examined, DMSO, gives substantial quenching, with WFb*CT nitrotoluene is exceptional compared with the rest of the ca. 90 and 67% for the MgFb- and ZnFb-dimers, respectively. solvents in giving shortened lifetimes. In this solvent, the Note that for both types of arrays, the values of kFb*CT and lifetimes of the MgU and ZnU control complexes are dramati- WFb*CT generally increase as the solvent polarity increases, cally reduced (to 14 ps and 52 ps, respectively) from the though the trend is not exactly linear as 2-nitrotoluene lifetimes in toluene (10.0 and 2.5 ns, respectively).Thus, the quenches slightly more than expected given its polarity. There similar shortening in the lifetimes in the arrays does not reflect is no reason to expect this trend to follow the solvent polarity enhanced energy transfer (or competitive electron transfer) but precisely, because other factors such as porphyrin electrochemi- is a consequence of a direct quenching interaction with the 2- cal potentials are of critical importance in determining the nitrotoluene which provides a very fast decay route for M*.extent of charge-transfer quenching, and these also are aected The invariance in rates of energy transfer upon changes in by the nature of the solvent.23b solvent is supported by the RR data. These data show that Collectively, the analysis of the fluorescence yield (Fig. 3) the magnitude of the electronic coupling in both the Mg- and and lifetime (Table 4) data of the MgFbU vs.ZnFbU array as Zn-containing arrays is not strongly aected by the nature of a function of solvent polarity indicate that the charge-transfer the solvent (polar vs.non-polar) or the coordination number process is more pronounced for the former arrays. The more of the metal ion (and hence the conformation of the porphyrin facile charge-transfer quenching observed for MgFbU vs. ring). The constancy of the electronic coupling under a variety ZnFbU is attributed to the larger driving force of conditions has important implications for the interpretation for the MgFb*�MgV+FbV- process relative to the of the eects of solvent on the photochemical properties of ZnFb*�ZnV+FbV- reaction.This dierence derives from the the arrays. ca. 300 mV greater ease of oxidation of the Mg-porphyrin relative to the Zn-porphyrin.10,23,32 Hence, Mg-containing Eects of solvent on fluorescence of the Fb-porphyrin. arrays oer some advantage over Zn-containing arrays in Although the rate of energy transfer and the magnitude of situations wherein charge transfer is a desired property of the electronic coupling remain constant upon changes in solvent, assembly.both the fluorescence yield and excited-state lifetime of the Fbporphyrin are diminished in polar solvents. Together these Energy-transfer properties of other dimers observations lead to the assessment that the quenching phenomenon is best attributed to a charge-transfer process A large number of dimers have been prepared for studies of (MFb*�MV+FbV-) that occurs following energy transfer energy transfer, with most containing Zn- and Fb- from the metallo(M)- to Fb-porphyrin (or following direct porphyrins.6,26 Among these, the most relevant to this dis- excitation of the Fb-porphyrin).The quenching of the Fb- cussion are those where charge-transfer quenching of the porphyrin excited state must occur within its nominal 12–13 ns Fb-porphyrin has been observed. Also relevant are those arrays lifetime. A model that accounts for this quenching is shown in where energy transfer has been studied between Mg- and Fb- Scheme 5.We previously reported quenching of the Fb- porphyrins, though these are far fewer in number.28k,29a porphyrin in ZnFbU in the polar solvent DMSO.13 The more Gust et al. prepared