Oxidation Reaction between Periodate and Polyhydroxyl Compounds

Department of Applied Chemistry, Graduate School of Engineering, Tokyo Metropolitan University, ... place in a strong alkaline solution without any sp...
0 downloads 0 Views 171KB Size
Download PDF
Anal. Chem. 1999, 71, 1760-1766

Oxidation Reaction between Periodate and Polyhydroxyl Compounds and Its Application to Chemiluminescence Jin-Ming Lin* and Masaaki Yamada

Department of Applied Chemistry, Graduate School of Engineering, Tokyo Metropolitan University, 1-1 Minami-Ohsawa, Hachioji, Tokyo 192-0397, Japan

The oxidation reaction between periodate and polyhydroxyl compounds was studied. A strong chemiluminescent (CL) emission was observed when the reaction took place in a strong alkaline solution without any special CL reagent. However, in acidic or neutral solution, it was hard to record the CL with our instrument. It was interesting to find that in the presence of carbonate the CL signal was enhanced significantly. When O2 gas and N2 gas were blown into the reagent solutions, both background and CL signals of the sample were enhanced by O2 and decreased by N2. The spectral distribution of the CL emission showed two main bands (λ ) 436-446 and 471-478 nm). Based on the studies of the spectra of CL, fluorescence and UV-visible, a possible CL mechanism was proposed. In strongly alkaline solution, periodate reacts with the dissolved oxygen to produce superoxide radical ions. A microamount of singlet oxygen (1O2*) could be produced from the superoxide radicals. A part of the superoxide radicals acts on carbonates and/or bicarbonates leading to the generation of carbonate radicals. Recombination of carbonate radicals may generate excited triplet dimers of two CO2 molecules ((CO2)2*). Mixing of periodate with carbonate generated were very few 1O2* and (CO2)2*. These two emitters contribute to the CL background. The addition of polyhydroxyl compounds or H2O2 caused enhancement of the CL signal. It may be due to the production of 1O2* during the oxidized decomposition of the analytes in periodate solution. This reaction system has been established as a flow injection analysis for H2O2, pyrogallol, and r-thioglycerol and their detection limits were 5 × 10-9, 5 × 10-9, and 1 × 10-8 M, respectively. Considering the effective reaction ions, IO4-, CO32-, and OH- could be immobilized on a strongly basic anion-exchange resin. A highly sensitive flow CL sensor for H2O2, pyrogallol, and r-thioglycerol was also prepared. Although there are many reports in the literature concerning the use of periodate as a reagent for degrading 1,2-glycols and related substances, and a number of works have examined the action of this oxidant on phenols, the uses of periodate in chemiluminescence (CL) are relatively few. Evmiridis reported a pyrogallol-periodate system for CL determinations of pyrogallol1 1760 Analytical Chemistry, Vol. 71, No. 9, May 1, 1999

and ethylene glycol.2 In his work, pyrogallol was used as the CL reagent and the CL mechanism was due to the formation of the singlet excited molecular oxygen species (1O2*).3,4 The CL intensity from the oxidation between periodate and pyrogallol could be enhanced by the addition of chromium(III).5 Instead of hydrogen peroxide, periodate was also used as an oxidant for the CL reactions of lucigenin6,7 and luminol.8,9 Almost all CL reactions of periodate with pyrogallol, luminol, or lucigenin needed a catalyst, for example, cobalt(II) and manganese(II). Although there are only a few reports on the CL applications of periodate, a number of papers on the oxidation of organic compounds by periodate have been published. The specific cleavage of 1,2-diols and the related compounds, R-diketones, R-ketols, R-amino alcohols and R-diamines have been widely exploited in the realm of carbohydrates and nucleic acids, and the mechanism has been established.10 Periodate was known to react with hydrogen peroxide in acid, neutral, and alkaline solutions to give IO3- and lower yields of singlet oxygen.11 The emission from this reaction was difficult to record. In our previously work,12 we found that, in strong alkaline solution, a bright luminescence from H2O2/IO4was observed in the presence of carbonate. This observation has been developed for the determination of trace amount of H2O2 in snow water. In this work, the IO4-/CO32-/OH- system was developed for the determination of trace amounts of some polyhydroxyl compounds. Pyrogallol and R-thioglycerol were used as typical samples for the optimization of the flow injection analytical system and the investigation of the CL reaction mechanism. Considering both of IO4- and CO32- are anion, the (1) Evmiridis, N. P. Analyst 1988, 113, 1051-1056. (2) Evmiridis, N. P. Talanta 1989, 36, 357-362. (3) Evmiridis, N. P. Analyst 1987, 112, 825-829. (4) Evmiridis, N. P.; Thansoulias, N. K.; Vlessidis, A. G. Talanta 1998, 46, 179-196. (5) Nakano, S.; Fukuda, M.; Kageyama, S.; Itabashi, H.; Kawashima, T. Talanta 1993, 40, 75-80. (6) Kamidate, T.; Kaneyasu, T.; Segawa, T.; Watanabe, H. Chem. Lett. 1991, 1719-1722. (7) Kamitade, T.; Ichihashi, H.; Segawa, T.; Watanabe, H. J. Biolumin. Chemilumin. 1995, 10, 55-61. (8) Gaikwad, A.; Silva, M.; Perez-Bendito, D. Analyst 1994, 119, 1819-1824. (9) Lin, Q.; Guiraum, A.; Escobar, R. Anal. Chim. Acta 1993, 283, 379-385. (10) Downs, A. J.; Adams, C. J. In Comprehensive Inorganic Chemistry; Bailar, J. C., Emeleus, H. J., Nyholm, S. R., Trotman-Dickenson, A. F., Eds.; Pergamon Press: New York, 1973; Vol. 2, pp 1452-1459. (11) Evans, D. F.; Upton, M. W. J. Chem. Soc., Dalton Trans. 1985, 1141-1145. (12) Lin, J.-M.; Arakawa, H.; Yamada, M. Anal. Chim. Acta 1998, 371, 171176. 10.1021/ac981341m CCC: $18.00

