Delayed chemiluminescence of luminol/ascorbic acid system triggered by ultrasound☆ Authors: Koki Azegami, Fumiki Takahashi, Hirosuke Tatsumi, Jiye Jin Published: 2026 ABSTRACT Graphical abstractUnlabelled ImageKeywords: Delayed chemiluminescence, Sonochemiluminescence, Luminol, Ascorbic acid Introduction Chemiluminescence (CL) has long been recognized as a valuable analytical tool for its high sensitivity, simplicity, and minimal background interference. Among various CL systems, the luminol-based reaction has been extensively studied and widely applied in chemical sensing, bioassays, and environmental monitoring [1], [2]. When coupled with ultrasonic irradiation, luminol can generate sonochemiluminescence (SCL) [3], [4], [5], [6], [7], a phenomenon attributed primarily to the formation of reactive oxygen species (ROS), such as the hydroxyl radical (•OH) and superoxide radical anion (O2•−), during acoustic cavitation. Ascorbic acid (H2A), a well-known antioxidant, is commonly regarded as a quencher in luminol SCL systems because it scavenges reactive oxygen species (ROS) [8], [9], [10]. In this study, we report a novel phenomenon, termed Delayed Chemiluminescence Induced by Ultrasound (DCLIU), observed in a luminol/H2A SCL system subjected to 1 MHz ultrasonic irradiation. Contrary to the typical quenching effect, trace amounts of H2A were found to induce a distinct secondary CL emission in the presence of trace Co2+ and carbonate ion after the cessation of ultrasonic irradiation. This delayed signal is absent in nitrogen-saturated solutions but is enhanced under aerobic conditions, indicating that superoxide anions (O2•−) play a central role in the emission mechanism. The aim of this study is to elucidate the underlying mechanism of DCLIU, focusing on the dual antioxidant/pro-oxidant behavior of H2A in an ultrasonic reaction environment. We propose that during ultrasound irradiation, H2A initially scavenges cavitation-generated ROS, thereby suppressing immediate chemiluminescence. Once H2A is oxidized, its radical intermediate (A•−) facilitates the generation of O2•− in the presence of trace Co2+ as a catalyst, accompanied by the production of hydroperoxide anions (HO2−) during a delayed phase. After depletion of H2A, the luminol/HO2−/Co2+ system produces delayed emission. Furthermore, we investigated the potential of DCLIU as a sensitive and selective analytical method for quantifying ultratrace levels of H2A, demonstrating its suitability for real-sample analysis, including commercial beverages. This study enhances understanding of sonochemiluminescence kinetics and introduces a promising strategy for controlling time-dependent chemiluminescence. Experimental Reagents All reagents were of analytical grade and used as received without further purification. Luminol, sodium carbonate (Na2CO3), uric acid, and glucose were purchased from FUJIFILM Wako Pure Chemical Corporation (Osaka, Japan). L(+)‑ascorbic acid (H2A) and Co2+ standard solution (1,000 ppm) were purchased from Nacalai Tesque (Kyoto, Japan). Caffeic acid was purchased from Tokyo Chemical Industry Co., Ltd. (Tokyo, Japan). All solutions were prepared using distilled water purified with a WS200 distillation system (Yamato Scientific Co., Japan). Apparatus and measurement setup A schematic diagram of the SCL measurement system is shown in Fig. 1. A 1 MHz lead‑zirconate‑titanate (PZT) transducer (diameter: 25 mm, Honda Electronics Co., Ltd., Toyohashi, Japan) was mounted on a stainless‑steel plate fixed at the bottom of a cylindrical quartz cell (inner diameter: 3 cm). The transducer was driven by a sinusoidal signal generated by a WF1974 function generator (NF Corporation, Japan) and amplified by a T145‑5015 power amplifier (THAMWAY Co., Japan). For all SCL experiments, the cell was filled with 20 mL of sample solution containing 0.1 M Na2CO3, 10 μM luminol, trace amounts of H2A, and 1 μM Co2+ as a catalyst. The sample was irradiated directly by PZT. All measurements were conducted inside a light‑proof box to eliminate interference from ambient light. Schematic diagram of miniaturized SCL measurement system using 1 MHz transducer. The SCL signals were detected by a side‑window photomultiplier tube (PMT) module (H6780, Hamamatsu Photonics, Japan) positioned perpendicularly to the cell wall. The PMT signal was recorded with a custom‑built data‑acquisition system controlled by LabVIEW software (National Instruments, USA). A SEC2000 photodiode spectrophotometer (BAS Japan, Tokyo, Japan) was employed for UV–Vis absorption measurements. The peak area was determined by integrating the secondary (delayed) CL signal over the full width of the emission peak. Results and discussion SCL and DCLIU behavior of luminol in the presence of trace H2A Fig. 2 shows the SCL responses from an aqueous solution containing 10 μM luminol, 0.1 M Na2CO3, and 1 μM Co2+ under 1 MHz ultrasound irradiation for 5 s at an electric output power of 10 