a ZnFb-dimer joined by a phenyl– comprehensive data base reported here for ZnFbU along with amide–phenyl linker and observed fast energy transfer [kEnT= (43 ps)-1] with WEnT=0.97 in CH2Cl2 .45 A dimer with the same linker but electron-deficient substituents on the Fbporphyrin exhibited slightly slower energy transfer [kEnT=(106 ps)-1] which was competitive with electron transfer (WET= 0.77, WET=0.18).In addition, the excited-state lifetime of the Fb-porphyrin was shortened from 8.5 to 2.7 ns which was Scheme 5 Charge-transfer quenching of the excited Fb-porphyrin (Fb*) by the ground-state metalloporphyrin (M) attributed to charge-transfer quenching (WFb*CT=0.68). J.Mater. Chem., 1997, 7(7), 1245–1262 1259Although both dimers were examined in the non-polar solvent Mg-containing arrays is due to the fact that magnesium is less electronegative than zinc (x=1.31 vs. 1.65); consequently, Mg- CH2Cl2, the electron-deficientgroups provide increaseddriving force for the charge-transfer process and thus play a role porphyrins are more easily oxidized than Zn-porphyrins.(3) MgII is a harder (less malleable) ion and strongly prefers similar to that of a polar solvent in enhancing this process. Two related ZnFb-dimers, each with an electron-rich Zn- oxygenic rather than nitrogenous ligands, while ZnII is softer and has similar anity for both.49 This preference for ligand porphyrin and an electron-deficient Fb-porphyrin, were examined in a wide variety of solvents.46 The charge-transfer quench- type and geometry leads to much easier acid-induced demetallation of Mg-porphyrins than Zn-porphyrins, a property with ing process of the Fb-porphyrin was not detected in toluene, became apparent in solvents such as ethyl acetate, and considerable practical implications (in chromatographic puri- fication of the arrays, silica is suciently acidic to demetallate increased dramatically in rate upon going to polar solvents such as DMSO.Mg- but not Zn-porphyrins; thus, the former generally cannot be chromatographed on silica).33 Several dimers28 and trimers29 containing Mg-porphyrins have been prepared, but in almost all cases electron transfer The viemerges from the above considerations is that the construction of molecular photonic devices based on Mg- rather than energy transfer has been the dominant photochemical process.Osuka et al. prepared a series of MgFb dimers or Zn-porphyrins will be based on a host of factors in addition to the photochemical properties of the two types of arrays.joined by hydrocarbon spacers of various lengths wherein energy transfer was studied.28k In one bis-spiroindane linked Clearly, either metal ion could be used if the rate of energy transfer is the only factor to be considered. However, Mg- MgFb dimer examined in DMF, fast energy transfer [kEnT= (62 ps)-1] and faster electron transfer [kET=(42 ps)-1] were containing arrays may oer certain advantages in cases where a succession of energy-transfer steps occur.For example, observed with quantum yields of 0.4 and 0.6, respectively. In other MgFb dimers with assorted linkers, energy transfer assuming the yields observed for MgFbU (99.7%) and ZnFbU (99.0%) carry over to extended arrays of metalloporphyrins, occurred with no observable competing electron-transfer processes.In the series of MgFb dimers, the rates of energy then upon 100 transfer steps, the all-Mg-containing arrays would give 74% eciency while the all-Zn-containing arrays transfer changed little (less than two-fold) in going from toluene to THF. In the same series in DMF, however, quench- would give 37% eciency. Such a large number of transfer steps would be required in realistic models of the natural light- ing of the Fb-porphyrin fluorescence was