As one of the most promising clean and renewable sources, solar energy is expected to solve energy and environment problems in the future [1]. Due to its natural form of spatial and temporal variability, solar energy must be stored or converted to buffer its intermittency [2]. The photoelectrochemical method that allows storage of solar energy in electrical or chemical form via semiconductors has been regarded as one of the most promising ways toward solar energy utilization [3]. As a key part of the photoelectrochemical reaction system, the study of semiconductor materials for photocathodes has become of great importance, and metal oxides have attracted wide attention as photocathode materials due to their advantages of low cost, environmental friendliness, and unique electrical properties [3,4]. In the family of metal oxides, copper oxide (CuO) is a direct-bandgap semiconductor, which affords it a large absorption coefficient, while its narrow bandgap of 1.1–1.7 eV permits it to absorb more photons close to the visible range [4,5]. These characteristics make it a promising material with high photon-to-current conversion efficiency and applications in photoelectrochemical reactions [6,7]. However, it is reported that oxides of copper lack long-term stability during light illumination, as their photocurrent density decreases with increasing illumination time [8]. This effect for CuO photocathodes has been investigated over the past few years [9,10], and the results indicate that CuO would be reduced following photoelectrochemical reactions, implying that the photocorrosion mechanism in CuO probably originates from the self-reduction reaction due to accumulation of photoexcited electrons [11]. Nevertheless, these findings on the photocorrosion mechanism would be further strengthened if those dynamic chemical events can be real-time detected during the reaction processes.

In the past, ex situ characterizations have been proven to be a rational approach to get a glimpse of trends in conventional electrochemical reactions and to provide a tentative framework for designing reaction pathways, etc. [12]. In the photoelectrochemical reaction process, however, the interfacial semiconductor itself may undergo reactions due to light illumination [13], such as the aforementioned photocorrosion effects [14,15]. These dynamic effects change in real time with illumination duration, making them difficult to capture for conventional ex situ methods [15]. Therefore, aiming at overcoming the obstacles to reveal dynamic chemical events during reaction processes, techniques based on in situ detection have been recently developed [16]. Various detection or characterization methodologies, including UV–Vis spectroscopy [17], X-ray absorption spectroscopy [18], infrared spectroscopy [19], Raman spectroscopy [20], and so on, have been pushed to coupling with electrochemical or photoelectrochemical systems. These methods are convincing for judiciously studying the underlying reaction mechanism; however, some of them deserve high instrument cost, cumbersome apparatus, and the complex design for coupling electrochemical cells. In contrast, using optical fiber sensors to probe interfacial electrochemistry with measurable light–matter interaction can well reflect the dynamic chemical events during real-time electrochemical reactions [21]. Despite the successful reconstruction of optical signals into electrochemical voltammograms in chemical analyses [22,23], biosensing [24,25], and battery monitoring [26,27], such a methodology has never been implemented under a photoelectrochemically driven dynamic material evolution such as the photocorrosion effect.

To fill these gaps, here we propose an in situ fiber-optic detection technique oriented to the dynamic monitoring of photoelectrochemical reactions processes at the interface. A gold nanofilm-coated plasmonic fiber-optic probe was used as the electrochemical electrode, and copper-based nanostructures were in situ electrochemically grown on its interface to develop a lab-on-fiber photoelectrochemical platform [Fig. 1(a)]. The evanescent field of surface plasmon on the gold nanofilm enabled the guided light signal to be sensitive to changes in the electrode–electrolyte interfacial environment [Figs. 1(b) and 1(c)], including the in situ electrochemical growth or conversion of copper and copper oxides. Moreover, the semiconductor characteristics of the interfacial copper oxides enable the functional fiber probe to absorb photons and initiate photoelectrochemical reactions via photogenerated carriers [Fig. 1(d)]. Leveraging the sensibility of the optical signals, the optical fiber is capable of in situ evaluating the photoelectrochemical reactions. By comparing the synchronized optical and electrical responses, we reveal the presence of self-reduction reactions in photocathodes due to the photocorrosion effect. Without the need for complex coupling designs, the methodology enables a combination of affordable fiber-optic sensing techniques in photoelectrochemistry, suggesting its potential in developing an “online monitoring” tool for the bottom-up processing and health conditions of the photocathode to respond to the needs of future energy applications.