W. Profiles (a) and (b) depict the response in the presence and absence of 100 nM H2A, respectively. During ultrasound irradiation, a pronounced SCL signal was observed, but its intensity decreased in the presence of H2A, suggesting that cavitation-generated reactive oxygen species were scavenged by H2A. Upon stopping irradiation, the CL intensity rapidly decayed in both cases. Surprisingly, only in the presence of H2A (profile b) was a distinct secondary CL signal observed following a quiescent, delayed period, confirming that this delayed emission, termed DCLIU, is specifically induced by H2A. To investigate the mechanism, we defined the interval between the end of ultrasound and the onset of secondary emission as Period I (delayed region), and the subsequent chemiluminescence as Period II. This division facilitates a stepwise discussion of the underlying reactions in DCLIU. Typical CL response profiles measured in 0.1 M Na2CO3 containing 10 μM luminol and 1 μM Co2+ (a) in the absence and (b) in the presence of 100 nM H2A, respectively. 1 MHz ultrasound was irradiated for 5 s. First, the effect of dissolved gases was investigated. Fig. 3 shows the DCLIU responses in 0.1 M Na2CO3 containing 10 μM luminol, 1 μM Co2+, and 100 nM H2A under (a) c-saturated, (b) air-saturated, and (c) nitrogen-saturated conditions. Each solution was purged with the respective gas for 20 min. The results clearly demonstrate that oxygen is essential for DCLIU activity. In the oxygen-saturated solution, the shortest delay time and highest DCLIU intensity were observed. The response under air saturation was intermediate, whereas no DCLIU signal was detected in the nitrogen-saturated system. We attribute the difference between oxygen and air-saturated conditions to the difference in O2 concentration. During sonication, reactive nitrogen species (e.g., NO2−, NO3−) may form and interact with reactive oxygen species or luminol intermediates. However, under our experimental conditions (with a shorter irradiation duration), the contribution of NO2− appears to be minor compared with the dominant oxygen-derived radical pathways. These findings suggest that molecular oxygen plays a critical role in the formation of reactive oxygen species—most likely superoxide (O2•−) and hydroperoxide anions (HO2•−), which subsequently oxidize luminol during Period II. Effect of dissolved gases on the response of CL. The measurements were conducted in 0.1 M Na2CO3 containing 10 μM luminol, 100 nM H2A, and 1 μM Co2+ after 1 MHz ultrasound irradiation for 5 s. Subsequently, the dependence of SCL and DCLIU intensities on ultrasound output power was examined, and the results are summarized in Table 1. No DCLIU signal was observed in the absence of ultrasonic irradiation, and the intensities of both SCL and DCLIU increased with increasing output power, indicating that DCLIU is an ultrasound-driven process. Higher acoustic power enhances cavitation activity, leading to more cavitation events and increased production of reactive oxygen species, such as •OH and O2•−. For SCL, this directly increases instantaneous chemiluminescence due to enhanced radical generation. For DCLIU, higher acoustic power promotes the oxidation of H2A and facilitates the accumulation of reactive intermediates during Period I. The increased concentration of these intermediates subsequently enhances the delayed CL signals, as discussed later. Table 1Dependence of SCL and DCLIU intensities on the electric output power supplied to the ultrasound transducer.US output power / WSCL intensity / a.u.Maximum DCLIU intensity / a.u.00.0000.00050.3650.079100.6950.107151.1360.123201.3200.179 Dependence of SCL and DCLIU intensities on the electric output power supplied to the ultrasound transducer. Changes in H2A concentration during period I Previous studies have shown that H2A interacts uniquely with ROS (e.g., •OH, O2•−) in aqueous solutions [9], [10]. To obtain kinetic information on H2A activity in Period I, the change in H2A concentration was monitored by UV–Vis spectrophotometry. Fig. 4(A) shows the change in the absorption spectra of a solution containing 0.1 M Na2CO3, 10 μM luminol, 100 μM H2A, and 1 μM Co2+ after 5 s of 1 MHz ultrasound irradiation. In alkaline medium, H2A exists predominantly as the ascorbate monoanion (HA−), which exhibits a characteristic absorption maximum at 266 nm [11]. The absorbance at 266 nm decreased gradually over time and became almost undetectable after 1200 s, indicating the consumption of HA− via its reaction with O2•− and •OH. Fig. 4(B) compares the profile of absorbance at 266 nm with the corresponding CL response as a function of time. The consumption time (1200 s) was consistent with the CL delay time under the same conditions. These results suggest that HA− inhibits the luminol CL reaction [12], resulting in a dark period without light emission (Period I). The delayed emission (DCLIU) would be generated in Period II only after