Zinc oxide nanostructure-based dye-sensitized solar cells

Abstract

Developing new technologies that could lead to alternatives to the traditional silicon-based solar panels, and to efficiently light the world in the future, is critically important because of limited natural petroleum resources. Dye-sensitized solar cells (DSSCs) are promisingly efficient and clean hybrid, organic–inorganic, low-cost molecular solar cell devices. The key components of DSSCs are the organic dyes that play the role of a photosensitizer—like the chlorophyll of a green plant that is responsible for photosynthesis—and nanostructured semiconductor metal oxides. Because of their unique, multifunctional properties, zinc oxide (ZnO) nanostructures are promising materials to use to create photoanodes for DSSCs. This review looks at recent developments in the field of ZnO-based DSSC devices; synthesis of ZnO nanostructures with variable morphologies, including nanorods, nanofibers, nanotubes, nano-/microflowers, thin sheets, and nanoaggregates; factors that control the growth and morphologies of these nanomaterials; and the role of crystallographic planes for the synthesis of versatile ZnO nanostructures. This review also covers photoelectrode fabrication, DSSC device components, nature and chemical features of the dyes used as photosensitizers, and operational principles. In addition, various photovoltaic parameters such as current density, open-circuit voltage, fill factor, photoconversion efficiency, and factors that influence these parameters for ZnO-based DSSCs are summarized and discussed.

Introduction

The consumption of non-renewable fossil fuel resources, such as oil, natural gas, and coal, and the subsequent damages to the environment and human health caused by their combustion in the twentieth and twenty-first centuries have forced the nations of the world to think about exploring alternative, renewable energy sources. The continuous and large-scale combustion of fossil fuels causes the emission of greenhouse gasses, resulting in global warming. Recently, it was reported that petroleum-based resources provide ~87% of total energy demands, whereas renewable energy resources like wind turbines, hydroelectric power, wave and tidal power, solar energy, and biomass-derived liquid fuels, including biomass-fired electricity generation, contribute only a small amount to the earth’s total energy requirements [1, 2].

To overcome this energy crisis and to provide a clean environment for future generations, the development of environmentally sustainable energy resources and technologies is required. Renewable energy resources, in particular, solar energy is one of the promising candidate for alternative energy sources, to either replace or reduce the dependence of world energy needs on fossil fuels. Earth receives about 174 × 10³ terawatts (TW) of solar radiation, whereas our global primary energy consumption is about 15 TW [3]. If even a fraction of solar energy can be harnessed, the world will never face an energy crisis. A number of technologies, such as solar photovoltaics, solar heating, and solar thermal electricity, are based on solar energy harvesting.

Photovoltaic techniques involve the direct conversion of solar energy to electrical energy. The photovoltaic phenomenon was discovered by French physicist Alexandre-Edmond Becquerel in 1839 using AgCl or AgBr as photosensitizers [4]. In 1883, Charles Fritts developed a double-junction photovoltaic cell by coating a highly pure selenium semiconductor with a thin layer of gold. The efficiency reported for this cell was only about 1% [5]. In 1954, Chapin, Fuller, and Pearson, at Bell Laboratories, created the first generation of single-junction, silicon-based photovoltaic cells, to efficiently convert sunlight directly into electricity [6].

Although silicon-based solar cells harness solar energy very efficiently, their production costs are very high. Thus, the need of the hour is to develop cheaper and highly efficient photovoltaic cells for widespread application. Recent developments in the field of nanotechnology and subsequent applications in harvesting solar energy have paved the way to tremendous opportunities for the development and manufacture of semiconductor-based, thin-layered, dye-sensitized photovoltaic solar cells, with improved efficiency, at low device fabrication cost. Tsubomura et al. [7] in 1976 developed a working dye-sensitized photovoltaic cell based on porous ZnO using a Pt counter electrode and iodide/triiodide electrolyte-based redox couple. In 1985, Desilvestro et al. [8] fabricated a TiO₂-based photoanode sensitized with a ruthenium-based dye with enhanced photoconversion efficiency. In a major breakthrough, in 1991, O’Regan and Grätzel [9] fabricated TiO₂-based dye-sensitized solar cells (DSSCs) with 7.1% photoconversion efficiency. Since then, a number of p-type and n-type metal oxides, sulfides, and selenide semiconductor nanomaterials, such as TiO₂ [10,11,12,13,14], SnO₂ [15,16,17], NiO [16, 18, 19], Nb₂O₅ [20], CuO [21, 22], Fe₂O₃ [23, 24], CdSe [25, 26], and WO₃ [27, 28] including ZnO, have been extensively explored for DSSC applications.

This article represents recent developments in the field of ZnO-based DSSC devices, photoelectrode fabrication, components of DSSC devices, nature and chemical features of the dyes used as photosensitizers, operational principles, various photovoltaic parameters, and the factors that influence these parameters. The current status of ZnO-based DSSCs is outlined.

