With increasing environmental pollution and depletion of traditional fossil energy reserves, it is imperative to explore new renewable energy resources to power a sustainable society and technology development. Semiconductor derived solar driven photocatalytic hydrogen production, which was discovered by Honda and Fujishima in 1972 [1], has shed light on one way to solve this challenge. Therefore, photocatalytic hydrogen production is considered one of the more promising technologies to provide renewable energy (i.e., hydrogen as an energy carrier) in an environmentally friendly approach. In the past decade, various photocatalytic semiconductors have been discovered by researchers, such as metal oxides (e.g., TiO2 [2, 3], CuO [4]), and metal sulfides (e.g., CdS [5, 6], ZnS [7, 8], ZnIn2S4 [9], and Zn0.5Cd0.5S [10, 11]). Compared to the reported inorganic photocatalysts, organic photocatalysts based on conjugated frameworks have attracted widespread attention recently because of their intrinsic advantages: low production cost, nontoxicity, and potential for functionality via molecular design. Among the novel organic photocatalysts now being developed, g-C3N4 exhibits the most excellent photocatalytic activity for H2 production [12-15], water oxidation [16, 17], organic pollutant degradation [18, 19], and CO2 reduction [20, 21]. However, pure g-C3N4 alone is not an ideal photocatalyst and shows poor photocatalytic activity. This is mainly attributed to the rapid recombination of the charges photogenerated. That is to say, much needs to be done in this field to improve further the photocatalytic activity of g-C3N4.
Compared to other strategies, co-catalyst modification provides a straightforward and efficient way to enhance the photocatalytic activities of photocatalysts. It is well known that noble metals (such as Pt [22], Pd [23], Ag [24, 25], and Au [26]) are effective co-catalysts for efficient photocatalytic hydrogen production. With the ability to tune the electronic structure at atomic scale provided by noble metal alloy co-catalysts (such as AuPd [27], PtCo [28] and NiCo [29]), they have attracted great attention for their potential to achieve more satisfactory photocatalytic performance. Furthermore, arising from the synergetic effect between different active sites, bimetallic alloy nanoparticles (NPs), (composed of two different metal elements) usually exhibit superior electronic, optical, and photocatalytic properties. Such properties often cannot be achieved in the corresponding monometallic nanoparticles [30-32]; therefore, bimetallic alloy NPs offer the potential for more efficient co-catalysts [33, 34].
Herein, we present a new strategy using the chemical deposition precipitation method to synthesize PtPd/g-C3N4 photocatalyst, which exhibits highly efficient visible-light activated H2 production from water. The prepared PtPd/g-C3N4 composite photocatalysts showed remarkably enhanced photocatalytic H2 evolution activity compared not only with monometallic Pt/g-C3N4 and Pd/g-C3N4, but also with pure g-C3N4. After full optimization of the loading ratio of PtPd alloy NPs, a photocatalytic H2 evolution rate of 1600.8 μmol g–1 h–1 was achieved with a 0.2 wt% PtPd/g-C3N4 composite photocatalyst. This is 3.55 times higher than with Pt/g-C3N4 and 1.34 times higher than with Pd/g-C3N4. In addition, the most extraordinary result was that H2 rapid precipitation occurred when K2HPO4 was added to the reaction system. To investigate the mechanism that enhances H2 evolution activity in the PtPd/g-C3N4 composite photocatalyst, its structural characteristics and optical properties were analyzed. The photocatalytic H2 evolution activity and stability of PtPd/g-C3N4 under visible light irradiation were also measured. Based on the characterization results and the photocatalytic activity, a possible photocatalytic mechanism for the enhanced H2 evolution activity was proposed. Our experimental results demonstrate that PtPd alloy NPs can be effective co-catalysts for g-C3N4. This means that they may also have potential benefits for other semiconductor photocatalysts for H2 evolution.
All reagents were of analytical grade and were used without further purification. The g-C3N4 nanosheets were prepared according to a reference, with slight modification [35]. Two-dimensional (2D) g-C3N4 nanosheets were prepared by heating 10 g of urea in a covered crucible at 550 ℃ for 4 h with a ramping rate of 5 ℃/min. After that, the g-C3N4 nanosheets were fabricated by thermal oxidation etching of as-obtained bulk g-C3N4 powders at 500 ℃ for another 2 h under the same condition. After heating and cooling, a light-yellow powder product was obtained. Then, the powders were washed in deionized water with centrifugation several times. The final products were obtained by drying them at 60 ℃ for 8 h in a vacuum oven. The as-synthesized g-C3N4 was subsequently used to prepare the PtPd/g-C3N4 nanocomposites.
