Supporting Information Photoelectrochemical Performance of a Mesoporous NiO-coated ZnO-CdS Core-Shell Photoanode Pranit Iyengar, Chandan Das, Balasubramaniam Kavaipatti* Department of Energy Science and Engineering, Indian Institute of Technology Bombay, Powai, Mumbai – 400076, India Email id:
[email protected] Phone: +91 22 25767808 Experimental Section Chemicals: The chemicals were all analytical grade and used without further purification. Cd(NO)3.4H2O (99%), CH4N2S (99%), Na2S.xH2O (min. 50% assay), Na2SO3 (98%), Na2SO4 (99%), Zn(NO3)2.6H2O (96-103%), ZnSO4.7H2O (99.5%), NiSO4.6H2O (97%), K2O8S2 (99%), Ethanol (99.9%) and NH3 (25-28%) were purchased from Emparta (India). Isopropyl alcohol (IPA, 99.5%) and H2SO4 (98%) were purchased from Vetec (Sigma Aldrich), India. The FTO-coated glass substrates (TEC-7, Sheet Resistance: 7Ω/sq) were procured from Dyesol, Australia. High purity Deionised (DI) water (ρ = 18 mΩ.cm) was used for materials’ synthesis and electrolyte preparation. FTO coated-Glass Cleaning Procedure: The FTO coated-glass substrates were ultrasonically cleaned in the diluted soap solution and DI water. Later, the substrates were dipped in concentrated H2SO4 for a few seconds (8-10 seconds) and instantly rinsed in DI water. The cleaning was further followed by ultrasonication in IPA-DI water mixture (50-50) to remove all the residues on the FTO surface. These substrates were dried with N2 gun and kept in an oven at 70 °C. ZnO Nanorod Synthesis for Device Fabrication: A seed layer of ZnO (180 nm) was first deposited on the cleaned FTO glass-substrates via dielectric sputtering. Later, the nanorods were grown on the ZnO seed layer by solution processed hydrothermal method.1 The bath for nanorod growth comprised of 3.5 ml aqueous Ammonia (25%) in 200 ml of 15 mM ZnSO4 solution. The sputtered seed layer was suspended in the solution keeping face-downwards and kept in a furnace at 95 °C for 1.5 hours. The given parameters resulted in the formation of a uniform ca. 930 nm thick ZnO nanorod film.
CdS Coating: A uniform CdS coat was achieved by adapting existing methods from literature.2,3 First, the ZnO nanorods were spin coated three times with a CdS precursor solution consisting of 0.5 M Cd(NO3 )2 and 0.5 M CH4 N2 S in ethanol at 2000 rpm for 30 seconds. After each spin coat the films were annealed at 120o C on a hot plate, cooled to room temperature, washed with DI water and dried using a Nitrogen gun. The films were then put through a spin SILAR (Successive Ionic Layer Adsorption and Reaction) regime in which they were separately spin coated alternately with 10 mM Cd(NO3 )2 and 10 mM Na2 S aqueous solutions at 3000 rpm for 20 seconds for 25 cycles. After the spin SILAR regime, the films were again annealed at 120o C on a hot plate, cooled to room temperature, washed with DI water and dried using a Nitrogen gun. Mesoporous NiO film Deposition: A well-established Chemical Bath Deposition method was used to grow a mesoporous NiO (referred to as m-NiO henceforth) layer over the prepared ZnO-CdS core-shell structure.4 80 ml 1M NiSO4 and 60 ml 0.25 M K2 S2 O8 aqueous solutions were separately prepared. 20 ml 25% Ammonia water was added to the NiSO4 solution with constant stirring at room temperature. The ZnO-CdS core-shell heterostructured films were dipped into this bath and after 1 minute K2 S2 O8 was added to the bath. NiOOH was then allowed to grow on the ZnOCdS films for 10 minutes, after which the substrates were removed from the bath, sonicated to eliminate loose particles, washed with DI water, dried using N2 and annealed at 80o C on a hot plate for 10 minutes. This resulted in the formation of a blackish NiOOH layer, which was oxidized to a transparent NiO layer on being annealed in a muffle furnace at 300 °C for 1.5 hours. Structural Characterization: The phase formation and crystallinity of the synthesized materials were investigated through Grazing-Incidence angle X-ray diffractometry using a Rigaku SmartLab instrument equipped with a scintillation detector. The Grazing-Incidence angle was kept at 0.5° to investigate our device, such that the incident X-rays can penetrate only at