observed with a rate increasing with shorter linkers and this also was attributed to harvesting arrays.These considerations must be balanced with the fact that Mg-porphyrins are more prone to oxidation, and charge separation (MgFb*�MgV+FbV-) as proposed here.this must be suppressed in a light-harvesting array. Conversely, the propensity toward oxidation of Mg-porphyrins might be Merits of Mg- vs. Zn-containing arrays for materials attractive in other types of devices such as switching elements. applications Indeed, we have applied the basic elements of this concept in the construction of a prototypical molecular optoelectronic Synthetic multiporphyrin nanostructures constitute a relatively new class of optical and photonic materials.The modularity gate.10 Finally, from a processing or device packaging standpoint, Zn-porphyrins are more robust toward demetallation. of the building block approach enables relatively easy preparation of diverse composite arrays containing dierent meta- However, Mg-chelates of electron-deficient porphyrins are less susceptible to demetallation than are those with electron-rich llo- or Fb-porphyrins.The studies reported herein indicate that the energy-transfer characteristics of the Mg-containing substituents. Thus, the incorporation of appropriate electronwithdrawing groups with the Mg-porphyrin may provide arrays are generally similar to those of their Zn-containing counterparts. The overall similarity in this property of the two protectiontoward demetallation, charge transfer, and photooxidation while maintaining the desired long lifetime of the types of arrays is on the surface surprising given the fact that magnesium (atomic number 12) is an alkaline earth metal singlet excited state.All of these factors need to be considered in the design of metalloporphyrin-based light-harvesting arrays while zinc (atomic number 30) is a transition metal with a filled 3d shell of electrons.For example, the MgII ion prefers and nanostructures for materials applications. an octahedral coordination sphere but also can accommodate a square-pyramidal geometry, while the presence of the mal- Conclusions leable 3d shell of electrons in ZnII leads to tolerance of a variety of coordination spheres.Nonetheless, MgII and ZnII Multiporphyrin arrays comprised of Fb- and Mg- or Znporphyrins can be constructed using a modular building block have nearly the same ionic radius (0.72 A° and 0.74 A° , respectively), 47 which is slightly larger than optimal for a comfortable approach. No significant dierences exist in the synthesis of Mg- or Zn-containing porphyrins or related arrays.Arrays fit in the porphyrin core.48 Regardless, these basic dierences in electronic structure and coordination-sphere geometry have containing Mg-porphyrins provide new models for biomimetic investigations of natural light-harvesting phenomena. In many very little influence on the linker-mediated electronic coupling between the metallo- and Fb-porphyrin which dictates the regards, magnesium and zinc can be used almost interchangeably in many light-harvesting arrays.The choice of energy-transfer rates. Certain dierences do exist in the photochemical and mate- metal can be based on subtle factors such as the slightly faster rate of energy transfer provided by Zn-porphyrins, the slightly rials properties of the Mg- vs.Zn-containing arrays. However, all of the dierences can be directly traced to dierences which higher yield of Mg-porphyrins emanating from the inherently longer lifetime of Mg-porphyrins, or the desire to favour are intrinsic to monomeric Mg- vs. Zn-porphyrins rather than being a consequence of array formation. These dierences are charge-transfer processes to which Mg-porphyrins are