The refractive index sensitivity of the fiber-optic device is first measured to verify its detection capability in optical sensing applications. A set of aqueous sucrose solution with refractive index (RI) ranging from 1.333 to 1.378 RIU is used for tests. The fabricated fiber-optic probe device is immersed into the measured liquid, while the resonant wavelength of SPR in the output spectrum is recorded. The obtained correlation curve of the wavelength shift with the change of surrounding RI is shown in Fig. 2(a), showing a high RI sensitivity of 2113.6 nm/RIU. In addition, the fundamental electrochemical properties of the fiber-optic probe as a working electrode are measured. Reversible [Fe(CN)6]3−/4− redox probe is used as electrolyte, and CV tests at different scan rates are conducted. As shown in Figs. 2(b) and 2(c), the voltammograms exhibit two reversible redox peaks which correspond to the oxidation and reduction behavior between [Fe(CN)6]3−/4−. Meanwhile, the peak currents of both anodic and cathodic are proportional to the square root of scan rates, suggesting that there is fast electron transfer kinetics on the optical fiber surface during redox reactions that thus undergoes diffusion-controlled processes. These results demonstrate the multimodal capability of the proposed lab-on-fiber system to host optical detection and electrochemical reactions.

The optical and electrochemical responses of electrochemical growth of copper on the optical fiber surface are investigated. Figure 2(d) illustrates the time-dependent results of the potential, the current, the measured resonant wavelength shift of SPR (Δλ), and the differential wavelength shift with respect to time (d(Δλ)/dt) during periodic potential scanning. The refractive index of the ambient surrounding around the optical fiber surface increases due to the formation of a new material layer (metallic Cu0) on the surface during the cathodic scanning processes, which results in a significant red shift of the resonant wavelength of SPR. Meanwhile, during anodic scanning processes, the processes of redox transitions (Cu0 to Cu2O) involving the change of material electronic structure lead to refractive index perturbations as well as slight shifts of SPR. The presence of wavelength blue shift occurring when the anodic scanning reaches higher potentials can be attributed to the decrease in the refractive index due to the dissolution of the aforementioned materials from the optical fiber surface as a result of electrooxidation. The differential responses [d(Δλ)/dt] calculated from resonant wavelength shift of SPR that characterize the rate of change of refractive index exhibit a pronounced potentiometric correlation with the redox current recorded by electrical signals [28,29]. Nevertheless, due to the limitation of the spectral acquisition rate of the optical spectrum analyzer, the recorded real-time optical response exhibits a slight delay lag with respect to the electrical signals. The measured current versus potential curves are also shown in Fig. 3(a), which present the redox nature of copper ions [30]. The electrochemical behavior is characterized by the presence of two redox pairs, where in the cathodic half cycle, the peak observed at ca. 0 V is associated with the electrochemical reduction of Cu2+ to Cu+, and the peak shown at ca.−0.2V is due to the electrodeposition of Cu+ to metallic Cu0 on the optical fiber surface. Thus, the corresponding two oxidation peaks can be assigned to Cu0/Cu+ and Cu+/Cu2+, respectively. The electrochemical peak current results further confirm that the optical measurements originate from the in situ growth of copper nanomaterials at the interface. Both the wavelength variation and electrochemical peak currents increase with the times of cycle, which suggests that the redox reactions within this potential window are not completely reversible and indicates an increase in copper species at the interface.

After successive scans, it can be found that both the anodic and cathodic peak currents obtained from the voltammograms increase with the times of cycle [Fig. 3(b)], which suggests that the redox reactions within this potential window are not completely reversible due to the accumulation of the Cu-based nanomaterials on the optical fiber surface. Figure 3(c) shows the output spectrum measured at the end of each cycle, which is presented as stacked images, and Fig. 3(d) shows the spectra of cycle 0 and cycle 10 extracted from Fig. 3(c). The results show that the sensitive plasmonic resonance undergoes a progressively larger red shift and the intensity of the resonance diminishes during successive scans. According to the phase-matching condition of surface plasmon resonance, a change in the refractive index will lead to a change in the phase-matching condition, which thus causes a change in the resonance wavelength. Besides, the weakening of the resonance intensity suggests that the optical fiber surface is enriched with highly refractive absorbing materials. Such spectral evolutions are similar to the results reported by Wang et al. [31], in which they measured the nucleation and growth behavior of copper based on plasmonic extinction spectroscopy with an optically transparent electrode. These results can be further rendered as a change in the complex refractive index of the surrounding environment, as shown in Fig. 3(e). The gradually increasing real part and decreasing imaginary part indicate the enrichment of high refractive index/absorption materials on the optical fiber surface. Thus, the optical results demonstrate that consecutive voltametric scanning results in the continuous growth of metallic Cu or Cu2O on the optical fiber surface.