Basic principles of DSSCs based on ZnO nanomaterials

The functioning of dye-sensitized solar cells is somewhat analogous to the naturally occurring phenomenon of photosynthesis in plants. In DSSCs, an absorbed metal-laden or metal-free organic dye on semiconductor material replaces the green, light-harvesting pigment chlorophyll. The semiconductor made with nanostructured materials with a wide band gap, such as ZnO, replaces the oxidized dihydro-nicotinamide adenine dinucleotide phosphate (NADPH) and also acts as the electron acceptor (analogous to CO₂ in photosynthesis) by accepting electrons in its conduction band. The ${\text{I}}^{ - } /{\text{I}}_{3}^{ - }$ electrolyte replaces the water [29]. It has been observed that the efficiency of DSSCs is affected by the electron injection efficiency, the extent of light absorption, and the carrier transport properties of the porous nanostructured semiconductor ZnO materials.

The key processes at work in DSSCs are light absorption and excitation of the dye molecules, injection of the electrons from excited dye sensitizer into the conduction band of the highly porous thin layer of ZnO nanomaterials with a very high specific surface area coated on a conducting electrode, charge collection and extraction at the conducting electrode, generation of electrical energy, and regeneration of the dye molecules [30,31,32,33]. When light is illuminated on the dye-sensitized transparent conduction ZnO layer, the photoexcitation of the electrons from the highest occupied molecular orbital (HOMO) to the lowest unoccupied molecular orbital (LUMO) of the dye molecules occurs, which subsequently injects electrons into the conduction band of the ZnO [34]. These electrons percolate through the porous thin layers of ZnO nanomaterials coated on a conduction substrate, where charge collection and extraction occurs in the conduction layer (CL) (Fig. 1). This extracted charge is responsible for the generation of electrical work in the external circuit (Eqs. 1–3).

$$ {\text{Dye}}{\mathop{\longrightarrow}\limits^{h\nu }}{\text{Dye}}^{*} $$

(1)

$$ {\text{Dye}}^{*} + {\text{ZnO}} \to {\text{ZnO}}({\text{e}}^{ - } ) + {\text{Dye}}^{ + } $$

(2)

$$ {\text{ZnO}}({\text{e}}^{ - } ) + {\text{CL}} \to {\text{ZnO}} + {\text{CL}}^{ - } + {\text{Electrical}}\,{\text{energy}} $$

(3)

The oxidized dye is regenerated by hole injection from a redox electrolyte that contains the ${\text{I}}^{ - } /{\text{I}}_{3}^{ - }$ couple. The regeneration of the ${\text{I}}^{ - } /{\text{I}}_{3}^{ - }$ redox couple occurs from the electrons from the counter electrode (Eqs. 4, 5).

$$ {\text{Dye}}^{ + } + \frac{3}{2}{\text{I}}^{ - } \to {\text{Dye}} + \frac{1}{2}{\text{I}}_{3}^{ - } $$

(4)

$$ \frac{1}{2}{\text{I}}_{3}^{ - } + {\text{CL}}^{ - } \to \frac{3}{2}{\text{I}}^{ - } + {\text{CL}} $$

(5)

Figure 1

Schematic representation of the working principle of ZnO-based DSSC

ZnO as photoanode material for DSSCs

ZnO can show different polymorphic forms, namely cubic zinc blende, cubic rock salt, hexagonal wurtzite, body-centered tetragonal (BCT), sodalite, cubane structures, and hexagonal boron nitride. An ideal and perfect tetrahedral arrangement is observed for zinc blende and wurtzite structures, whereas other forms are distorted with slight variations in angles but having similar average axial bond lengths [35]. At atmospheric pressure, hexagonal wurtzite structure is the most stable but can be transformed into cubic rock salt type at a very high pressure of ~10 GPa. The cubic zinc blende polymorph can only be stabilized by growing it on a substrate with cubic lattice structure [36].

Wurtzite hexagonal ZnO, an II–VI semiconductor nanomaterial, has a special place among other polymorphs of ZnO and metal oxide semiconductors because of its wide band gap of 3.37 eV, large exciton binding energy (60 meV), high electron mobility, piezoelectric and pyroelectric properties, biocompatibility, non-toxic nature, very high specific surface area due to versatile morphologies (which is the key factor for a large amount of dye adsorption and subsequent light-harvesting from the source), good chemical and thermal stability (though far more chemically unstable compared to TiO₂ and Nb₂O₅), and easy, low-cost fabrication [37,38,39,40]. Because of these versatile properties, ZnO has been explored for applications as gas sensors, biosensors, photocatalysts, solar cells, surface acoustic wave filters, light-emitting diodes, photodetectors, and photodiodes, among other applications. DSSCs, based on ZnO nanomaterials, are the most promising candidates for photovoltaic device architectures. Polymorphic forms other than wurtzite hexagonal of ZnO are not reported in the literature for the fabrication of photoanode for DSSCs due to either high-pressure stabilities or drastic growth conditions. Unlike silicon solar cells, DSSCs absorb energy from sunlight using photosensitizer dyes.