© 1999 American Chemical Society Published on Web 03/19/1999

immobilization of IO4- and CO32- on anion-exchange resins was also studied. A reagentless CL flow sensor for pyrogallol and R-thioglycerol has been prepared. EXPERIMENTAL SECTION Reagents. All reagents were of analytical reagent grade. Water was obtained from a Milli-Q purification system (Millipore). Potassium periodate and pyrogallol (PG) were obtained from Kanto Chemical Co., Inc. (Tokyo, Japan). R-Thioglycerol (TG) and other reagents were the products of Tokyo Kasei (Tokyo). Ionexchange resins, Amberlite IRA-458, -598, and -68 were purchased from Organo (Tokyo). Oxygen and nitrogen gases were of 99.9% purity. KIO4 solution (0.01 M) was prepared by dissolving a 2.3 g of KIO4 in 1 L of water. This solution was stored in a brown bottle to avoid photochemical decomposition. H2O2, PG, and TG solutions were prepared before using. It should be noted that PG is significantly toxic, with the potential for poisoning and death occurring from percutaneous absorption. Apparatus and Procedure. The flow injection analysis system was the same as in our previous work.12 KIO4 solution, KOH/K2CO3 solution, and carrier (H2O) were delivered to a spiraltype flow cell by pumps. A 90-µL loop injector was placed close to the luminometer. The batch method for the CL profile was carried out at a Lumicounter 600 (Microtec NITI-ON, Funabashi, Japan). A F-4010 fluorescence spectrophotometer (Hitachi, Tokyo, Japan) and a Shimadzu UV-2200 UV-visible spectrophotometer were used. CL spectral analyses from 400 to 600 nm were performed using cutoff filters (Toshiba Electric Co.) Immobilizations of periodate and carbonate on anion-exchange resins were carried out separately by suspending the resin particles with 0.04 M periodate solution and 0.1 M KOH/0.4 M K2CO3 solution for 24 h. The exchanged resins were filtered and rinsed with water. Then the same weights of resins exchanged with periodate and KOH/K2CO3 were mixed. A 0.3-g sample of homogeneously mixed resins was packed into a 5.0 mm i.d. × 3.0 cm length glass tube. This flow cell was placed in front of the photomultiplier tube. A SJ1211 peristaltic minipump was used to deliver the carrier (H2O) through a 50-µL loop valve injector. RESULTS AND DISCUSSION Optimal Conditions for the FIA-CL System. Before carrying out the flow injection method, the batch method for the CL profiles was used. As shown in Figure 1, without any special CL reagent, the mixing of 50 µL of a 0.01 M KIO4 solution and 50 µL of 0.1 M KOH/0.1 M K2CO3 solution gave out an evident CL signal. On injection of 50 µL of H2O2, TG, or PG into the above mixing solution, a strong CL emission was recorded. The peak heights of the CL emission were proportional to the concentration of H2O2, TG, or PG. The CL signal from mixing the KIO4 solution with KOH/K2CO3 solution will be the background of the flow injection analysis. To establish the optimal conditions for the flow injection analysis of PG, TG, and other polyhydroxyl compounds, the ratio of the peak height of CL signal to noise (S/N) was measured as a function of the concentrations of KIO4, KOH, and K2CO3 and