Preparation of the PtPd/g-C3N4 composite catalysts proceeded as follows. The 0.5 g of g-C3N4 nanosheet powder was dispersed ultrasonically in 40 mL of ethylene glycol for up to 1 h. The product was used as stock solution. Subsequently, a certain weight ratio of 2 g L–1 H2PtCl6 (Pt) and 4 g L–1 PdCl2 (Pd) were added to the stock solution with continuous stirring (Pt:Pd = 0.5:0.5). After 5 min, 500 μL of 80% hydrazine hydrate was added to the reaction mixture and stirring was continued for another 4 h at room temperature. The product was washed and centrifuged with deionized water and ethanol several times, respectively, and then dried at 60 ℃ for 8 h in a vacuum oven [36]. The obtained samples of PtPd/g-C3N4 were further heated at 300 ℃ for 2 h under Ar flow in a tube furnace [37]. Finally, PtPd/g-C3N4 composite photocatalysts with different weight ratios were obtained. The weight percentages of the PtPd alloy NPs in the prepared composite photocatalysts were 0.0, 0.1, 0.2, 0.5, 1.0, and 2.0 wt%, respectively. Next, PtPd/g-C3N4 with other Pt/Pd ratios was synthesized under the same experimental conditions and same methods. These included Pt/g-C3N4, Pd/g-C3N4, Pt0.9Pd0.1/g-C3N4, Pt0.7Pd0.3/g-C3N4, Pt0.3Pd0.7/g-C3N4, and Pt0.1Pd0.9/g-C3N4 composite photocatalysts with 0.2 wt% loading.
An X-ray diffractometer (XRD) using Cu Kα as radiation was applied to investigate the crystallographic texture of all samples, with 2θ ranges from 5° to 60°. Transmission electron microscopy (TEM) and high-resolution transmission electron microscopy (HRTEM) were carried out using an FEI Tecnai G2 F20 microscope. UV-vis diffuse reflection spectroscopy (DRS) of all samples was detected using a Hitachi UV-4100 spectrophotometer, and BaSO4 was used as the reflectance standard. X-ray photoelectron spectroscopy (XPS, Thermo ESCALAB 250XI) was applied to analyze the element-valence state of the samples. The signal of carbon at 284.8 eV was used as a reference to calibrate the binding energies. A photoluminescence (PL) spectrum was provided by an FLS980 Series of fluorescence spectrometer.
The photocatalytic H2 evolution test was performed in a Perfectlight Labsolar IIIAG system equipped with a 300 mL quartz reactor. The temperature of the system was maintained at 4 ℃ by connecting it to a low-temperature thermostat in a bath of circulating water. The employed visible light source was a PLS-SXE 300W Xe arc lamp combined with a UV-cutoff (≥ 400 nm) filter. The light source was positioned 15 cm away from the reactor to trigger the photocatalytic reaction. The irradiation light intensity selected was 32 mW/cm2 and was confirmed by a radiometer. In a typical procedure, 50 mg of catalyst was dispersed in 100 mL of mixed aqueous solution containing 10 mL of triethanolamine (TEOA) and 90 mL of deionized water. This was subjected to constant ultrasonic treatment for about 5 min before the reactor was installed in the system. In addition, 10 g of K2HPO4 was added to the above solution to promote photocatalytic performance. Moreover, before irradiation, the system was evacuated for at least 0.5 h to remove dissolved oxygen and ensure an anaerobic condition. The mixture was magnetically stirred meanwhile. The gas produced during the reaction was detected by a gas chromatograph (Beifen 3420A, China) using high purity argon as the carrier gas. The chromatograph was equipped with a 420 ± 5 nm band pass cut filter. The apparent quantum yield (AQY) was calculated using Eq. (1):
The surface photovoltage (SPV) measurements were carried out on a surface photovoltage spectroscope (PL-SPS/IPCE1000, Beijing Perfect Light Technology Co., Ltd.). The measurement system included a source of monochromatic light, a lock-in amplifier (SR830, Stanford Research Systems, Inc.), a light chopper (SR 540, Stanford Research Systems, Inc.), and a sample cell. The measurements were done under ambient conditions.