In this section, the structural properties that allow ZnO nanomaterials to be used for DSSC applications are discussed. Wurtzite is the most favored and studied form of ZnO at ambient conditions. With lattice constant parameters a = 3.249 Å and c = 5.207 Å, and ratio of c/a = 1.602, wurtzite ZnO corresponds to the P6 ₃ mc space group with two formula units per primitive cell. All of the atoms in the ZnO crystal belong to the C₃v point symmetry [39]. These lattice constant values are sensitive to the crystal point defects that arise as a result of vacancies and interstitial sites on the surface of ZnO [40], and which control the various properties of ZnO nanomaterials such as mechanical, thermal, electrical, and optical. Excess zinc atoms function as donor interstitials, causing n-type conductivity. In two interconnecting hexagonal close-packed (hcp) sublattices in hexagonal lattice Zn²⁺ and O²⁻, each O²⁻ ion is surrounded by four Zn²⁺ cations at the corners of a tetrahedron, and vice versa, with sp³ covalent bonding (Fig. 2).

Figure 2

ZnO unit cell (wurtzite structure) with ionic positions of Zn²⁺ and O²⁻ ions, with various crystal planes of ZnO wurtzite structure

Because of the polar symmetry, surfaces of wurtzite ZnO [0001] terminated by Zn atoms and $[000\bar{1}]$ terminated by O atoms have different bulk terminations. These two orientations are the most common crystal orientations of ZnO, and they bear different physical and chemical properties. The repeating units of the crystal structure during hydrothermal growth are perpendicular to the c-axis, with [0001] and the $[000\bar{1}]$ polar surfaces. Because of this, these bulk-truncated surfaces are not stable. Despite this inherent instability, the polar [0001] and the $[000\bar{1}]$ surfaces are among the most common ZnO crystal orientations.

These unstable active polar faces, [0001] $[000\bar{1}]$, along with some other stable nonpolar faces such as $[01\bar{1}0]$, $[0\bar{1}10]$, $[1\bar{1}00]$, $[\bar{1}100]$, $[\bar{1}010]$, and $[10\bar{1}0]$, make the ZnO crystals acquire a number of morphologies, including nanorods [30,31,32,33, 41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58], nanotetrapods [59], nanoprisms [60], nanotubes [61,62,63], branched and unbranched nanowires [64,65,66,67,68,69,70,71,72], nanofibers [73], nanograsses [74], nanoaggregates [75,76,77,78], nanoparticles [79,80,81,82,83,84,85,86], micro-/nanoflowers [87, 88], thin film sheets and coatings [34, 89,90,91,92,93,94,95], cactus architecture [96], hierarchical structures [97, 98], nanobullets and flakes [99], and nanocombs [100] (Tables 1, 2). These nanomaterials have been extensively studied for DSSC applications because their high specific surface area can adsorb a lot of photosensitizer dyes, leading to increased light-harvesting capabilities from the source. For improving the photoconversion efficiencies, a variety of composite heterostructures and doped ZnO nanomaterials are also reported [101, 102].

Table 1 One-dimensional (1D) ZnO nanostructures: morphology, methods of preparation, size, and other growth parameters

Table 2 ZnO nanostructures with various morphologies, methods of preparation, size, and other growth parameters

Synthesis and growth of ZnO nanostructures for DSSCs

One-dimensional (1D) ZnO nanomaterials (nanorods, nanotubes, nanowires, and nanofibers)

One-dimensional (1D) ZnO nanostructures with variable morphologies, including nanorods, nanotubes, nanowires, and nanofibers, have been extensively studied for DSSC applications and are more widely reported in the literature than the other morphologies. A number of methods for fabricating these nanomaterials have been reported. It is important to mention that the quantity and quality of ZnO nanomaterials vary widely between processes and experimental conditions [103]. Growth processes for the synthesis of 1D ZnO nanostructures can be broadly classified into two categories: (a) wet processing routes, including hydrothermal methods, sonochemical growths, and chemical bath depositions (these methods may or may not involve the use capping agents), and (b) vapor-phase processing routes, which include various thermal evaporation and vapor-phase transport processes. Processing details for the fabrication of 1D ZnO nanostructures are summarized in Table 1.

ZnO nanorods

According to Vayssieres [104] and Greene et al. [105], solution methods offer large-scale, cost-effective, and eco-friendly alternatives for ZnO nanorods at low temperature, typically <100 °C, as compared to chemical vapor deposition and electrodeposition methods. Fang et al. [41] synthesized branched ZnO nanorod arrays on FTO substrates using a solution growth method and studied the effect of Zn(NO₃)₂ and hexamethylenetetramine precursor concentrations. It was found that the concentration of the precursor compounds greatly affects the morphology of the branched nanorod array. The presence of diaminopropane (DAP), a structure-directing agent, induces the branched growth of the nanorods. Primary growth was carried out using a solution containing 20 mM Zn(NO₃)₂/HMT at 60 °C for 6 h. Constant DAP concentration resulted in the formation of various morphologies for different Zn(NO₃)₂/HMT concentrations of 25, 50, 75, and 100 mM, including petal-like branches, dense needle-like branches, columnar facet nanorods densely covered with tapered nanostructures, and unbranched columnar facets of the rod, respectively (Fig. 3). A high concentration of Zn(NO₃)₂/HMT during the secondary growth enhanced the apical growth of the primary ZnO nanorods. Al-Hajry et al. [42] and Umar et al. [30] reported the synthesis of hexagonal ZnO nanorods using Zn (NO₃)₂·6H₂O/NaOH and Zn(CH₃COO)₂·2H₂O/NH₄OH solutions, respectively. ZnO nanorods can be synthesized even in the absence of a capping agent or base, as reported by Cakir et al. [31].