Figure 1. CL profiles in the batch system: (1) 50 µL of 0.01 M KIO4 + 50 µL of 0.1 M KOH/0.1 M K2CO3; (2) 1 + 50 µL of 5 × 10-7 M H2O2; (3) 1 + 50 µL of 1 × 10-6 M R-thioglycerol; (4) 1 + 50 µL of 1 × 10-6 M pygrogallol.

the flow rates. In acidic or neutral medium, there was no CL phenomenon recorded. Several basic solutions, for example, 0.01 M Na2CO3/NaOH, 0.01 M K2CO3/KOH, and 0.01 M Na2HPO4/ Na3PO4 adjusted to pH 11.0 by the second chemical of each pair of reagents, were compared. The brightest luminescence was observed in KOH/K2CO3 solution. The S/N ratios corresponding to the concentrations of KIO4, KOH, and K2CO3 were also studied. In the absence of KIO4, no CL was recorded. With an increase in the concentration of KIO4, the S/N ratio was respectively increased up to 0.01 M KIO4 and then decreased slowly. A 0.01 M KIO4 solution was used as one of FIA-CL conditions. The CL intensity was strongly dependent on the concentrations of K2CO3 and KOH. There was difficultly observing the CL by using a fresh KOH solution without addition of K2CO3. Although the higher the concentration of K2CO3 the higher CL intensity, the background and the noise were also increased. The S/N became low when the concentration of K2CO3was over 0.4 M. The concentrations of 0.1 M KOH and 0.4 M K2CO3 were the most suitable for the following experiments. The optimum flow rates of the flow injection were also examined. The flow rates for KIO4 solution and KOH/K2CO3 solution were 1.2 mL/min. The carrier (H2O) flowed at 2.5 mL/min. The satisfactory conditions and the analytical figures of merit for the FIA-CL determinations of H2O2, PG, and TG are listed in Table 1, indicating that these compounds may be determined down to 5 × 10-9-1 × 10-8 M. Especially, the detection limits of H2O2 with those of other CL techniques were compared. Under the optimal conditions, some polyhydroxyl compounds of interest were determined. As shown in Table 2, we found that there was no CL observed for compounds without vicinal hydroxyl groups. The CL intensity was decided by the chemical structure of the polyhydroxyl compound. The polyphenol with vicinal hydroxyl groups gives a relatively strong CL emission. Flow CL Sensor Response Optimization. Although the above flow injection analysis has a high sensitivity and simple operation for determining H2O2, TG, and PG, the use of KIO4 solution and K2CO3/KOH solution was not only a waste of reagents but also could cause pollution of the environment. We Analytical Chemistry, Vol. 71, No. 9, May 1, 1999

1761

Table 1. Optimum Conditions and Analytical Figures of Merit of the FIA-CL System optimum conditions 0.01 M KIO4 of 1.2 mL/min flow rate 0.1 M KOH/0.4 M K2CO3 of 1.2 mL/ min flow rate sample carrier (water) of 2.5 mL/ min flow rate dynamic linear range/M H2O2 pyrogallol R-thioglycerol detection limit (S/N ) 3)/M H2O2 pyrogallol R-thioglycerol relative standard deviation (%) H2O2 (5 × 10-8 M, n ) 14) pyrogallol (5 × 10-8 M, n ) 10) R-thioglycerol (1 × 10-7 M, n ) 10) sample frequency/h comparison of the detection limits of H2O2 with CL techniques luminol/K3Fe(CN)6/pH 10.5 13 fluorophore/1,1-oxalydiimidazole14 luminol/Co2+ 15 KMnO4/octylphenyl poly(glycol ether)/ H+ 16

Table 3. Comparison of Different Ion-Exchange Resins for CL Flow Sensora S/N

5 × 10-9-1 × 10-5 5 × 10-9-1 × 10-7 1 × 10-8-1 × 10-6

resin

TG

PG

H2O2

IRA-458 (strongly basic, gel type) IRA-958 (strongly basic, MR type) IRA-68 (weakly basic, gel type)

68 27 1

96 39 2

102 38 2

a Conditions: sample, 5 × 10-6 M; carrier, 2.0 mL/min water; injection volume, 50 µL.