the surface of the device. This helped in identifying the surface materials, which are porous and less crystalline compared to the underneath highly crystalline materials. Morphology of the films was obtained from Field Emission Scanning Electron Microscopy (FESEM) analysis. The machine has an accelerating voltage from 0.1 – 30 kV with magnification up to 900,000x. Photoelectrochemical (PEC) Measurements: The PEC experiments were performed in a three-electrode system, using Pt as counter electrode, Ag/AgCl [in 4M KCl] reference electrode and the working electrode, equipped with a Biologic (SP-150) potentiostat/galvanostat. The potentials applied were normalized to RHE by correcting with the Nernst equation, ERHE = EAg/AgCl + 0.059 × pH + 0.197. The electrolyte was comprised of 0.25 M Na2S and 0.35 M Na2SO3 in DI water (pH ≈ 12.5). Prior
to performing the PEC experiments, the electrolyte was purged with high purity Ar gas (99.999%) for 15 minutes. Linear sweep voltammetry (LSV) experiments were carried out with a scan rate of 10 mV/sec. under light chopping mode (on/off) to observe the photoresponse of the device. Also, chronoamperometric experiments under chopping and continuous light insolation were performed to investigate the stability of the material. During 2 all the PEC experiments, 100 mW/cm illumination from a Xe-lamp equipped with an AM 1.5G filter was used from the rear side of the device. Optimization of the ZnO Nanorod Structure: To take maximum advantage of the visible range absorption of CdS, the ZnO nanorod structure was required to be made as sparse as possible to increase the porosity, without heavily compromising the vertical orientation. In order to achieve this, the following parameters were optimized: 1. Concentration of Zinc salt 2. Choice of Zinc Salt 3. Duration of hydrothermal deposition and ZnO seed layer thickness In the following discussion, images of the synthesis results along with brief descriptions of the parameters used have been provided. The general procedure for hydrothermal synthesis remained the same as that described above in the ZnO nanorod synthesis section, with the changing parameters. 1. Concentration Variation Figure
Zinc salt / Conc. (mM)
25 % Seed Temperature Duration of Nanorod Ammonia layer hydrothermal film o volume thickness ( C) procedure thickness (ml) (μm) (nm) (hours)
S1
Zn(NO3)2 / 30
7
180
90
3.5
1.32
S2
Zn(NO3)2 / 20
4.5
180
90
3.5
1.19
Table S1: Growth parameters for the ZnO nanorods grown using 2 different concentrations
Figure S3: ZnO nanorods grown using 30 mM Zn(NO3 )2 . Other parameters in Table S1
Figure S3: ZnO nanorods grown using 20 mM Zn(NO3 )2 . Other parameters in Table S1
Comparing Figures S1 and S2, we may observe that the density of the nanorod film has visibly reduced on account of the reduced concentration, thereby increasing the porosity of the nanostructure. We move on to further reducing the concentration down to 15 mM for the following optimization experiments. 2. Salt Variation It has been reported that the choice of Zinc Salt for the solution processing of ZnO nanorods 5 determines the polarity of the grown nanorods, and also their shape. The hydrothermal procedure was followed using two salts Zn(NO3)2 and ZnSO4. Figure
Zinc salt / Conc. (mM)
25 % Seed Temperature Duration of Nanorod Ammonia layer hydrothermal film o volume thickness ( C) procedure thickness (ml) (μm) (nm) (hours)
S3
Zn(NO3)2 / 15
3.5
50
95
3.5
0.57
S4
ZnSO4 / 15
3.5
50
95
3.5
1.5
Table S2: Growth parameters for the ZnO nanorods grown using 2 different Zn 2+ salts
In the nanorod film grown using the Zinc Sulphate (S3), we may see that the aspect ratio of the nanorods is much better than that of the Zinc Nitrate (S4) case. The nanorods in the Sulphate case grow in the hexagonal shape as opposed to the needle like nature of the the Nitrate case. The growth rate while using Sulphate also seemed to be 3 times as large as the Nitrate case. However, the vertical orientation of the nanorods in the Sulphate case was disturbed to a very large extent. This is perhaps due to the fact that the seed layer was too thin to harbor nanorods as long as 1.5 μm in length.