The potential dependence of electrochemical growth of Cu is investigated. Figures 4(a)–4(d) show the chronoamperogram and real-time spectra under pre-set stepped potentials of 0.4 to −0.2V and 0.4 to −0.4V. The spectra are normalized, scaled in logarithmic coordinates, and plotted against time as 2D maps to better illustrate the spectral differences. Figure 4(e) shows the typical spectra measured after applying the potentials of 0.4 V, −0.2V, and −0.4V for 600 s. The results show that even though both potentials are below the deposition potential of copper, it is clear that the spectra change more rapidly at the more negative potential. Besides, when examining the real-time spectra during the first 300 s after the potential step from 0.4 to −0.4V, as shown in Fig. 4(f), it is observed that the measured spectra exhibit a more rapid red shift in resonant wavelength and decrease in spectral contrast, suggesting a more drastic electrochemical growth behavior on the optical fiber surface. Managing the measured spectral changes as changes in complex refractive index variations, as shown in Fig. 4(g), it can be observed that the results are similar to those measured after multiple voltametric cycles. The enhanced spectral change may be attributed to the fact that the application of constant potential does not involve dissolution of Cu0/Cu+ associated with anodic scanning in voltammetry, which causes more metallic Cu material to grow on the optical fiber surface. The electronic structure of the Cu materials is investigated by X-ray photoelectron spectrum (XPS) analysis, as shown in Figs. 4(h) and 4(i). The high-resolution spectrum shows Cu 2p has significantly split spin-orbit components and can be divided into Cu 2p1/2 (952.6 eV) and Cu 2p1/2 (932.7 eV) [32,33]. The weak satellite features can be attributed to the inevitable surface oxidation of sample. Furthermore, the spectral changes during the electrochemical growth of copper nanomaterials at different ambient temperatures are experimentally investigated, as shown in Fig. 4(j). The results show higher temperatures cause a faster red shift of the resonant wavelength. The changes in temperature will lead to faster diffusion rates of the reacting substances in the electrolyte as well as faster fluid convection, which increases the mass transfer rate of the redox molecules to the optical fiber surface [34] and ultimately affects the electrochemical growth behavior as well as the spectral results. These results also demonstrate the capability of fiber-optic sensing to in situ detect the electrochemical growth behaviors at different potentials.

CV behaviors of the optical fiber probe deposited with Cu/Cu2O nanomaterials in alkaline environments are examined in order to conduct the electrochemical conversion of Cu/Cu2O on the optical fiber into semiconductor copper oxides. Typical cyclic voltammograms are shown in Fig. 5(a). In the anodic half cycles, the peak observed at ca.−0.15V is associated with the electrochemical oxidation of Cu+ to Cu2+ [Cu2O to Cu(OH)2 and finally equilibrates to be CuO], while the cathodic peak at ca.−0.3V corresponds to the electrochemical reduction of Cu2+ to Cu+ [35]. In addition, both the anodic and cathodic peak currents increase with the times of cycle [Fig. 5(b)]. The decreasing peak currents for the consecutive voltametric scanning can be attributed to the formation of CuO-based nanomaterial film on the optical fiber surface. The output spectrum from the optical fiber is measured at the end of each cycle, as shown in Fig. 5(c). The results show that the wavelength positions of the SPR-induced dip vary with the times of cycle, while the contrast of the valley remains virtually unchanged, as shown in the zoom-in results of Fig. 5(d). The spectral variation is then rendered as change in complex refractive index [Fig. 5(e)] to illustrate how the electrochemical conversion behaviors occurring at the optical fiber surface affect the plasmonic responses. As documented by Ref. [36], the transition from Cu2O to CuO nanoparticle results in an large increase in the real part of complex refractive index from ca. 2.66 to ca. 2.81 as well as a relatively small increase in the imaginary part from ca.−0.05 to ca.−0.04. Thus, this accounts for the measurement results and suggests that the phenomenon relating to composition transitions between Cu2O and CuO (e.g., subsequent studies in the photocorrosion effect) primarily involves changes in the real part of the complex refractive index.