more inclined.Regardless, the choice of metal is now a design issue as follows. (1) Although both MgII and ZnII are diamagnetic and support metalloporphyrin singlet excited states with nano- rather than a synthetic consideration. The ability to tune the photodynamic properties of the arrays through choice of metal second lifetimes, Zn-porphyrins have a shorter lifetime (and commensurably diminished fluorescence yield) than Mg- is an important handle for controlling the flow of energy in porphyrin-based nanostructures.Finally, metalloporphyrins porphyrins (2–2.5 ns vs. 8–10 ns, respectively). The shorter lifetime of Zn-porphyrins is due to the increased rate of containing metals with far larger dierences than magnesium and zinc should also be accessible via this modular synthetic intersystem crossing, which stems from the heavy-atom eect, not dierences in radiative decay [MgTPP and ZnTPP have approach, in turn broadening the scope of photonic and electronic properties that can be elicited in these porphyrin- identical fluorescent radiative decay rates, kf ca.(60 ns)-1].36 (2) The somewhat increased charge-transfer propensity of the based materials. 1260 J. Mater. Chem., 1997, 7(7), 1245–126225 M. Gouterman, in T he Porphyrins, ed. D. Dolphin, Academic This work was supported by a grant from Division of Chemical Press, New York, 1978, vol. 3, pp. 1–166. Sciences, Oce of Basic Energy Sciences, Oce of Energy 26 For reviews, see: (a) S. G. Boxer, Biochim.Biophys. Acta, 1983, 726, Research, Department of Energy (J.S.L.), the LACOR Program 265; (b) D. Gust and T. A. Moore, Science, 1989, 244, 35; (c) V. V. (D.F.B.) from Los Alamos National Laboratory, and grants Borovkov, R. P. Evstigneeva, L. N. Strekova and E. I. Filippovich, GM36243 (D.F.B.) and GM34685 (D.H.) from the National Russ. Chem. Rev., 1989, 58, 602; (d) D. Gust and T.A. Moore, T op. Institute of General Medical Sciences. Laser desorption mass Curr. Chem., 1991, 159, 103; (e)M. R. Wasielewski, in Chlorophylls, ed. H. Scheer, CRC Press, Boca Raton, FL, USA, 1991, spectra were obtained by Dr.Walter Svec at Argonne National pp. 269–286; ( f ) M. R. Wasielewski, Chem. Rev., 1992, 92, 435; (g) Laboratory. D. Gust, T. A. Moore and A. L. Moore, Acc.Chem. Res., 1993, 26, 198; (h) S. E. Gribkova, R. P. Evstigneeva and V. N. Luzgina, Russ. Chem. Rev., 1993, 62, 963; (i ) H. Kurreck and M. Huber, Angew. References Chem., Int. Ed. Engl., 1995, 34, 849. 27 (a) K. N. Solovev, M. P. Tsvirko, A. T. Gradyushko and D. T. 1 (a) A. W. D. Larkum and J. Barrett, Adv. Bot. Res., 1983, 10, 1; (b) Kozhich, Opt. Spectrosc., 1972, 33, 480; (b) T.Yamamura, Chem. Photosynthetic L ight-Harvesting Systems, ed. H. Scheer and L ett., 1977, 773; (c) K. Yamashita, N. Kihara, H. Shimidsu and S. Siegried, W. de Gruyter, Berlin, 1988; (c) D. C. Mauzerall and H. Suzuki, Phochem. Photobiol., 1982, 35, 1; (d) T. Katagi, N. L. Greenbaum, Biochim. Biophys. Acta, 1989, 974, 119; (d) C.N. T. Yamamura, T. Saito and Y. Sasaki, Chem.L ett., 1982, 417; (e) Hunter, R. van Grondelle and J. D. Olsen, T rends Biochem. Sci., W. Szulbinski and M. Lapkowski, Inorg. Chim. Acta, 1986, 123, 1989, 14, 72; (e) G. McDermott, S. M. Prince, A. A. Freer, A. M. 127; (f ) K. Takahashi, H. Katsurada, T. Komura and H. Imanaga, Hawethornthwaite-Lawless, M. Z. Papiz, R. J. Cogdell and N. W. Bull. Chem. Soc. Jpn., 1990, 63, 3315; (g) S.Yamauchi, T. Ueda, Isaacs, Nature (L ondon), 1995, 374, 517; (f ) S. Karrasch, P. A. M. Satoh, K. Akiyama, S. Tero-Kubota, Y. Ikegami and Bullough and R. Ghosh, EMBO J., 1995, 14, 631. M. Iwaizumi, J. Photochem. Photobiol. A: Chem., 1992, 65, 177; (h) 2 R. Jimenez, S. N. Dikshit, S. E. Bradforth and G. R. Fleming, Y. Iseki and S. Inoue, J. Chem. Soc., Chem. Commun., 1994, 2577; J.Phys. Chem., 1996, 100, 6825. (i ) K. Wynne, S. M. LeCours, C. Galli, M. J. Therien and R. M. 3 J. S. Lindsey, S. Prathapan, T. E. Johnson and R. W. Wagner, Hochstrasser, J. Am. Chem. Soc., 1995, 117, 3749. T etrahedron, 1994, 50, 8941. 