5 × 10-9 5 × 10-9 1 × 10-8 2.5 4.2 2.0 ∼60 7 × 10-9 M (S/N ) 2) 1 × 10-8 M 5 × 10-9 M (S/N ) 3) 6 × 10-9 M

Table 2. CL Signal of Polyhydroxyl Compounds (1 × 10-5 M)

noted that all of the effective ions, IO4-, OH-, and CO32-, were anions. They could be immobilized on anion-exchange resins. Three different anion-exchange resins, Amberlite IRA-68, -458, and 1762 Analytical Chemistry, Vol. 71, No. 9, May 1, 1999

-958 were tested. As shown in Table 3, with the same conditions, the strongest CL was observed on the Amberlite IRA-458. There was no CL observed by using IRA-68 resin, a weakly basic resin. Although IRA-958 is also a strongly basic anion-exchange resin, its transparency is much poorer than that of IRA-458. Also, the CL signal from IRA-958 is only half that of IRA-458. Several immobilization methods were studied. First, the resin was treated with the mixture 0.01 M KIO4/0.1 M KOH/0.4 M K2CO3. Second, IO4-, OH-, and CO32- were separately immobilized on resins, and then the same weight of these three resins was mixed and packed into the cell. Third, the resins were separately treated with 0.01 M KIO4 solution and 0.1 M KOH/0.4 M K2CO3 solution and then these two resins were mixed in the cell. The sensor prepared with the first method had no CL observed. This result may be due to the consumption of KIO4 in the alkaline/carbonate media. The CL emissions from the second and third methods were almost the same, but the third treatment method was simpler than the second one. The treatment of anion-exchange resin with different concentrations of KIO4 solution and KOH/K2CO3 solution was examined. KIO4 in the concentration range of 0.01-0.04 M gave out a stable signal when the resin was treated for 24 h for all different concentrations. Even the concentration of KIO4 over 0.01 M had almost no effect on the S/N ratio; the higher concentration of KIO4 will cause a longer lifetime of the flow sensor. In this work, in consideration of the solubility of KIO4 in water, a 0.04 M KIO4 solution was used for the sensor preparation. The effect of KOH on the S/N ratio was also examined. KOH in the concentration of range of 0.01-0.1 M showed no CL enhancing contribution to the peak height. However, with an appropriate amount of KOH in the K2CO3 solution, the CL sensor was more stable than that of resin treated with only K2CO3. Resin treated with 0.1 M KOH without K2CO3 also had a CL response. This result could be explained by the small amount of CO2 dissolved in KOH solution. Using a 0.1 M KOH and K2CO3 mixing solution, the CL S/N ratio increased with increase in the concentration of K2CO3 from 0 to 0.2 M and then had almost no change in the range 0.2-0.4 M K2CO3. A 0.1 M KOH/0.4 M K2CO3 solution was used. The CL signal was also dependent on the flow rate of carrier (H2O). The S/N ratio increased at a higher flow rate because the higher flow rate would impact the rate of contact of sample molecules with the ion-exchange resin. The lower flow rate caused broadening of the peak and slowing down of the sampling rates. On the other hand, although the CL signal increased with increasing flow rates, it is not recommended to use flow rates higher than 2.0 mL/min because the high flow rate leads to a

Table 4. Effect of the Dissolved O2 and N2 on the CL Intensitya CL intensity/mV backgroundb

Figure 2. Chemiluminescent signals of CL flow sensor using IRA458 anion-exchange resin. The flow rate of carrier (H2O) was 2.0 mL/ min. Sample, 50 µL of 1 × 10-6 M pyrogallol. The high voltage for photomultiplier was -850 V.

usually carrier (H2O) and KIO4 bubbled with O2 for 5 min all of carrier, KIO4, and buffer bubbled with O2 for 5 min carrier and KIO4 bubbled with N2 for 5 min all of carrier, KIO4 and buffer bubbled with N2 for 5 min

H2O2c

PGd

TGe

0.12 ( 0.01 0.14 ( 0.01

1.50

1.99 ( 0.03 2.20

0.19 ( 0.02

1.85

2.10 ( 0.02 2.44

1.71

1.32 ( 0.03 1.43

0.076 ( 0.005 0.060 ( 0.005

a The experiment was carried out by the flow injection method. The flow system was the same as our previous work.12 b Instead of the sample, 50 µL of 0.1 M KOH/0.4 M K2CO3 solution was injected into the spiral CL cell. The flow line for 0.1 M KOH/0.4 M K2CO3 solution was stopped. c [H2O2], 1 × 10-6 M. d [PG], 2 × 10-6 M pyrogallol. e [TG], 3 × 10-6 M R-thioglycerol.