HA− was completely exhausted. It was also noted that the coexistence of trace Co2+ catalyzes the oxidation of HA−, thereby shortening the reaction delay time (see Fig. 1S and Fig. 2S in the Supplementary Data). (A) Time-dependent absorption spectra of ascorbate anion. (B) Relationship between the absorbance at 266 nm and the CL response profile. The measurements were conducted in 0.1 M Na2CO3 containing 10 μM luminol, 100 μM H2A, and 1 μM Co2+ after 1 MHz ultrasound irradiation for 5 s. We selected 5 s of irradiation as a compromise between sufficient ROS production and avoiding excessive heating or effects caused by H2A depletion. Preliminary experiments (now described in the revised manuscript, Section XX) showed that shorter irradiation times (< 3 s) yielded insufficient DCLIU signals, whereas longer irradiation times (>10 s) led to shorter delay times; in particular, at low concentrations of ascorbic acid, the delay time became unobservable. Mechanism of the DCLIU of luminol/H2A/Co2+/CO32– system Accordingly, a plausible mechanism for the reaction in Period I is proposed in Scheme 1(A). In the first step, highly reactive species such as •OH and O2•− are generated by ultrasonic cavitation. HA− acts as an effective antioxidant toward these species through a rapid one-electron transfer from HA− to the ascorbate radical (HA•) [13]. Because the irradiation time is short (5 s), ultrasound acts merely as an initiator; the subsequent reactions continue even after sonication is stopped. HA•, being strongly acidic (pKa ≈ −0.45) [14], rapidly deprotonates to yield the ascorbyl radical anion (A•−) [15]. The ascorbyl radical is relatively stable with E0 ≈ −1.01 V vs. NHE [16], so it can reduce dissolved oxygen (E0 ≈ −0.16 V vs. NHE) to O2•− in the presence of Co2+ as a catalyst, as shown in step 2 in Scheme 1(A). The regenerated O2•− subsequently reacts with HA−, forming the hydroperoxide anion (HO2−) and dehydroascorbate (DHA) in step 3 [17]. Thus, redox cycling between ascorbate and dissolved oxygen can promote further ROS generation, shifting its role from an antioxidant to a prooxidant [18]. The overall reaction sequence culminates in the formation of HO2− as the final reduced oxygen species in Period I. Plausible mechanism of DCLIU in the luminol/H2A/Co2+/CO32− system. Because ROS are preferentially consumed by HA− rather than reacting with luminol, luminol chemiluminescence (CL) was suppressed during Period I. Once HA− is depleted, the CL reaction involving the luminol/HO2−/Co2+/CO32− system becomes prominent, as illustrated in Scheme 1(B). The DCLIU signals were significantly enhanced in the presence of carbonate ions. In Period II, luminol undergoes oxidation through two-electron steps under alkaline conditions (typical pH > 10 in CO32− solution). Co2+ forms a peroxide complex with HO2− [19], producing O2•− and •OH via Fenton-like reactions [20], [21]. CO32− reacts with •OH to generate carbonate radicals (CO3•−), which specifically react with luminol at a rate constant of 9 × 108 M−1s−1 [22]. Consequently, carbonate ions function as co-catalysts in this system. Both •OH and CO3•− oxidize the luminol monoanion (since luminol is a diprotic acid with pKa values of 6 and 13), leading to the formation of the excited dianion 3-aminophthalate (3AP2−*) [23], [24]. Light emission occurs when this excited species relaxes: 3AP2−* → 3AP2− + hν. Thus, trace amounts of Co2+ and concentrated CO32− are essential for DCLIU, as they facilitate the overall chemiluminescence reaction. Additional details on the specific role of carbonate ions are provided in the Supplementary Data (Fig. 3S and Fig. 4S). The radiative decay of the excited state to the ground state produces light, which is observed as the chemiluminescence signal during Period II. We selected an ultrasound frequency of 1 MHz because it provides stable and reproducible cavitation conditions suitable for mechanistic studies. Compared with lower frequencies, a 1 MHz transducer produces more spatially uniform cavitation while minimizing temperature rise, making it particularly advantageous for analytical applications. Analytical performance of DCLIU for trace H2A quantification The potential of DCLIU as a novel analytical method for H2A determination was explored. Fig. 5 shows the dependence of the DCLIU responses on H2A concentration. As the H2A concentration increased, both the delay time and the DCLIU intensity increased. The extended delay time reflects the longer time required to oxidize more H2A. Because HO2− is produced during H2A oxidation in Period I, higher H2A concentrations likely enhance the luminol oxidation in Period II, producing stronger DCLIU signals. A linear relationship was observed between H2A concentration and the integrated DCLIU peak area. The relative standard deviation (RSD) for 50 nM H2A (n = 10) was 3.9%, and the limit of detection (LOD) was estimated to be 1.2 nM, based on a signal‑to‑noise ratio greater