a
b
Figure S3: a) Cross section and b) surface morphology of the ZnO nanorods grown using Zn(NO)3. Other parameters in Table S2.
a
b
Figure S4: a) Cross section and b) surface morphology of the ZnO nanorods grown using ZnSO4. Other parameters in Table S2.
To obtain an amount of CdS in the nanostructure consistent with it’s absorption coefficient,6 we aimed to synthesize a uniform ZnO nanorod film with thickness of 1 μm. The following experiments consist of attempts in which the seed layer was made thicker and the deposition time was also reduced. 3. Seed layer thickness (slt) and hydrothermal procedure duration (hpd) variation
Figure S5: slt 50 nm, hpd 3.5 hrs
Figure S6: slt 180 nm, hpd 2.5 hrs
Figure S7: slt 180 nm, hpd 1.5 hrs *Other parameters for deposition corresponding to S5, S6 and S7 in Table S3
Figure
Zinc salt / Conc.
25 % Seed Temperature Duration of Nanorod Ammonia layer hydrothermal film volume thickness procedure thickness
(mM)
(ml)
o
(nm)
( C)
(hours)
(μm)
S5
ZnSO4 3.5 / 15
50
95
3.5
1.5
S6
ZnSO4 3.5 / 15
180
95
2.5
1.23
S7
ZnSO4 3.5 / 15
180
95
1.5
.930
Table S3: Growth parameters for the ZnO nanorods grown using different seed layers and durations
It can be seen that increasing the seed layer thickness and decreasing the deposition time have resulted in a shorter, well aligned yet porous ZnO nanorod film. The ZnO nanorod film obtained using the parameters as in S7 is the one selected for the heterostructured photoanode synthesis. Continuous NiO base layer
Figure S8: CBD procedure used to grow mesoporous NiO FTO-coated glass
The procedure used for the m-NiO film synthesis was also repeated on FTO-coated glass.4 In Figure S8, we may observe that the surface of the FTO is completely covered by the NiO base, thus offering protection from film degradation. This NiO base then grows into flakes as the growth proceeds, resulting in the increased surface area. In the following section, we will visit evidence of how the m-NiO deposition is highly contingent on the uniformity of the substrate. Keeping this in view, the CdS deposition had to be optimized in order to achieve a uniform surface morphology, providing a suitable surface for the NiO film growth.
Effect of CdS coating optimization on NiO deposition and PEC performance: A combination of two methods adapted from literature were attempted for CdS coating on the ZnO nanorod structure. They were as follows: 1. Spin 3x: The ZnO nanorods were spin coated three times with a CdS precursor solution consisting of 0.5 M Cd(NO3 )2 and 0.5 M CH4 N2 S in ethanol at 2000 rpm for 30 seconds. After each spin coat the films were annealed at 120o C on a hot plate, cooled to room temperature, rinsed with DI water and dried using a Nitrogen gun. 2. Spin 3x + Spin SILAR: This consisted of method 1 (Spin 3x) followed by the already thrice spin coated films being put through a spin SILAR regime in which they were separately spin coated alternately with 10 mM Cd(NO3 )2 and 10 mM Na2 S aqueous solutions at 3000 rpm for 20 seconds for 25 times each. After the spin SILAR regime, the films were again annealed at 120o C on a hot plate, cooled to room temperature, rinsed with DI water and dried using a Nitrogen gun.
a
b
c
Figure S9: Surface morphology and cross section on the ZnO-CdS structure obtained with Spin 3x (a and c respectively) and Spin 3x + Spin SILAR (b and d respectively).