Figure 3. Calibration graphs for the CL flow sensor. Conditions are the same as in Figure 2.

shortening of the sensor lifetime. Under these conditions, the CL signals from the CL flow sensor using IRA-458 anion-exchange resin were shown in Figure 2. The peak heights of 36 replicate injections of 50 µL of 1 × 10-6 M PG corresponded to a 3.3% relative standard derivation. Figure 3 shows a characteristic calibration graph for the range from 1 × 10-6 to 2 × 10-8 M PG and 1 × 10-6 to 5 × 10-8. M TG. For a new immobilization sensor, the detection limits (S/N ) 3) for PG and TG were 2 × 10-8 and 5 × 10-8 M, respectively. The proposed sensor can be reused at least 200 times for samples of lower than 1 × 10-6 M with a 5% relative standard derivation. Possible Mechanism of the Present CL Reaction. The present work indicated that it was difficult to observe the CL from the reaction of KIO4 with PG or TG without carbonate. The effect of carbonate on the CL signal was evident. Indeed, many CL analyses were carried out using carbonate solution17-24 not only (13) Bostick D. T.; Hercules, D. M. Anal. Chem. 1975, 47, 447-452. (14) Stigbrand, M.; Ponten, E.; Irgum, K. Anal. Chem. 1994, 66, 1766-1770. (15) Price, D.; Worsfold, P. J.; Mantoura, R. F. Anal. Chim. Acta 1994, 298, 121-128. (16) Feng, M. L.; Li, Z.; Lu, J. R.; Jiang, H. L. Mikrochim. Acta 1997, 126, 7376. (17) Lin, J.-M.; Hobo, T. Anal. Chim. Acta 1996, 323, 69-74. (18) Zhang, F.; Lin, Q. Talanta 1993, 40, 1557-1561 (19) Duran, N.; Mansila, H.; Leite, L. C. C.; Faljoni, A. J. Inorg. Biochem. 1988, 34, 105-115. (20) Wierzchowski, J.; Slawinska, D.; Slawinski, J. Z. Phys. Chem. Neue Folge 1986, 148, 197-214.

based on its buffer function but also the CL enhancement. To explain the CL phenomena of the oxidation reaction between KIO4 and polyhydroxyl compounds as well as the CL enhancement of carbonate, the following experiments were carried out. Effect on O2 and N2 in Solution. As shown in Table 4, when the 0.1 M KOH/0.4 M K2CO3 solution, KIO4 solution, and carrier were bubbled with O2 gas, the CL intensities of background and analytes were increased. Otherwise, bubbled with N2 gas, both background and CL signal decreased, most especially, the background emission was only the half of the usual value. These results indicated that the dissolved oxygen took an important role in the CL reaction. But we also found from Table 4 that the signal observed for H2O2 was higher than usual when all solutions were bubbled with N2; this phenomenon may be due to the interaction of H2O2 with the dissolved oxygen. The CL mechanisms of H2O2 with KIO4 and PG (or TG) with KIO4 were different. CL and UV Spectra. The CL spectra from the reactions of H2O2, PG, and TG with IO4- in KOH/K2CO3 solution are shown in Figure 4. There were two peak bands, 436-446 and 471-478 nm. The first one (λ ) 436-446 nm) cannot be ascribed to any component or product of the system investigated. Many investigations18,20-24 have confirmed that CO32- was a luminous species when it existed with a strong oxidant in basic solution. The CL spectral band at 436-446 nm possibly resulted from the decomposition of carbon dioxide dimer to carbon dioxide.22,25 Production of carbon dioxide dimer has been interpreted as a side chain reaction between •OH and/or •O2- radicals and CO32- ions.20 As reported in our recent paper,12 CL emission from mixing the KIO4 solution with KOH/K2CO3 solution was strongly enhanced by the addition of trace amounts of Mn2+, Fe2+, or Fe3+. Other metal (21) Vladimirov, Y. A.; Gavrilov, V. B.; Losev, G. M.; Azizova, O. A.; Olenev, V. I. Zh. Fiz. Khim. 1980, 54, 504-506. (22) Stauff, J.; Bergmann, U. Z. Phys. Chem. Neue Folge 1972, 78, 263-276. (23) Elbanowski, M.; Paetz, M.; Slawinski, J.; Ciesla, L. Photochem. Photobiol. 1988, 47, 463-466. (24) Stawinska, D.; Stawinski, J. J. Biolumin. Chemilumin. 1998, 13, 13-19. (25) Cordes, H. F.; Richter, H. P.; Heller, C. A. J. Am. Chem. Soc. 1969, 91, 7209.