After series of voltametric treatments, in situ growth as well as conversion of Cu-based nanomaterials is performed on the optical fiber surface. What finally remains on the surface is mostly CuO nanomaterials, which can be demonstrated by characterizing the structure and composition of the optical fiber surface with SEM and Raman spectroscopy. Figure 6(a) shows the SEM image of the optical fiber probe without any electrochemical treatment, in which the Au nanofilm-coated silica waveguide exhibits a bare and smooth surface. In contrast, a dense and uniform film layer is observed on the optical fiber surface deposited with Cu-based nanomaterials, as shown in Fig. 6(b). From the magnified area [Fig. 6(c)], it is observed that the copper oxide nanoparticles are grown on the Au layer on the optical fiber through the electrochemical redox reactions and they are homogeneously distributed on the surface. Due to the high coverage rate, a portion of nanoparticles fused, leading to an increase in particle size and the formation of particle-stacked film. The optical images obtained from these device samples are also shown in the Figs. 3(d) and 5(d), showing the gold color on the surface of the bare optical fiber probe, the red color on that with Cu/Cu2O nanomaterials, and the black color on that with CuO nanomaterials, respectively. Figure 6(d) shows the Raman spectra of the bare optical fiber probe and that deposited with Cu-based nanomaterials. There are no Raman scattering signals in the spectrum measured from the bare probe device, while the spectrum measured from the device with nanomaterials exhibits characteristic phonon frequencies of the CuO [37]. The strong Raman peaks at 278cm−1 are attributed to the A_g mode of CuO. The characteristic peak at 602cm−1 and the weak peak at 366cm−1 originated from the B_g mode of CuO. The weak peak at 506cm−1 is ascribed to a minor Cu2O phase. These results demonstrate the realization of bottom-up fabrication of CuO nanomaterials on the optical fiber by means of electrochemical methods and elucidate that the spectral variations of fiber-optic detection during the fabrication processes originate from such electrochemical growth and conversion processes of Cu-based nanoparticles.

Then, the photoelectrochemical responses of the fabricated lab-on-fiber CuO photocathode are investigated. The photocurrent measurements under chopped illumination are conducted to further examine the photoresponses of the CuO photocathode. As shown in Fig. 6(e), the current falls and rises rapidly when the illumination from the xenon lamp is turned on and off, respectively, which indicates that the fabricated CuO nanoparticle possesses significant cathodic photocurrent properties of the p-type semiconductor. In addition, there is a significant decay in the photocurrent of the CuO photocathode after the chopped illumination persists over time. The XPSs [Fig. 6(f)] of the CuO photocathode with different reaction times under continuous illumination show both the 2p peak (954.2 and 934.3 eV) and satellite peak (962.6 and 943.8 eV) attributed to Cu2+ decreased significantly, accompanied by the enhancement of the peaks corresponding to Cu+ at 952.6 and 932.8 eV and the chemical shift in Cu LMM Auger peaks, which demonstrate the change of valence state during reaction [32,33]. The electrochemical potential for the reduction of CuO to Cu2O/Cu is located within the CuO bandgap, which probably enables the photoelectron generated during the light illumination to cause the self-photoreduction reactions of the CuO photocathode, as described by Eqs. (1) and (2) [11,38]: 2CuO+H2O+2e−↔Cu2O+2OH−,(1)Cu2O+H2O+2e−↔2Cu+2OH−.(2)

To explore such photoinduced corrosion behavior, the output plasmonic spectrum from the optical fiber is recorded simultaneously during the photoelectrochemical tests. Figure 6(g) shows the temporal spectra under chopped illumination with increasing period. The results show that when the illumination is turned on, there is first jumping photocurrent originated from the photoelectrons, followed by a gradual loss in the photocurrent due to photoinduced redox reactions of the CuO photocathode. Meanwhile, the measured SPR dip in spectrum varies with the times of illumination and, intriguingly, shifts in the opposite direction after the illumination is turned off. The blue shift of SPR during illumination can be attributed to the self-photoreduction reaction, in which the CuO nanomaterials on the optical fiber surface are reduced to Cu2O and metallic Cu. As is known from the previous subsections, these transitions cause a significant decrease in the real part of the refractive index. Besides, it is reported that Cu2O is quite unstable even in the absence of light [39]: in the potential window of 0.6–0.85 V versus reversible hydrogen electrode (RHE), it can be electrochemically oxidized to CuO in aqueous electrolytes [i.e., the inverse reaction of Eq. (1)]. Since the bias voltage in the photoelectrochemical tests here is set to 0 V versus SCE (i.e., 0.6545 V versus RHE), the Cu2O nanomaterials generated on the optical fiber surface are subsequently self-oxidized to CuO after the light is turned off, which caused the slight increase in the refractive index and thus the red shift of SPR.