Figure 3

FESEM images of branched ZnO nanorods grown using constant concentration of DAP and Zn(NO₃)₂/HMTA with concentrations of a 25, b 50, c 75, and d 100 mM, respectively. Reprinted with permission from [41], © 2011 Elsevier B.V

It is important to mention that post-annealing the nanorods increases in crystallinity. Cakir et al. [31] processed ZnO nanorods synthesized through an additive-free, low-temperature, solution-based, autoclave and microwave process for the fabrication of crystalline ZnO nanorods with diameters and lengths ranging from 14–36 and 330–558 nm, respectively. Regardless of the method of synthesis, a lateral growth is observed during the growth process. Figure 4 clearly demonstrates the effect of calcination on the morphology of the nanorods synthesized through different processes.

Figure 4

SEM images of ZnO nanorods before and after calcination synthesized through a, b simple solution method, c, d autoclave method, and e, f microwave oven method. Reprinted with permission from [31], © 2011 Elsevier B.V. and The Society of Powder Technology Japan

Another effective approach reported in the literature for the synthesis of ZnO nanorods is seeded growth. These ZnO seeds serve as energetically favored sites and building blocks for the growth of ZnO nanorods. Even distribution of the seeded ZnO nanoparticles on the substrate results in the formation of vertically grown nanorods [106]. For ZnO seed layer growth on substrates like FTO [32], ITO-coated glass [33, 43, 45], Ti foil [44] Si [52], stainless steel wire [50], and glass substrate [55], different methods, including sol–gel [32, 33, 45, 52], spin coating [46, 47, 50], atomic layer deposition [51], and sputtering method [53] are applied, using different precursors, such as Zn(CH₃COO)₂·2H₂O [33, 43,44,45, 53] or diethyl zinc [50]. The precursor compounds for seed layer growth are dispersed in solvents such as ethanol and monoethanolamine (MEA) [47, 48]. Cai et al. [32] and Valls et al. [33] synthesized ZnO nanorod arrays on annealed ZnO-seeded fluorinated tin oxide (FTO) and indium titanium oxide (ITO)-coated glass substrates, respectively, using zinc nitrate hydrate HMTA as a precursor compound.

Huang et al. [46] reported the formation of oblique nanorods with a larger diameter by adding polyethyleneimine (PEI) to the primary growth solution for seed layer formation, as well as in the secondary growth solution for ZnO nanorod formation (Fig. 5). Four types of nanorod ZnO thin films (films 1–4) were formed under different conditions. No PEI was used in ZnO sol as well as growth solution for film 1, whereas for film 4 PEI was used in ZnO sol as well as growth solutions. For films 2 and 3, ZnO sol solution was without and with PEI solution, respectively. It was suggested that the presence of PEI in primary and secondary growth solutions results in the preferential adsorption to different crystal faces, affecting both surface free energy and growth rate [107]. This can lead to inhabited radial growth, but it also permits axial growth of the primary nanorods.

Figure 5

SEM images of ZnO nanorod arrays grown using ZnO seed layers onto FTO under different conditions top views (left) and 30° tilted view (right). Reprinted with permission from [46], © 2011 Elsevier B.V

Raja et al. [48] reported the synthesis of ZnO nanorods and bundled nanorods in a two-step sol–gel dip-coating method deposited onto conducting ITO substrate. Hexamethylenetetramine (HMTA) plays an important role in controlling the concentration of ${\text{HO}}^{ - }$ ions in the growth solution. $[ {\text{Zn(OH)}}_{ 4} ]^{ - 2}$ growth units are formed as a result of a series of chemical reactions as shown below (Eqs. 6–11)

$$ {\text{C}}_{6} {\text{H}}_{12} {\text{N}}_{4} + {\text{H}}_{2} {\text{O}} \to 6{\text{CH}}_{2} {\text{O}} + 4{\text{NH}}_{3} $$

(6)

$$ {\text{NH}}_{3} + {\text{H}}_{2} {\text{O}} \to {\text{NH}}_{4}^{ + } + {\text{OH}}^{ - } $$

(7)

$$ {\text{Zn}}({\text{NO}}_{3} )_{2} \to {\text{Zn}}^{2 + } + 2{\text{NO}}_{3}^{ - } $$

(8)

$$ {\text{Zn}}^{2 + } + 4{\text{NH}}_{3} \to {\text{Zn}}({\text{NH}}_{3} )_{4}^{2 + } $$

(9)

$$ {\text{Zn}}^{ 2+ } + 4 {\text{OH}}^{ - } \to {\text{Zn(OH)}}_{ 4}^{ 2- } $$

(10)

$$ {\text{Zn(OH)}}_{ 4}^{ 2- } \to {\text{ZnO}} + {\text{H}}_{ 2} {\text{O}} + 2 {\text{OH}}^{ - } $$