Analytical Chemistry, Vol. 71, No. 9, May 1, 1999

1763

Figure 5. Absorption spectra of pyrogallol solution (upper) and R-thioglycerol solution (lower) in the presence and absence of periodate: reference solution, 0.1 M KOH/0.4 M K2CO3 solution; (1) 5 × 10-4 M pyrogallol in 0.1 M KOH/0.4 M K2CO3 solution; (2) to 3.0 mL of (1), added 100 µL of 0.01 M KIO4 solution; (3) 3 × 10-4 M R-thioglycerol in 0.1 M KOH/0.4 M K2CO3 solution; (4) to 3.0 mL of (3), add 100 µL of 0.01 M KIO4 solution; (5) to 3.0 mL of reference solution, added 100 µL of 0.01 M KIO4 solution.

Figure 4. Chemiluminescent spectra of H2O2, pyrogallol, and R-thioglycerol. The concentrations of H2O2, pyrogallol, and R-thioglycerol were 1 × 10-6, 2 × 10-6, and 5 × 10-6 M, respectively. The flow injection method was used. The flow rates of 0.01 M KIO4 solution and 0.1 M KOH/0.4 M K2CO3 solution were 1.2 mL/min. The carrier (H2O) flow was 2.5 mL/min. The cutoff filter was placed between flow cell and photomultiplier tube.

ions of lower than 10-5 M have almost no effect on this reaction system. The transition metal ions, Mn2+, Fe2+, or Fe3+, were considered as the catalyst of the formation and the decomposition of carbon dioxide dimers in KIO4 solution.20 This simply CL reaction system could be used for determination of Mn2+, Fe2+, and Fe3+ from ∼1 × 10-5 M to the detection limits (S/N ) 3) of 2 × 10-9, 4 × 10-9, and 1 × 10-7 M, respectively. Another CL band (471-478 nm) may be assigned to the transition from the singlet oxygen molecular pair, (O2)2*. Singlet molecular oxygen became a subject of intense laboratory study as a chemical reagent following the interpretation by Khan and Kasha26 of the chemiluminescence of the hypochlorite-oxygen reaction as due to liberated singlet oxygen. During the last 30 years, many CL investigations27,28 concerned with (O2)2* have been reported based on their pioneering work. In our previous work,29,30 the CL wavelength near 480 nm was also recorded. All of them (26) Khan, A. U.; Kasha, M. J. Chem. Phys. 1963, 39, 2105-2106. (27) Wasserman, H. H.; Murray, R. W. In Singlet Oxygen; Academic Press: New York, 1979. (28) Chemistry of Oxygen Species, (Kikan Kagaku Sosetsu); The Chemical Society of Japan: Tokyo, 1990. (29) Wu, X. Z.; Yamada, M.; Hobo, T.; Suzuki, S. Anal. Chem. 1989, 61, 15051510. (30) Lin, J.-M.; Hobo, T. Talanta 1995, 42, 1619-1623.

1764 Analytical Chemistry, Vol. 71, No. 9, May 1, 1999

could be well explained by the formation of the (1O2)2* species. In this work, the formation of (1O2)2* may occur two ways. One was from the recombination of •O2- radicals31 which were the intermediate products of the reaction of IO4- and the dissolved oxygen in basic solution.

IO4- + O2 + 2OH- f 2•O2- + IO3- + H2O

(1)