than 3. The results demonstrate that the DCLIU enables accurate and sensitive quantification of H2A at trace concentration levels. The selectivity of DCLIU was evaluated by comparing responses from ascorbic acid with those from potential interferents: caffeic acid, uric acid, and glucose (Fig. 6). A significant DCLIU signal was observed only in the presence of ascorbic acid, confirming high selectivity toward H2A. Concentration dependence of H2A on the DCLIU response observed in 0.1 M Na2CO3 containing 10 μM luminol and 1 μM Co2+ after irradiation with 1 MHz ultrasound for 5 s. The peak areas were determined by integrating the delayed chemiluminescence signal in Period II. Specifically, the integration time ranges were as follows: 65.06–532.42 s (10 nM), 58.03–612.49 s (30 nM), 69.81–734.02 s (50 nM), 85.73–808.21 s (70 nM), and 103.19–813.11 s (90 nM). Effect of interferents on DCLIU intensity. Measured in 0.1 M Na2CO3 solutions containing 10 μM luminol, 1 μM Co2+, and 1 μM H2A or an interferent after 1 MHz ultrasound irradiation for 5 s. To evaluate the practical applicability, the proposed method was applied to commercial samples: a lemon drink (labeled H2A content: 1.2 mg mL−1) and vitamin C tablets (167 mg per tablet). The lemon drink was filtered through a 0.45 μm membrane filter and diluted 1000-fold with distilled water. A 100 μL aliquot was then added to a solution containing 0.1 M Na2CO3, 10 μM luminol, and 1 μM Co2+. The tablet sample was dissolved in 100 mL of distilled water, centrifuged to remove insoluble components, and diluted 1000-fold prior to analysis. The results are summarized in Table 2. The concentrations determined by the DCLIU showed good agreement with the labeled values, confirming the method’s accuracy and reliability for real-sample analysis. Table 2Determination of H2A in real samples by DCLIU.Real sampleDCLIURecommended valueRSD (n = 3)Soft drinka)1.15 ± 0.02 mg / mL1.20 mg / mL1.8%Vitamin supplementb)158 ± 4 mg/tablet167 mg/tablet2.7%a) C.C. Lemon (Suntory, Japan).b) Vitamin C 2000 (Minrak Co., Ltd., Japan). Determination of H2A in real samples by DCLIU. a) C.C. Lemon (Suntory, Japan). b) Vitamin C 2000 (Minrak Co., Ltd., Japan). Conclusions In this study, we report a new phenomenon, termed Delayed Chemiluminescence Induced by Ultrasound (DCLIU), in the luminol/H2A/Co2+/CO32− system. Unlike conventional SCL, in which emission decays immediately upon stopping irradiation, this system exhibits a distinct secondary emission after a specific delay. Investigation of the reaction kinetics and mechanism revealed the following key insights: 1) the delay period is governed by the oxidative consumption of H2A. We demonstrated that H2A plays a dual role in this ultrasonic reaction field. Initially, it acts as an antioxidant, efficiently scavenging ROS via one-electron transfer and suppressing immediate luminol chemiluminescence. However, the resulting ascorbyl radical intermediate exhibits pro-oxidant behavior, reacting with molecular oxygen to regenerate O2•−. This radical cycling promotes the accumulation of HO2−, which serves as the oxidant for the luminol CL reaction once H2A is depleted. 2) Dissolved oxygen was confirmed to be essential for the DCLIU phenomenon, supporting the involvement of O2•−. Additionally, Co2+ and CO32− ions are crucial catalysts that sustain the delayed emission. The DCLIU method provides a robust platform for quantifying H2A. Both the delay time (lag time) and the integrated CL intensity exhibit linear responses to H2A concentration over the nM range. The method achieved a low detection limit (1.2 nM) and demonstrated excellent selectivity against common interferents, including caffeic acid, uric acid, and glucose. CRediT authorship contribution statement Koki Azegami: Writing – original draft, Investigation, Formal analysis, Data curation. Fumiki Takahashi: Writing – review & editing, Validation, Investigation. Hirosuke Tatsumi: Writing – review & editing, Validation, Data curation. Jiye Jin: Writing – review & editing, Supervision, Project administration, Methodology, Investigation, Conceptualization. Declaration of competing interest The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: Jiye Jin reports financial support was provided by Shinshu University. The authors declare that there are no additional relationships or activities that could be perceived as a potential conflict of interest. If there are other authors, they declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Footnotes This article is part of a special issue entitled: ‘AOSS-6/JSS-34’ published in Ultrasonics Sonochemistry. The following are the supplementary materials, including additional experimental results related to this article. https://doi.org/10.1016/j.ultsonch.2026.107883. Supplementary data The following are the Supplementary data to this article: Supplementary Data 1 References Associated Data Supplementary Materials Supplementary Data 1