(11)

The rate of formation and the morphology of the grown nanorods largely depend on nucleation as well as crystal growth rate [48, 108]. Meng et al. [49] deposited ZnO thin film through electrodeposition. In the initial step, ZnO nanorods were deposited by cathodic electrochemical deposition on FTO glass substrates at 70 °C. The deposited process lasted for 4 h under the influence of 0.5 mA current. It has been observed that the density of the ZnO nanorods greatly depends on the thickness of the seed layer [51]. Fang et al. [51] deposited ZnO seed layers with 10, 20, 50, 75, and 100 nm thicknesses on FTO conductive glass substrates through atomic layer deposition (ALD). The ZnO nanorod arrays were grown on these seeded layers using zinc nitrate and HMTA through a hydrothermal method. The ZnO nanorods grown on 10–50-nm-thick seed layers had a diameter range between 100 and 250 nm, whereas seed layers with a thickness greater than 75 nm produced more uniform nanorods, with a finer diameter of 100 nm. The morphology of substrates also plays an important role in the ZnO nanorod growth process, but sufficiently thicker seed layers can outweigh the effect of substrates [109]. Figure 6 clearly demonstrates the effect of the seed layer thickness on the density and diameter of ZnO nanorods deposited on FTO glass substrate. However, the length of the ZnO nanorods was almost uniform regardless of the thickness of the seed layers.

Figure 6

FESEM and corresponding cross-sectional FESEM images of ZnO nanorod arrays grown on ZnO seed layers with thickness: a, b 10 nm, c, d 20 nm, e, f 50 nm, g, h 75 nm, and i, j 100 nm, respectively. Reprinted with permission from [51], © 2014 Elsevier Ltd

The presence of surfactants in the seed solution, as well as in the growth solutions, also affects the morphology of the ZnO 1D materials. Pawar et al. [55] deposited a uniform ZnO seed layer using an alcohol solution of zinc acetate and diethanolamine (DEA) on ultrasonically cleaned glass substrates. Growth solution that contains capping agents like diaminopropane (DAP), polyacrylic acid (PAA), polyethyleneimine (PEI), or their mixtures resulted in highly crystalline and aligned ZnO nanorods, faceted microrods, nanoneedles, and nanotowers at a relatively low growth temperature. ZnO nanorods are modified to nanoneedles in the presence of DAP because it causes etching of the ZnO nanorod tips [110]. PAA, on the other hand, improves crystallinity and forms a complex with Zn²⁺ ions through its carboxyl group [111]. A mixture of DAP and PAA surfactant results in the formation of polydispersed ZnO nanotowers. PEI, a cationic and polar polyelectrolyte with a large number of amino groups, hinders radial growth but promotes longitudinal growth along the c-axis of the nanorods, and also stabilizes nanorods against aggregation by adsorbing on the facets of ZnO nanorods [112, 113]. It also has been demonstrated that a faster growth rate along the c-axis than on radial growth induces tapering at the ends of the nanorods, changing the morphology to nanograsses [74] or nanoneedles [114].

In general, annealing or calcination of the seed layer is carried out before the growth of secondary nanorod structures. It is suggested that annealing attributes to the improved adhesion between the substrate and the seed layer, influencing the growth of the ZnO nanorods. This affects the morphology, porosity, surface area, and hence the photovoltaic efficiencies of the ZnO nanorods. To investigate the effect of annealing of the seed layer on the growth and morphology of the ZnO nanorods, Chung et al. [58] grew ZnO nanorods on annealed and unannealed seed layers.

Figure 7a and b illustrates the morphologies of the ZnO nanorod arrays grown on the seed layer annealed in the presence of N₂/H₂. The hexagonal nanorods with diameter and length of 40 nm and 1.8 μm, respectively, were formed. These nanorod arrays were grown vertically and uniformly on the entire surface of the FTO glass. On the other hand, nanorods and seed layers were desorbed from the surface of the FTO substrate during hydrothermal synthesis at 60 °C in the precursor solutions for unannealed seed layers (Fig. 7c). However, the morphology of the nanorods grown was not affected by the annealing of the seed layers in the presence of N₂/H₂ gases. These results, however, are not in agreement with the previously reported results, in which the annealing of ZnO nanorods in a gaseous environment reduced the diameter of the nanorods and deteriorated the surface quality of the nanorods [115].

Figure 7

FESEM images of a top and b cross-sectional views of ZnO nanorods grown on the seed layer annealed in the presence of N₂/H₂, and c top view of nanorods on the unannealed seed layers. Reprinted with permission from [58], © 2010 Elsevier B.V

The morphology of the ZnO nanomaterials also can be controlled by the concentration and growth temperature of the precursor solutions. Jana et al. [60] reported the formation of hexagonal prism-shaped ZnO nanomaterials grown on a hexagonal facetted structure through a single-step process using 0.3 M zinc acetate. The average size of the as-synthesized nanoprisms was 600 nm (Fig. 8a, b). The ZnO rods were formed by the two-step process using the initial formation of zinc hydroxide from 0.3 M zinc acetate solution, followed by heating. The average length and diameter of the as-grown nanorods were reported to be 5–6 and 1.1 μm, respectively (Fig. 8c, d). On the other hand, a 0.0375 M zinc acetate solution on heating resulted in the formation of smaller sized nanorods, with length and diameter of 2.5 μm and 150 nm, respectively (Fig. 8e, f).