2•O2- + 2•O2- + 4H2O f (O2)2* + 2H2O2 + 4HO- (2)31

The other way was from the decomposition of the cyclic intermediate of the reducing reaction product of the neighboring hydroxyl compounds. Therefore, in rigid systems, the cis-hydroxyl groups beneficially formed the cyclic intermediate. There was no CL emission from the reduction reaction of polyhydroxyl compounds of only trans-hydroxyl groups. As shown in Figure 5(upper), the UV absorption of PG at 341 and 427 nm decreased after the addition of KIO4. Contrarily, the absorption peak at 232 nm was significantly increased after the addition of a small amount of KIO4 into the PG solution. It means that PG was translated into another compound when it reacted with KIO4. The fluorescence spectra of PG with and without the presence of KIO4 are shown in Figure 6. The FL intensity at wavelength 524 nm was increased with the increase in KIO4 concentration. In the absence of KIO4 in the PG solution, only a very small peak was recorded. The UV-visible spectra of 3 × 10-4 M TG in 0.1 M KOH/0.4 M K2CO3 solution and in 0.1 M KOH/0.4 M K2CO3/3 × 10-4 M KIO4 solution are were in Figure 5 (lower). Their peak heights at 238 nm were significantly different. As with PG, TG was also destroyed in the presence of (31) Afanas’ev, I. B. Superoxide ion: Chemistry and Biological Implications, CRC Press: Boca Raton, FL, 1989.

Figure 7. Stimulating effect of carbonate and formate ions on the chemiluminescence Conditions, 50 µL of 1 × 10-8 M H2O2. The experiment was carried out by the flow injection method and the formate solution was mixed with 0.1 M KOH/0.4 M K2CO3 solution: (1) 0.1 M KOH/0.4 M K2CO3; (2) 0.1 M KOH/0.4 M K2CO3/0.01 M HCOONa; (3) 0.1 M KOH/0.1 M HCOONa; (4). 0.1 M K2CO3/0.1 M HCOONa.

Figure 6. Fluorescent spectra of pyrogallol with and without the addition of periodate: (1) 1.0 mL of 1 × 10-3 M pyrogallol solution + 1.0 mL H2O; (2) 1.0 mL of 1 × 10-3 M pyrogallol solution + 1.0 mL of 0.1 M KOH/0.4 M K2CO3 solution. Spectra 3-7 correspond to the additions of 0.10, 0.20, 0.30, 0.40, and 0.50 mL of 0.01 M KIO4 solution into the 5 × 10-4 M pyrogallol/0.05 M KOH/0.2 M K2CO3 solution, respectively.

formate. Without the presence of dissolved oxygen in the solution, the CL background of KIO4 mixing with KOH/K2CO3/HCOONa solution was very small. This indicates that reaction 4 needs oxygen. In this way, the efficiency of superoxide radical generation could be doubled. Reactions of •OH and •O2- radicals with carbonates lead to the generation of carbonate radicals.20,35

O2- + H2O2 f OH- + •OH + O2

(5)36

OH + HCO3- f OH- + •HCO3

(6)

•

KIO4. For TG, there was no fluorescence recorded by our instrument with or without KIO4. Effect of Sodium Formate on the CL Signal. To identify the effect of carbonate on the CL signal, we also demonstrated that, in the presence of formate, the CL background when KIO4 was mixed with KOH/K2CO3/HCOONa solution was increased. The CL signals of the analytes also improved significantly (Figure 7). The simulating effect of carbonate and formate ions on the CL provided an attractive explanation for the chemiexciation.20 Formate ions were known to react with •OH radicals and, subsequently, with O2 to form •O2- and CO2.20,32-34

Reaction 3 is much faster than reaction 6, which will produce more •O - ions in the solution. In other words, addition of formate could 2 protect the •O2- ions translating to •OH radicals. The recombination of formate ion radicals may generate directly excited oxalate dianions:

HCOO- + •OH f •COO- + H2O

(3)

2•COO- f [(COO)22-]*

(4)

Reactions 3 and 4 are relatively fast. The product of radical chain reactions, •OH radical ions could be effectively reacted with

with ∆E ) 360 kJ mol-1 high enough to promote emission at wavelength 325 nm. Therefore, the addition of formate into the reaction solution caused enhancement of both the CL background and the CL signal; the S/N ratio was almost not affected by the concentration of formate at low concentration. Nitro Blue Tetrazolium Reduction. The reaction for •O2with nitro blue tetrazolium (NBT) was frequently used as a

(32) Rabani, J.; Stein, G. Trans. Faraday Soc. 1962, 58, 2150-2159. (33) Tacconi, N. R.; Wenren, H.; McChesney, D.; Rajeshwar, K. Langmuir 1998, 14, 2933-2935. (34) Goldstein, S.; Czapski, G. J. Am. Chem, Soc. 1998, 120, 3458-3463.

(35) Stauff, J.; Sander, U.; Jaeschke, W. In Chemiluminescence and Bioluminescence; Hercules, D. M., Cormier, M., Lee, J., Eds.; Plenum Press: New York, 1973; pp 131-140. (36) Hodgson, E. K.; Fridovich, I. Arch. Biochem. Biophys. 1976, 172, 202-205.