Further, the measured SPR spectrum as well as the current during self-photoreduction and self-oxidation processes is analyzed separately for each on/off light cycle. Figure 6(h) shows the shift of resonant wavelength of SPR during three consecutive on/off light cycles, where the wavelength shifts during the light illumination and no-illumination processes are presented independently. The results show that the value of wavelength blue shift of the light illumination is significantly larger than that of wavelength red shift of the no-illumination process during the first cycle, which suggests that the redox reactions caused by the photocorrosion effect are not completely reversible. Meanwhile, such difference decreases with times of cycle. After successive cycles, the wavelength shifts induced by light illumination significantly decrease, while the wavelength shift of no-illumination increases and finally approaches the same level as that of light illumination during the third cycle. Such electrochemical and optical results are summarized in Figs. 6(i) and 6(j), which show the average photocurrent and average SPR wavelength reduce with increasing times of cycles, indicating the light illumination does cause the CuO nanomaterials on the optical fiber surface to be continuously and irreversibly corroded. Besides, Figs. 6(i) and 6(j) show that the loss of photocurrent and the wavelength shift per cycle decrease with increasing number of cycles. This can be further rendered as the weakening of the self-photoreduction reactions and the enhancement of the self-oxidation reactions, which eventually tend to an equilibrium and manifest as a gradual stabilization of the decaying photocurrent. These findings potentially account for the results regarding decaying but gradually stabilizing photocurrent of previous reports [11] and potentially provide a “fingerprint” of the photocorrosion processes of CuO.

In summary, we propose an in situ fiber-optic detection technique for dynamic monitoring of interfacial nanomaterial growth, conversion, and photoelectron-chemical reaction processes. By using a gold-coated optical fiber as both the electrochemical electrode and the optical sensing probe, which enable tracking the electrochemical and photoelectrochemical events on the optical fiber surface with the evanescent field of surface plasmon, a lab-on-fiber photoelectrochemical platform is realized. The experimental results of the electrochemical growth of copper indicate the formation of a new material layer (metallic Cu0) on the fiber surface during cathodic scanning, in which the refractive index of the new metal-based medium layer increases compared to the original aqueous electrolyte-based medium layer and thus results in a red shift of the plasmonic resonant wavelength via the change in phase-matching condition. Meanwhile, the high extinction coefficient (i.e., the imaginary part of complex refractive index) of the metal nanomaterials will also cause the intensity of the resonance to change significantly. These optical parameter variations accompanying the reaction process thereby endow the optical fiber with the capacity of accurately and in situ tracking the dynamic processes including in situ electrochemical growth (Cu2+ to Cu/Cu2O) and in situ electrochemical conversion (Cu/Cu2O to CuO). Further, by leveraging the sensibility of the plasmonic resonance signals to the redox processes of copper species, we performed “online monitoring” of the photoelectrochemical reaction processes of CuO photocathode under light illumination. Based on the synchronized optical and electrical responses, we demonstrate the presence of self-photoreduction reactions in CuO nanomaterials (CuO to Cu/Cu2O) induced by photogenerated electron, which leads to a corrosion of the photocathode and to a decay in the photocurrent. Meanwhile, we unveil the presence of self-oxidation reactions in Cu2O (Cu2O to CuO) during non-illumination processes, which strengthens after the continuation of light illumination and moderates the decay of photocurrent caused by the photocorrosion effect. These results provide us with a “fingerprint” of the bottom-up fabrication and photocorrosion process of CuO nanomaterials, which facilitates the online monitoring and diagnosis for the processing and health condition of photocathode. In future work, the limitation of the sampling rate of the spectrum analyzer can be addressed to track the dynamic information of electrochemical reaction more accurately, for example, by using a high-speed acquisition oscilloscope for intensity interrogation since the refractive index change is accompanied by both a shift of resonance wavelength and a change of the light intensity.