Figure 8

SEM images of the as-prepared ZnO samples: a, b single step without heating using 0.3 M zinc acetate solution; c, d two step with heating and using 0.3 M zinc acetate solution; and e, f two step with heating using 0.0375 M zinc acetate solution. Reprinted with permission from [60], © 2014 Elsevier Ltd

An ionic surfactant like cetyltrimethylammonium bromide (CTAB) ionizes completely in the solution, directs growth, and prevents the agglomeration for the anisotropic growth of 1D crystalline ZnO nanorods with high crystallinity and high aspect ratio. The resulting tetrahedron cation has a long hydrophobic tail and is electrostatically attached to $[{\text{Zn}}({\text{OH}})_{4} ]^{ - 2}$ growth units, which form ion pairs. These growth units are generated for ZnO crystals using a hydrothermal process from precursor compounds [116, 117]. It is supposed that this growth unit is transported by CTAB to the polar [0001] plane, resulting in the elongation of ZnO nanorods (Fig. 2). In the presence of CTAB, ZnO nanorods undergo self-assembling into nanoflowers. As the concentration of CTAB is increased, the surface tension of the ZnO is reduced and ZnO nanoflowers become denser [56].

Sudhagar et al. [57] explored jacks-like ZnO nanorod architectures, with diameters of 100–150 nm and lengths of up to 1–1.2 μm, with high internal surface areas, larger pores, and highly interconnected branches to use for photoanodes in DSSCs. Branch-free ZnO nanorods synthesized in the presence of HMTA were uniform in diameter and were randomly distributed over the substrate. These nanorods are several micrometers in length, with diameters ranging between 200 and 300 nm. The photovoltaic performances of as-synthesized branched and jacks-like ZnO nanorod-based DSSCs were compared with those of unbranched ZnO nanorod-based photoanodes. Zhu et al. [54] synthesized ZnO nanorod arrays on a zinc foil using an unseeded alkali hydrothermal process. In a secondary growth process, the hydrothermal reaction of these ZnO nanorods was performed using Zn(NO₃)₂, HMTA, and trisodium citrate. This resulted in the formation of ZnO nanorod–nanosheet hierarchical architectures.

ZnO nanotubes

ZnO nanotubes are supposed to be better suited for DSSCs because they have a higher surface-to-volume ratio and a porous structure compared to ZnO nanorods and nanowires [78, 118]. Nanotubes with hollow and oriented structures may contribute simultaneously to the extension of porosity and to the requirement for the large surface area for photosensitizer dye adsorption, leading to efficient light absorption that can enhance the efficiency of dye-sensitized photovoltaic cells [119]. The morphology and hence the surface-to-volume ratio of the ZnO nanotubes can be controlled by reaction parameters, including precursor concentration, growth temperature, and growth time. Precursor concentration controls the density and aspect ratios, whereas growth time and growth temperature control the morphology of the ZnO nanotubes [120, 121].

Xi et al. [61] investigated the effect of growth temperature and growth time on ZnO nanotubes grown on ITO-coated optical fibers. Seeded ZnO nanotube arrays were first fabricated at 95 °C for 4 h. Additional growth was carried out in a nutrient solution that contained 1:1 zinc nitrate and HMTA at 50 °C for different growth times of 1–24 h. It was observed that the length of nanotube structures increased with growth time; however, diameters and wall thickness remained unaffected. No significant growth was observed for a growth time of 1 h (Fig. 9a), but for a growth time of 4 h, a good tubular shape of ZnO nanotubes was formed (Fig. 9b). Growth continued for still higher growth times of 6 and 16 h (Fig. 9c, d). As-formed ZnO nanotube arrays were largely destroyed at a growth time of 24 h (Fig. 9e). The result further indicated that nanotubular length increased, whereas wall thickness decreased, with growth time. However, no significant change in the diameters of the nanotubes was observed. It is apparent that the growth morphologies were largely affected by the growth temperatures. When the growth temperature was increased to 65 °C, ZnO nanowires were formed instead of nanotubes.

Figure 9

SEM images of ZnO nanotubes grown at concentration of 20 mmol L⁻¹, initially at 95 °C for 4 h, then at 50 °C for a 1 h, b 4 h, c 6 h, d 16 h, and e 24 h. Reprinted with permission from [61], © 2012 Elsevier B.V

Another approach to ZnO nanotube synthesis is the chemical etching of ZnO nanorods at low temperature using alkaline solutions. Liu et al. [62] utilized electrodeposited ZnO seed layers for the synthesis of oriented single-crystal ZnO nanorods through an aqueous growth solution using zinc nitrate and HMTA precursor solutions at 95 °C. Chemical etching using a 0.1 M KOH solution at 85 °C for 1 h of the as-grown ZnO nanorods resulted in the formation of center hollow ZnO nanotubes. It was further observed that the etching time has a marked effect on the depth of the nanotubes: as the etching time increased, the depth of the hollow tube also showed a gradual increase. At still higher etching time, the walls of the nanotubes were also etched (Fig. 10a, b). To conclude, ZnO nanorods can be entirely converted to hollow nanotube arrays by controlling the solution concentration, growth temperature, and etching time.