The photocurrent measurements were conducted on an electrochemical workstation (CHI, China) in a standard three electrode framework with a 0.5 mol/L Na2SO4 electrolyte solution. A Pt sheet was used as the counter electrode and an Ag/AgCl electrode as the reference electrode. The work electrode was prepared as follows. The photocatalyst (10 mg) was dispersed in a solution of 500 μL of absolute ethanol and 15 μL of Nafion solution. The obtained slurry was then dripped onto pretreated FTO glass with a controlled area of 1 cm x 1 cm. It was then dried in air at 150 ℃ for 2 h to form a film electrode. Photocurrents were obtained using a 300 W xenon arc lamp with light passing through an AM 1.5 G filter.
The chemical composition and crystal structure of the as-prepared photocatalysts were investigated by X-ray diffraction (XRD), and the results are shown in Fig. 1. The pure g-C3N4 nanosheets have two diffraction peaks located at 12.8° and 27.4°, corresponding to the (100) and (002) lattice planes of g-C3N4, respectively [38]. Aside from the diffraction peaks of g-C3N4, the PtPd/g-C3N4 sample has two additional diffraction peaks at 40.1° and 46.6°. These were indexed to the (111) and (200) crystal planes of PtPd (JCPDS 65-6418), indicating the existence of PtPd alloy. The co-existence of PtPd and g-C3N4 in the as-prepared PtPd/g-C3N4 composites indicates that the PtPd/g-C3N4 composites were successfully synthesized. The clear crystal diffraction pattern of g-C3N4 demonstrates that its crystal structure was not destroyed during the coupling of the PtPd phase.
To confirm the presence of PtPd NPs in the PtPd/g-C3N4 composite sample, the morphology and crystallography information of the samples were investigated using TEM and HRTEM. From the TEM images shown in Fig. 2, its can be seen that the pure g-C3N4 nanosheets are thin film and that the PtPd/g-C3N4 composite sample has a typical thin film morphology with black nanoparticles distributed on the surface. The thin film observed in TEM can be attributed to g-C3N4 due to its intrinsic 2D molecular structure. To determine the composition of the black nanoparticles observed on the surface of g-C3N4, a close investigation of the sample was done using HRTEM. When the HRTEM focus was on the black nanoparticles, a set of clear lattice fringes could be seen in the image, from which a lattice spacing of 0.225 nm was determined. The lattice spacing revealed by HRTEM matches well with the interplanar crystal spacing of the (111) crystal plane of the PtPd alloy (JCPDS 65-6418), which provides additional support for the successful synthesis of the PtPd/g-C3N4 composites. Fig. 2(d) shows the element-distribution maps of Pt and Pd, illustrating the composition of Pt and Pd in the bimetallic catalysts.
X-ray photoelectron spectroscopy (XPS) was further utilized to investigate the surface chemical composition and electronic states of the PtPd/g-C3N4 composite sample. The XPS spectra of C 1s, and N 1s, Pt 4f, and Pd 3d are shown in Fig. 3. The results of the survey scan of the PtPd/g-C3N4 composite show the presence of C, N, Pt, and Pd in the composite (Fig. 3a). Fig. 3(b) displays the XPS spectra of C 1s: the two peaks located at 284.8 and 288.4 eV were related to C–C and N–C=N, respectively [27]. There were no differences between the pure g-C3N4 nanosheets and the PtPd/g-C3N4 composite sample, which means that there was no obvious effect on their chemical bonding after loading the PtPd alloy NPs onto the surface of the g-C3N4 nanosheets. The N 1s XPS spectra were deconvoluted into three peaks at 398.51, 399.25, and 400.65 eV (Fig. 3c and 3d), which represent sp2-bonded N in C=N–C, tertiary nitrogen N–C3, and nitrogen, respectively, in the C–N–H functional group in the polymeric g-C3N4 structure [18, 28]. A slight shift can be observed for N 1s toward higher binding energy compared with pure g-C3N4, which is attributed to changes in the surface electron density. The XPS spectra of Pt 4f for the Pt/g-C3N4 and PtPd/g-C3N4 sample are shown in Fig. 3(e). The Pt 4f XPS spectra have two peaks at 70.62 and 73.92 eV in the Pt/g-C3N4 sample, which correspond to Pt 4f7/2 and Pt 4f5/2, respectively [39]. However, the PtPd/g-C3N4 sample shows two peaks at 71.45 and 74.74 eV in the Pt 4f XPS spectra, which exhibit a positive upshift (~ 0.83 eV) compared to that of the Pt/g-C3N4 sample. This positive upshift in the binding energies for Pt 4f can be attributed to the alloying effect between Pt and Pd. Similarly, in Fig. 3(f) the two Pd 3d XPS peaks in the PtPd/g-C3N4 sample correspond to Pd 3d5/2 at 336.20 and Pd 3d3/2 at 341.35 eV, and also show a 0.83 eV shift compared to the Pd/g-C3N4 sample (335.37 and 340.52 eV) [37]. From the shift of binding energy for Pt 4f and Pd 3d in the PtPd/g-C3N4 sample, we can conclude that the PtPd alloy NPs were successfully synthesized and deposited on the surface of the g-C3N4 nanosheets.