K ) 3.2 × 107 M-1 s-1 20 •

O2- + HCO3- f HO2- + •CO3-

(7)

(8)

K ) 2.7 × 109 M-1 s-1 at pH 4-10 33 •

COO- + O2 f CO2 + •O2-

K ) (2.0-4.2) × 109 M-1 s-1 34

Analytical Chemistry, Vol. 71, No. 9, May 1, 1999

1765

method for detecting •O2- radicals.37-39 The chemical reactions that produce •O2- can reduce NBT to its deep-blue diformazan pigment. When NBT was mixed with KIO4/H2O2 in alkaline solution, the color change from yellow to blue was very evident. Even the reaction of NBT with •O2- is nonspecific for superoxide, as an assistant detection method, we believed that •O2- ions existed as intermediates in the mixing solution. Based on these results, we suggest the following mechanism for CL reaction of H2O2 with KIO4. In basic solution, H2O2 directly reacts with KIO4 to produce •O - radicals.11 2

IO4- + H2O2 f IO3 + •O2- + H2O

(9)

The recombination of part of the •O2- radicals may generate energy-rich precursors of excited molecules (O2)2*,31 which decompose to O2, and a bright luminescence appeared with a 450nm maximum wavelength. When in the presence of CO32-/HCO3in KIO4 solution, part of the •O2- radicals acts on carbonate and gives free •CO3- radicals.

CO32- + •O2- f •CO3- + O22-

(10)

HCO3- + •O2- f •CO3- + HO2-

(11)

The carbonate-free radicals recombine to generate the excited triplet dimers of two CO2 molecules [(CO2)2*]. With the decomposition of this unstable intermediate to CO2, the energy is released.25,40 The decomposition energy of the (CO2)2 dimer was calculated by the EHMO method and found to be 132 kcal mol-1,41 which is high enough to promote emission at a wavelength higher than 220 nm. The CL spectrum of the 436-446-nm band could be suggested as the result of this energy transportation. When H2O2 was replaced by polyhydroxyl compounds, reaction 1 played a key process in the chain reactions. Both dissolved oxygen and alkaline solution were needed. The polyhydroxyl (37) Bielski, B. H. J.; Shiue, G. G.; Bajuk, S. J. Phys. Chem. 1980, 84, 830-833. (38) Bielski, B. H. J.; Richter, H. W. J. Am. Chem. Soc. 1977, 99, 3019-3023. (39) Takabe, T.; Miyakawa, M.; Nikai, S. Bull. Chem. Soc. Jpn. 1978, 51, 321. (40) DeCorpo, J. J.; Baronavski, A.; McDowell, M. V.; Saalfeld, F. E. J. Am. Chem. Soc. 1972, 94, 2879-80. (41) Bollyky, L. J. J. Am. Chem. Soc. 1970, 92, 3230-32.

1766 Analytical Chemistry, Vol. 71, No. 9, May 1, 1999

compounds were oxidized with the following reactions:

(12)

(13)

We noted that only the two vicinal hydroxyl groups were on the same side, the reaction giving out CL emission. And if the polymer is a fluorescent medium, it may accept the energy from the excited molecular oxygen species, forming the excited organic molecules and generating or enhancing the emission.

21O2* + P f P* + 2O2

(14)

P* f P + hν

(15)

The CL phenomena of the oxidation of PG with molecular oxygen or H2O2 in aqueous solution were thought to be due to the excited singlet oxygen molecules; the mechanism was proposed many years ago.3,42,43 The CL enhanced by carbonate ions may be due to the effective formation of O22- or HO2-, which takes part in the oxidation reaction of PG. The CL wavelength at 510-520 nm (Figure 4b) can be considered as the intermolecular energy transfer. The energy of a part of the excited singlet oxygen molecules is changed to the fluorescence polymer emission. Under the same conditions, if there is no fluorescent compound formation from the reaction of analyte with oxidant, there is of course no fluorescent emitter for the energy transfer. Therefore, the CL spectra of TG and H2O2 were only two peak bands. ACKNOWLEDGMENT This work was partially supported by a grant-in-aid for scientific research (09650894) from the Ministry of Education, Science and Culture of Japan, which we gratefully acknowledge. Received for review December 1, 1998. Accepted February 9, 1999. AC981341M (42) Meluzova, G. B.; Vassil’ev, R. F. Mol. Photochem. 1970, 2, 251. (43) Slawinska, D.; Slawinski, J. Anal. Chem. 1975, 47, 2101-2109.