Figure 10

SEM image of ZnO nanotube arrays at a 30-min etching time, and b 120-min etching time. Reprinted with permission from [62], © 2010 Elsevier Ltd

In a similar approach, Ameen et al. [63] illustrated that the orientation of nanotubes on FTO substrate can be controlled by synthesizing the ZnO nanotubes through seeded growth, thereby forming aligned ZnO nanotubes with hexagonal nanotubes with ∼150 and ∼300 nm average inner and outer diameters, respectively. Unseeded, nonaligned ZnO nanotubes with an average diameter of 800 nm were synthesized on FTO substrates by low-temperature solution process followed by etching, using a 0.01 M NaOH alkali solution. Figure 11a represents the schematic illustration of the synthesis of seeded ZnO nanotubes, whereas Fig. 11b–e represents the FESEM images of aligned and nonaligned ZnO nanotubes.

Figure 11

a Schematic representation of the growth mechanism and the formation of ZnO NTs; b, c low-magnification and high-magnification FESEM images of aligned ZnO nanotubes; and d, e low-magnification and high-magnification FESEM images of nonaligned ZnO nanotubes. Reprinted with permission from [63], © 2010 Elsevier Ltd

ZnO nanowires

In addition to ZnO nanorods and nanotubes, ZnO nanowire arrays have been suggested as suitable candidates for photoelectrode configurations. ZnO nanowires with suitable length can provide a direct pathway for rapid transport of the photoelectrons to the conducting surface of the photoanode [122,123,124,125]. The key factor that affects photo-to-current conversion efficiency is the surface area, and for nanowires, this can be significantly increased by increasing the length and hence aspect ratio. Homogeneous nucleation resulting in the formation of ZnO in the bulk solution rapidly depletes the reactants, thereby limiting the length and slowing down the growth rate of ZnO nanowires [126]. Thus, the main focus on achieving the high aspect ratio for ZnO nanowires should be to control the homogeneous nucleation rate [64]. Many approaches to accomplish this have been adopted and reported in the literature. One of these involves the presence and concentration of ${\text{NH}}_{ 4}^{ + }$ ions in the growth solution. Gao et al. [127] synthesized ZnO nanowires up to 14 μm in length by changing the concentrations of NH₄OH in the growth solution. Cho et al. [128] reported ZnO nanowires with a length of 13.6 μm and a very high aspect ratio of 97 by adding NH₄Cl in the growth solution. In addition to ${\text{NH}}_{ 4}^{ + }$ ions, the presence of PEI in the growth solution also suppresses homogeneous nucleation [129]. ZnO nanowires of about 40 μm have been synthesized via a preheating, PEI-assisted hydrothermal growth by Qiu et al. [130]. Huang et al. [46] also reported that the aspect ratio of ZnO nanorods increases with the addition of PEI in growth solution, transforming the morphology to nanowires. Refreshing the growth reaction solution at regular intervals also increases ZnO nanowire length [112]. Chen et al. employed a facile continuous flow injection process for the synthesis of high-quality ZnO nanowires [131]. Growth time may also result in the formation of longer nanowires. Chen et al. [64] synthesized seeded ZnO nanowires on FTO substrate. Prior to the growth of nanowires, the seeded layer was annealed at 500 °C in air for 1 h. ZnO nanowires were synthesized using a simple one-step hydrothermal method using Zn(NO₃)₂·6H₂O, HMTA, and ammonia hydroxide (NH₃·H₂O). The ZnO nanowires were grown for various growth time intervals ranging from 1 to 20 h. Figure 12 represents the variations of the ZnO nanowires as a function of growth time.

Figure 12

Variation of the length of ZnO nanowires versus growth time. The inset shows the cross-sectional FESEM image of ZnO nanowires. Reprinted with permission from [64], © 2013 Elsevier B.V

The growth process for ZnO nanowires involves an initial formation of ZnO nanowires in up to 8 h growth time followed by etching if growth is prolonged beyond 8 h. A high concentration of ${\text{Zn}}({\text{NH}}_{3} )_{4}^{ + 2}$ favors the precipitation of ZnO at 95 °C during the first 8 h, resulting in the fast growth of ZnO nanowires (Eqs. 12–14) [64, 129, 132, 133].

$$ {\text{Zn}}^{2 + } + 4{\text{NH}}_{3} \cdot {\text{H}}_{2} {\text{O}} \to {\text{Zn}}({\text{NH}}_{3} )_{4}^{2 + } + 4{\text{H}}_{2} {\text{O}} $$

(12)

$$ {\text{Zn}}({\text{NH}}_{3} )_{4}^{2 + } + 2{\text{OH}}^{ - } \to {\text{ZnO}} + 4{\text{NH}}_{3} + {\text{H}}_{2} {\text{O}} $$