28 (a) S. G. Boxer and G. L. Closs, J. Am. Chem. Soc., 1976, 98, 5406; 4 S. Prathapan, T. E. Johnson and J. S. Lindsey, J. Am. Chem. Soc., (b) C.K. Chang, J. Heterocycl. Chem., 1977, 14, 1285; (c) M. R. 1993, 115, 7519. Wasielewski, W. A. Svec and B. T. Cope, J. Am. Chem. Soc., 1978, 5 R.W.Wagner and J. S.Lindsey, J. Am. Chem. Soc., 1994, 116,9759. 100, 1961; (d) T. L. Netzel, P. Kroger, C. K. Chang, I. Fujita and 6 R. W. Wagner, T. E. Johnson and J. S. Lindsey, J. Am. Chem. Soc., J. Fajer, Chem. Phys. L ett., 1979, 67, 223; (e)M.R.Wasielewski and 1996, 118, 11166. W. A. Svec, J. Org. Chem., 1980, 45, 1969; ( f ) T. L. Netzel, M. A. 7 N. Nishino, R. W. Wagner and J. S. Lindsey, J. Org. Chem., 1996, Bergkamp, C. K. Chang and J. Dalton, J. Photochem., 1981, 17, 61, 7534. 451; (g) I. Fujita, J. Fajer, C. K. Chang, C. B. Wang, M. A. 8 J. S. Lindsey, in Modular Chemistry, ed. J. Michl, Kluwer Bergkamp and T.L. Netzel, J. Phys. Chem., 1982, 86, 3754; (h) T. L. Academic, Dordrecht, in press. Netzel, M. A. Bergkamp and C. K. Chang, J. Am. Chem. Soc., 1982, 9 R. W. Wagner, T. E. Johnson, F. Li and J. S. Lindsey, J. Org. 104, 1952; (i ) R. E. Overfield, A. Scherz, K. J. Kaufmann and M. Chem., 1995, 60, 5266. R. Wasielewski, J. Am. Chem. Soc., 1983, 105, 4256; (j ) 10 R. W.Wagner, J. S. Lindsey, J. Seth, V. Palaniappan and D. F. H. Kamogawa, S. Miyama and S. Minoura,Macromolecules, 1989, Bocian J. Am. Chem. Soc., 1996, 118, 3996. 22, 2123; (k) A. Osuka, F. Kobayashi, K. Maruyama, N. Mataga, 11 D. Fenyo, B. T. Chait, T. E. Johnson and J. S. Lindsey, J. Porph. T. Asahi, T. Okada, I. Yamazaki and Y. Nishimura, Chem. Phys. Phthalocy., 1997, 1, 93. L ett., 1993, 201, 223; (l ) P.Hilderbrandt, H. Tamiaki, A. R. 12 J. Seth, V. Palaniappan, T. E. Johnson, S. Prathapan, J. S. Lindsey Holzwarth and K. Schaner, J. Phys. Chem., 1994, 98, 2192; (m) and D. F. Bocian, J. Am. Chem. Soc., 1994, 116, 10 578. H. Nakamura, M. Terazima, N. Hirota, S. Nakajima and 13 J. S. Hsiao, B. P. Krueger, R. W. Wagner, T. E. Johnson, J. K. A. Osuka, Bull.Chem. Soc. Jpn., 1995, 68, 2193; (n) H. Berg, Delaney, D. C. Mauzerall, G. R. Fleming, J. S. Lindsey, D. F. M. Rachamim, T. Galili and H. Levanon, J. Phys. Chem., 1996, Bocian and R. J. Donohoe, J. Am. Chem. Soc., 1996, 118, 11 181. 100, 8791. 14 J. Seth, V. Palaniappan, R. W.Wagner,T. E. Johnson, J. S. Lindsey 29 (a) S. G. Boxer and R. R. Bucks, J. Am. Chem. Soc., 1979, 101, 1883; and D.F. Bocian, J. Am. Chem. Soc., 1996, 118, 11 194. (b) R. R. Bucks, T. L. Netzel, I. Fujita and S. G. Boxer, J. Phys. 15 A. A. Bothner-By, J. Dadok, T. E. Johnson and J. S. Lindsey, Chem., 1982, 86, 1947; (c) A. Osuka, B. Liu and K. Maruyama, J. Phys. Chem., 1996, 100, 17551. J. Org. Chem., 1993, 58, 3582; (d ) A. Osuka, S. Marumo, 16 (a) L. Milgrom, J. Chem. Soc., Perkin T rans. 1, 1983, 2535; S. Taniguchi, T. Okada and N. Mataga, Chem. Phys. L ett., 1994, (b) J. Davila, A. Harriman and L. R. Milgrom, Chem. Phys. L ett., 230, 144; (e) Y. Kobuke and H. Miyaji, Bull. Chem. Soc. Jpn., 1996, 1987, 136, 427. 69, 3563. 17 (a) O. Wennerstrom, H. Ericssom, I. Raston, S. Svensson and 30 (a) H. Miyaji, Y. Kobuke and J. Kondo, Chem. L ett., 1996, 497; (b) W. Pimlott, T etrahedron L ett., 1989, 30, 1129; (b) A. Osuka, B. Liu H. Tamiaki, M. Amakawa, Y. Shimono, R. Tanikaga, A. R. and K. Maruyama, Chem. L ett., 1993, 949. Holzwarth and K. Schaner, Photochem. Photobiol., 1996, 63, 92. 