Sodium chloride (NaCl) and cupric chloride dihydrate (CuCl2·2H2O) were provided by Aladdin Chemistry Co., Ltd. Potassium chloride (KCl) and potassium ferricyanide (K3[Fe(CN)6]) were provided by Macklin Biochemical Co., Ltd. All chemicals were analytical grade and used as received without any further purification. 5 mM (1 M = 1 mol/L) K3[Fe(CN)6] in 0.5 M KCl solution, 1 and 2 mM CuCl2 in 0.1 M NaCl solution, 0.1 M NaOH, and 0.1 M NaCl solution were prepared by adding ultrapure water from a water purifier (Thermo Scientific).

A fiber-optic opto-electrochemical platform was used in the experiment. The platform is similar to a traditional three-electrode electrochemical system, but different in that the working electrode is replaced by a fiber-optic probe. The optical fiber probe consists of a silica waveguide (HP 600/630-37/750 E, YOFC) with an Au nanofilm coated on the surface via an ion sputtering system (SBC-12, Beijing KYKY Technology Co. Ltd.), which is bent into a U-shape to form an inserted probe device [40]. Optimized gold nanofilm thickness (ca. 45 nm) and probe device length (ca. 25 mm) improve sensitivity and spectral resonance contrast. Au is a relatively corrosion-resistant material and is able to effectively excite surface plasmon resonance. It has been demonstrated to be effective in performing electrochemical reactions as it has low electrical impedance properties comparable to commercial electrodes [40]. Therefore, it was chosen as the coating material for the optical fiber in this work. Nevertheless, there are materials that can be used as alternatives to Au, such as indium tin oxide, which has low resistivity and high corrosion resistance and can also excite the lossy-mode resonance, which has been demonstrated in previous reports [23,24]. The interrogation optical signal from a BBS (HL-2000, Ocean Insight Inc.) is launched into the optical fiber, in which the signal carries plasmonic spectrum modulated with ambient information after passing through the probe device and is eventually collected by an OSA (AQ6373, Yokogawa). The potential of the probe device is controlled by a CHI electrochemical workstation (CHI 660E, Shanghai Chenhua), in which the wires from the working electrode end are connected to the gold film of the probe device via silver paste. The counter electrode and reference electrode of the electrochemical test system are a platinum sheet electrode (purity: 99.99%, Taizhou Zenno Material Technology) and a saturated calomel electrode (SCE, CHI 150, Shanghai Chenhua), respectively. The electrochemical and optical signals are analyzed simultaneously in real time using a computer.

Cyclic voltammetry (CV) tests for electrochemical growth of Cu are conducted in the electrolyte of 2 mM CuCl2 in 0.1 M NaCl solution at a scanning rate of 50 mV/s. CV experiments are performed over a potential range of −0.4V to 0 V, which covers the redox window from Cu2+ to Cu0/1+. The stepped potential ranges for the chronoamperometry are pre-set to +0.4V to −0.2V and +0.4V to −0.4V. CV tests for electrochemical conversion of CuO are conducted in the electrolyte of 1 M NaOH solution at a scanning rate of 50 mV/s. The selected potential range is −0.5 to 0.1 V, which covers the redox potential window from Cu0/1+ to Cu2+. Finally, the structure and composition of surface nanomaterials are characterized by scanning electron microscopy (SEM, Apreo S, Thermo Scientific), Raman spectroscopy (DXR, Thermo Scientific), and X-ray photoelectron spectroscopy (XPS, K-Alpha, Thermo Scientific).

Photoelectrochemical tests coupled with in situ fiber-optic detection are performed in a custom-made quartz glass cell. The fiber-optic probe device grown with CuO nanomaterials obtained in the previous experiments is used as a photocathode in the electrolytic cell here. A 300 W xenon lamp (CEL-HXF300, Beijing Au-light) is used to simulate natural sunlight during photoelectrochemical reactions. First, the time-current curve tests are conducted in the electrolyte of 0.1 M NaCl solution at a constant potential of 0 V under fast chopped lighting with a period of 2 s in order to investigate the photocurrent-time response of the optical fiber-based CuO photocathodes. Subsequently, the time-current curve tests are performed under slow chopped lighting with a period of 600 s. The longer illumination duration is to amplify the photocorrosion effect that occurs during the photoelectrochemical reaction processes in order to demonstrate the changes in the photocurrent response as well as the fiber-optic detection results. After 30 min, the average photocurrent and resonant wavelength during the light-on period are taken over three on/off light cycles, as well as the photocurrent reduction and wavelength shift results.