The UV-vis diffuse reflectance spectra (DRS) of the as-prepared samples were recorded to investigate their optical properties and the results are shown in Fig. 4. Compared to pure g-C3N4, the optical absorption of the PtPd/g-C3N4 composite samples was substantially enhanced in the wavelength range 350–800 nm. This is consistent with the change of color observed in the prepared samples from yellow (pure g-C3N4) to gray (PtPd/g-C3N4) [11]. With increasing loading ratio of PtPd alloy NPs, the absorption of PtPd/g-C3N4 in the visible range also increased. The enhanced light absorption efficiency of these composite photocatalysts indicates that they have the potential to harvest more photons to drive photocatalytic reactions, which would be expected to deliver improved photocatalytic activity [38]. The band gap of the as-prepared samples was derived using Eq. (2):
where A is the proportionality constant, h is Planck's constant, ν is the incident light frequency, and Eg represents the band gap energy. The value of n is related to the type of band gap. Here, a value of '1' was used for n because g-C3N4 is reported to have a direct band gap [35]. The band gap of pure g-C3N4 was estimated to be 2.69 eV using Eq. (2), as shown in the Fig. 4 inset. This is consistent with the results in previous studies [27, 37]. The DRS results indicate that the PtPd alloy NPs chemically deposited in situ had considerable effect on the g-C3N4 and resulted in more efficient utilization of visible light.
The photocatalytic H2 evolution activities of the prepared samples were characterized and the results are shown in Fig. 5. The pure g-C3N4 sample shows a relatively low H2 production rate of 2.2 μmol g–1 h–1. After loading the PtPd alloy NPs onto the surface of the g-C3N4 nanosheets, the photocatalytic activity of the sample for hydrogen production was significantly enhanced. After loading only 0.1 wt% PtPd alloy NPs, the rate of H2 production was remarkably increased to 1207.1 μmol g–1 h–1: nearly 600 times higher than that of pure g-C3N4. Along with the gain from increase of the PtPd alloy NP weight ratios, the H2 evolution activity of the PtPd/g-C3N4 composite sample was also improved when the weight fraction of the PtPd alloy was increased to 0.2 wt%. The 0.2 wt% PtPd/g-C3N4 sample gave the highest H2 evolution activity (rate of 1600.8 μmol g–1 h–1), 800 times higher than that of pure g-C3N4. It can be seen that deposition of the PtPd alloy NPs can remarkably improve the hydrogen evolution activity of pure g-C3N4. However, further increase in the deposition ratio of PtPd alloy NPs to higher than 0.2 wt% results in a decrease in the H2 production rate. The degraded photocatalytic H2 evolution activity might be caused by the light shielding effect of the excess black PtPd alloy NPs, which would attenuate the light intensity at the surface of the g-C3N4. To explore the conditions needed to provide the greatest potential of the PtPd/g-C3N4 composite sample for photocatalytic H2 evolution, K2HPO4 was added to the electrolyte as a sacrificial agent. When 10 g of K2HPO4 was added to the photocatalytic system, the H2 production rate of the PtPd/g-C3N4 composite sample reached 2885.0 μmol g–1 h–1. The explanation of this enhancement is that HPO42– can act as a mediator that directly takes part in the photocatalytic H2 production [27, 40]. From Table 1, it can be noted that better hydrogen production rates were obtained in this study than with the other monometallic and bimetallic photocatalysts reported in the literature.