18 D. L. Ocer, A. K. Burrell and D. C. W. Reid, Chem. Commun., 31 (a) A. T. Gradyushko and M. P. Tsvirko, Opt. Spectrosc., 1971, 31, 1996, 1657. 291; (b) A. Harriman, J. Chem. Soc., Faraday T rans. 2, 1981, 77, 19 (a) T. Nagata, A. Osuka and K. Maruyama, J. Am. Chem. Soc., 1281; (c) T. G. Politis and H. G. Drickamer, J. Chem. Phys., 1982, 1990, 112, 3054; (b) K. Ichihara and Y. Naruta, Chem. L ett., 1995, 76, 285; (d) O. Ohno, Y. Kaizu and H. Kobayashi, J. Chem. Phys., 631. 1985, 82, 1779. 20 (a) A. Osuka, N. Tanabe, R. P. Zhang and K. Maruyama, Chem. 32 (a) J. H. Fuhrhop and D. Mauzerall, J. Am. Chem. Soc., 1969, 91, L ett., 1993, 1505; (b) A. Osuka, N. Tanabe, S. Nakajima and 4174; (b) J. H. Fuhrhop, Z. Naturforsch., T eil B, 1970, 25, 255; (c) K. Maruyama, J. Chem. Soc., Perkin T rans. 2, 1996, 199; (c) N. Carnieri and A. Harriman, Inorg. Chim. Acta, 1982, 62, 103. S. Anderson, H. L. Anderson and J. K. M. Sanders, Angew. Chem., 33 J. S. Lindsey and J. N.Woodford, Inorg. Chem., 1995, 34, 1063. Int. Ed. Engl., 1992, 31, 907. 34 D. F. O’Shea, M. A. Miller, H. Matsueda and J. S. Lindsey, Inorg. 21 H. Aota, Y. Itai, A. Matsumoto and M. Kamachi, Chem. L ett., Chem., 1996, 35, 7325. 1994, 2043. 35 A. J. Gordon and R. A. Ford, in T he Chemist’s Companion, Wiley 22 (a) S. Anderson, H. L. Anderson, A. Bashall, M. McPartlin and Interscience, New York, 1972, pp. 374–375. J. K. M. Sanders, Angew. Chem., Int. Ed. Engl., 1995, 34, 1096; (b) 36 P. G. Seybold and M. Gouterman, J. Mol. Spectrosc., 1969, 31, 1. H. L. Anderson, S. Anderson and J. K. M. Sanders, J. Chem. Soc., 37 M. D. Ediger, R. S. Moog, S. G. Boxer and M. D. Fayer, Chem. Perkin T rans. 1, 1995, 2231; (c) S. Anderson, H. L. Anderson and Phys. L ett., 1982, 88, 123. J. K. M. Sanders, J. Chem. Soc., Perkin T rans. 1, 1995, 2255; (d) 38 J. A. Riddick and W. B. Bunger, Organic Solvents, T echniques of E. Alessio, M. Macchi, S. Heath and L. G. Marzilli, Chem. Chemistry, Wiley-Interscience, New York, 3rd edn., 1970 vol. 2. Commun., 1996, 1411. 39 C. Kirmaier and D. Holten, Biochemistry, 1991, 30, 609. 23 (a) D. G. Davis, in T he Porphyrins, ed. D. Dolphin, Academic 40 T. E. Johnson, Ph.D. Thesis, Carnegie Mellon University, 1995, Press, New York, 1978, vol. 5, pp. 127–152; (b) R. H. Felton, in T he p. 149. Porphyrins, ed. D. Dolphin, Academic Press, New York, 1978, 41 H. Brochmann and N. Risch, in Chlorophylls, ed. H. Scheer, CRC vol. 5, pp. 53–126. Press, Boca Raton, FL, 1991, p. 108. 24 W. R. Scheidt, in T he Porphyrins, ed. D. Dolphin, Academic Press, New York, 1978, vol. 3, pp. 463–512. 42 (a) R. P. Haugland, J. Yguerabide and L. Stryer, Proc. Natl. Acad. J. Mater. Chem., 1997, 7(7), 1245–1262 1261Sci. USA, 1969, 63, 23; (b) L. Stryer and R. P. Haugland, Proc. Natl. 46 J. M. DeGraziano, A. N. Macpherson, P. A. Liddell, L. Noss, J. P. Sumida, G. R. Seely, J. E. Lewis, A. L. Moore, T. A. Moore and Acad. Sci. USA, 1967, 58, 719. 43 J. S. Lindsey, P. A. Brown and D. A. Siesel, T etrahedron, 1989, D. Gust, New J. Chem., 1996, 20, 839. 47 N. N. Greenwood and A. Earnshaw, Chemistry of the Elements 45, 4845. 44 J. Rodriguez, C. Kirmaier and D. Holten, J. Am. Chem. Soc., 1989, Pergamon, Oxford, 1984. 48 J. W. Buchler, in T he Porphyrins, ed. D. Dolphin, Academic Press, 111, 6500. 45 D. Gust, T. A. Moore, A. L. Moore, F. Gao, D. Luttrull, J. M. New York, 1978, vol. 1, pp. 389–483. 49 J. Bjerrum,Metal Ammine Formation in Aqueous Solution, P. Haase DeGraziano, X. C. Ma, L. R. Makings, S. J. Lee, T. T. Trier, E. Bittersmann, G. R. Seely, S. Woodward, R. V. Bensasson, and Sons, Copenhagen, 1957, pp. 137–164. M. Rouge�e, F. C. De Schryver and M. Van der Auweraer, J. Am. Chem. Soc., 1991, 113, 3638. Paper 7/00146K; Received 7th January, 1997 1262 J. Mater. Chem., 1997, 7(7), 1245–12