In Figure S9 we may observe the difference between the surface morphology as well as cross section of the ZnO-CdS structure synthesized using method 1 (Spin 3x) seen in S9a and S9c, and method 2 (Spin 3x + Spin SILAR) seen in S9b and S9d. The surface obtained by Spin 3x still has large sections of the ZnO nanorods exposed outside the CdS coating, and the same obtained by Spin 3x + Spin SILAR exhibits a uniform cover of the ZnO nanorods by the CdS coat. As a result, the surface morphology in the latter is much more suitable for a uniform mNiO layer deposition. From Figure S10 we can clearly observe the difference in the m-NiO deposition attempts on the two CdS coating methods. In Figure S10a we can see large cracks in the surface of the m-NiO whereas the same is much more contiguous in Figure S10b, which translates into better protection from degradation for the ZnO-CdS core-shell film surface in the m-NiO deposited on the Spin 3x + Spin SILAR CdS coated ZnO. a
Figure S10: Surface morphology of the CBD synthesised m-NiO covered ZnO core shell with CdS deposited using a) Spin 3x and b) Spin 3x + Spin SILAR
b
a)
c)
b)
d)
Figure S11: Chronoamperometric plots of a) ZnO-CdS(Spin 3x)-NiO @ 1.23 V, b) ZnO-CdS(Spin 3x + Spin SILAR)-NiO @ 1.23 V, c) ZnO-CdS(Spin 3x)-NiO @ 0.00 V and d) ZnO-CdS(Spin 3x + Spin SILAR)-NiO @ 0.00 V. All w.r.t. RHE. These tests were performed in an S2-/SO4 2- solution under 100 mW.cm-2 A.M 1.5 illumination.
Effect of the same on PEC performance is also evident from tests shown in Figure S11. We can see that the m-NiO does not inhibit dissolution of the ZnO-CdS structure in the case of CdS synthesized by Spin 3x. Chronoamperometry under illumination of this photoanode can be seen in Figure S11a at 1.23 V and Figure S11c at 0.00 V w.r.t RHE. In Figure S11a we can see that the current response in the Light On period is appreciable in magnitude at greater -2 -2 than 3 mA.cm , but the the response in the Light Off period is not returning to 0 mA.cm . This indicates that the high positive voltage supplied to the device is causing alkaline anodic corrosion/dissolution of the ZnO-CdS surface as it has not been sufficiently protected. In contrast looking at Figure S11b, we can see the much more stable performance of the photoanode synthesized using Spin SILAR in addition to Spin 3x which shows better protection against degradation than the former case.
Also, at 0.00 V w.r.t RHE, the current response shown in Figure S11c is much lower than that of the better m-NiO deposited on CdS coated using Spin 3x + Spin SILAR in S11d. This could be due to inefficient charge separation in the large tracts of the surface not covered by m-NiO. Hence the CdS coat by Spin 3x + Spin SILAR was chosen for the photoanode synthesis.
Device Structure and Charge Transfer Schematic
Figure S12: Heterostructured device schematic of the synthesised ZnO-CdS-NiO photoanode indicating the different layers (left), and Charge transfer mechanism in the photoanode indicated with the respective band positions of the three layers in the heterostructure (right).
To summarize, the optimal method for the synthesis of the m-NiO covered ZnO-CdS core shell heterostructured photoanode was arrived at after performing a range of synthesis experiments.
References 1 2 3 4 5 6
Y. Tak and K. Yong 2005 J. Phys. Chem. B 109 19263 J. Joo, D. Kim, D.-J. Yun, H. Jun, S. Rhee, J. S. Lee, K. Yong, S. Kim and S. Jeon 2010 Nanotechnology 21 3256041 Y. Li, X. Zhang, S. Jiang, H. Dai, X. Sun and Y. Li 2015 Sol. Energy Mater. Sol. Cells 2015 132 40 X. H. Xia, J. P. Tu, J. Zhang, X. L. Wang, W. K. Zhang and H. Huang 2008 Sol. Energy Mater. Sol. Cells 92 628 N. Rathore, D. V. S. Rao and S. K. Sarkar 2015 RSC Adv. 5 28251 N. Maticiuc, T. Potlog and N. Spalatu 2010 Semicond. Conf. (CAS), Int. 02 537