In addition, the atomic ratio between Pt and Pd in the PtPd alloy NPs greatly influenced the photocatalytic activity of the PtPd/g-C3N4 composite sample. As shown in Fig. 6, the rate of H2 production was only 450.1 μmol g–1 h–1 for the 0.2 wt% Pt/g-C3N4 sample. By increasing the atomic ratio of Pd in the PtPd bimetallic alloy NPs, the photocatalytic hydrogen evolution of PtPd/g-C3N4 was significantly enhanced. The optimal atomic ratio of Pt:Pd in the PtPd bimetallic alloy NPs to achieve the highest photocatalytic activity in the PtPd/g-C3N4 composite sample was demonstrated to be 1:1. The 0.2 wt% Pt0.5Pd0.5/g-C3N4 sample showed the highest H2 evolution rate (1600.8 μmol g–1 h–1): about 3.6 and 1.5 times higher than that of Pt/g-C3N4 (450.1 μmol g–1 h–1) and Pd/g-C3N4 (1040.5 μmol g–1 h–1), respectively. The enhanced activity observed with Pd might be related to the following properties [41]. (1) Pd has a higher density of states in the vicinity of the semiconductor's Fermi level than Pt does. The Fermi level of Pd is around 0.2 eV higher than that of Pt (~10.8 eV) [49]. (2) Pd has a much lower electron affinity (hence electron trapping capability), which may enable more facile electron transfer from Pd to protons [50]. This suggests that alloy NPs have the potential for much higher photocatalytic activity compared to their monometallic counterparts.
The AQY of the PtPd/g-C3N4 composite sample for photocatalytic H2 production was measured at 420±5 nm. One molecule of H2 was assumed to be generated by absorption of two photons. The AQY characterization result is shown in Fig. 7. From the measured light intensity and the average rate H2 evolution rate of 527.4 μmol g–1 h–1, the AQY of the PtPd/g-C3N4 composite sample at 420 nm was determined to be 5.5%. This high AQY for photocatalytic H2 evolution indicates efficient conversion of visible light energy, suggesting that the PtPd/g-C3N4 composite photocatalyst is highly active in electron-hole pair generation and separation. Thus, it can be seen that PtPd/g-C3N4 is an efficient photocatalyst for visible-light-driven photocatalytic H2 evolution.
Stability is an important parameter by which to define the performance of a photocatalyst. The stability of the photocatalyst with optimal 0.2 wt% PtPd/g-C3N4 was evaluated by performing cycle runs of photocatalytic H2 evolution, and the results are displayed in Fig. 8. It can be seen clearly that there is no substantial decrease in the H2 evolution rate after four successive runs of photocatalytic reaction, which demonstrates the high durability of 0.2 wt% PtPd/g-C3N4 for photocatalytic H2 production.
In order to determine the reason for the enhancement of photocatalytic activity by the PtPd decorated g-C3N4 sample, steady-state PL spectroscopy of the 0.2 wt% PtPd/g-C3N4 composite photocatalyst was performed to investigate the interfacial charge transfer process. As shown in Fig. 9(a), pure g-C3N4 nanosheets exhibit a strong emission peak at approximately 460 nm, indicating a high rate of recombination of the photogenerated electron-hole pairs. This generally results in low photocatalytic activity [51]. Obviously, after loading the PtPd alloy NPs, the PL emission intensity decreased, demonstrating more efficient charge separation and slower rate of recombination of the photogenerated electron-hole pairs in the PtPd/g-C3N4 composite sample. Time-resolved fluorescence decay spectra were captured to evaluate the kinetic properties of the photo-generated charge carriers in the PtPd/g-C3N4 composite sample (Fig. 9b and 9c). The average lifetime of the photogenerated hole-electron pairs could be calculated using Eq. (3):
where A is the corresponding amplitude and τ is the PL emission lifetime [35, 52, 53]. The average charge lifetime τa of pure g-C3N4 and the PtPd/g-C3N4 composites was found to be 5.0920 and 5.5392 ns, respectively. The longer charge lifetime of the PtPd/g-C3N4 composites could improve the potential for the involvement of photo-generated charge carriers in photocatalytic reactions before they recombine, which leads to enhanced photocatalytic activity [54].
Furthermore, surface photovoltage (SPV) spectra were used to investigate the separation efficiency of the photo-generated charge carriers. It is well known that a stronger SPV peak represents higher separation efficiency of the electron-hole pairs [55]. As shown in Fig. 10, both the pure g-C3N4 nanosheets and PtPd/g-C3N4 composite photocatalysts exhibit positive photovoltage response in the wavelength range between 300 and 400 nm. This matches the photo-active wavelength range of DRS. Thus, the photovoltage values are only noise in the range 400–600 nm, which means that the light in this wave band is unused in both pure g-C3N4 and PtPd/g-C3N4. In addition, the PtPd/g-C3N4 composite shows a much stronger SPV response than pure g-C3N4 does. This result demonstrates higher charge separation efficiency after loading the PtPd alloy NPs onto the surface of the g-C3N4 nanosheets. High photocatalytic activity benefits from the high charge separation rate. This result is consistent with the results on photocatalytic performance.
The transient photocurrent response of a photoelectrochemical system under chopped light illumination can give information about the kinetics of the photo-generated carriers from a different perspective. For this reason, the transient photocurrent response of the samples was recorded, and is shown in Fig. 11. It can clearly be seen that the photoelectrode composed of the PtPd/g-C3N4 composite photocatalyst shows higher photocurrent than the pure g-C3N4 nanosheets do, which suggests improved charge separation efficiency in the g-C3N4 decorated with PtPd alloy NPs. Both the low photocurrent value and the rapid decay characteristic of pure g-C3N4 can be assigned to fast recombination kinetics [56].
Based on the characterization analysis and photocatalysis experimental results, a possible photocatalytic mechanism for overall water splitting over PtPd/g-C3N4 photocatalyst is proposed (Fig. 12). Under visible light irradiation, g-C3N4 absorbs photons and excites electron (e–)-hole (h+) pairs. The electrons transfer from the valence band (VB) to the conduction band (CB), leaving the holes in the VB. However, without co-catalyst, the photo-generated e- and h+ are likely to recombine rapidly, which results in low photocatalytic H2 production activity. The PtPd alloy NPs on the g-C3N4 nanosheet surfaces act as electron traps and provide protons reduction sites for H2 production. At the same time, the h+ diffuse to the semiconductor/electrolyte interface and take part in another half-redox reaction for H2 evolution, that is, oxidizing TEOA to TEOA+. Moreover, using K2HPO4 as sacrificial agent is an effective way to enhance the photocatalytic H2 production rate even further. In the presence of K2HPO4, HPO42– acts as a mediator that participates directly in the photocatalytic reaction, such that H+ from HPO42– instead of H2O combines with the photogenerated electrons to produce H2. After H2 evolution from HPO42–, the newly formed PO43– immediately combines with H+ from H2O to regenerate HPO42– and complete the proton-reduction cycle [40].
We successfully synthesized a series of PtPd/g-C3N4 composites via the chemical deposition precipitation method. The decoration with PtPd NPs did not result in damage to the morphology or crystal structure of the g-C3N4 photocatalyst. Compared to pure g-C3N4, the PtPd/g-C3N4 composite photocatalysts exhibited enhanced photocatalytic activity for H2 production. The 0.2 wt% PtPd/g-C3N4 sample exhibited the highest photocatalytic activity. An even higher H2 evolution rate of 2885.0 μmol·g-1·h-1 was achieved when K2HPO4 was used as a sacrificial agent. In addition, the 0.2 wt% PtPd/g-C3N4 sample also showed good photocatalytic stability in the time-circle experiments. In order to explain the observed H2 evolution activity enhancement in PtPd alloy-NP decorated g-C3N4 nanosheets, a possible mechanism based on suppressed charge recombination and promoted charge transfer at the PtPd/g-C3N4 interface is proposed. This was confirmed by the SPV, PL, and time-resolved fluorescence characterizations. Therefore, our experimental results demonstrated that PtPd alloy NPs are a promising co-catalyst for g-C3N4, which can also be regarded as a potential functional co-catalyst for enhancement of the photocatalytic H2 production activities of other semiconductor photocatalysts.
2019,
Vol. 40 Issue (3): 352-361 
PDF
2019,
Vol. 40 Issue (3): 352-361 PDF