(13)

Prolonged reaction time affects etching results as a result of the following reaction.

$$ {\text{ZnO}} + 4{\text{NH}}_{3} + {\text{H}}_{2} {\text{O}} \to {\text{Zn}}({\text{NH}}_{3} )_{4}^{2 + } + 2{\text{OH}}^{ - } $$

(14)

Similar approaches have been adopted by Fu et al. [68] and Tian et al. [70]. Nayeri et al. [71] studied the effect of Al-doped ZnO seed layers deposited on the ITO-coated glass through RF sputtering. It was observed that vertically well-aligned ZnO nanowires grown on Al-doped seed layers had a higher surface area than the nanowires grown on undoped ZnO seed layer. These results reveal that the hydrothermal growth rate of the ZnO nanowire is strongly dependent on the structure and composition of the seed layer [71].

Lupan et al. [65] fabricated aligned ZnO nanowire arrays on conductive indium-doped tin oxide (ITO) glass substrate by the electrochemical deposition method using a ZnCl₂ aqueous solution at 85 °C. Guérin et al. [66] reported that the treatment of the FTO glass substrate with nitric acid prior to the electrochemical deposition largely affects the length and aspect ratio of the ZnO nanowires. Fast ZnO nucleation permits the growth of denser arrays of ZnO nanowires, it was observed. There are many other reports on the synthesis of ZnO nanowires on substrates, such as ITO/PEN conductive flexible substrate through hydrothermal method [67] and zinc foil through electrochemical deposition [69].

Multistep seeding and growth can result in branching of the nanowires structure, thereby increasing the surface area for dye adsorption [134, 135]. Suh et al. [72] demonstrated the synthesis of branched ZnO nanowire for DSSCs via a two-step sequential seeding process using hot-wall chemical vapor deposition (CVD) on silicon wafers and FTO substrates. The diameters of the as-grown nanowires were 50–100 nm, and lengths were in the range of 5–6 µm. Figure 13a represents the FESEM image of uniformly distributed seeds with 3–4 nm diameter on Si wafer grown using a Zn(CH₃COO)₂·2H₂O precursor solution, whereas Fig. 13b represents the schematic growth of branched ZnO nanowires.

Figure 13

a FESEM image of uniformly distributed seed layer on Si wafer; and b the schematic representation of the growth of branched ZnO nanowires. Reprinted with permission from [72], © 2007 Elsevier B.V

Recent research in the field of DSSCs, based upon 1D ZnO nanowires, has undoubtedly improved the photoconversion efficiency, but commercialization of these DSSCs is still awaited. More research is still required to further improve the efficiency of these cells for potential commercial applications.

ZnO nanofibers

Another versatile, though less studied ZnO morphology suitable for the photoanode material in DSSCs is the porous nanofibrous morphology. Li et al. [73] synthesized porous ZnO nanofibers on Si wafer substrate using a simple one-step thermal evaporation technique and vapor-phase transport methods using a mixture of Zn and CuCl₂·2H₂O as precursors. The presence of CuCl₂·2H₂O in the precursor is thought to increase the porosity of the nanofibers. It was proposed by Li et al. [73] that with the increase in temperature, ZnO nanofibers on the Si wafer were oxidized by the oxygen gas and reacted with the ${\text{Cl}}^{ - }$ ions and H₂O of the CuCl₂·2H₂O, resulting in the formation of polycrystalline ZnO nanofibers with intermediate products, including ${\text{Zn(OH)}}_{ 2}$ and ${\text{ZnCl}}_{ 2}$ hydrate phases with low crystallinity. With further heating at high temperatures in the 600 to 700 °C range, these unstable intermediate products decompose, leading to increased porosity of the ZnO nanofibers [73].

ZnO nanomaterials with other morphologies

ZnO nanoaggregates

One-dimensional ZnO nanostructures, including ZnO nanotubes, nanofibers, nanorods, and nanowires, have an insufficient specific surface area for improved light-harnessing and dye adsorption and thus have relatively low photocurrent conversion efficiency [136,137,138].

It has been suggested that the enhancement of light-scattering functions occurs from submicron ZnO aggregates produced as a result of the self-assembly of ZnO nanoparticles. Photoconversion efficiency also can be improved by forming a light-scattering layer, thereby enhancing the optical absorption in photoelectrode films [75,76,77, 139, 140]. The excellent efficiencies of ZnO aggregates are due to the fact that these aggregates have a size comparable to the wavelengths of incident photons from source light and moreover because these aggregates consist of nanosized particles, which have a high capability of dye adsorption. Hence, a number of efforts have been made recently toward synthesizing and utilizing ZnO aggregates for DSSCs applications.

Gao et al. [75] reported an interface precipitation method for the synthesis of hierarchically structured ZnO nanocrystalline aggregates. A slower reaction at the interface of the ethanol and water as solvents, shown in Fig. 14, was supposed to take place, leading to highly porous hierarchical structures with ~100 nm to ~1 μm size distributions. As-synthesized ZnO nanostructures have a larger surface area, and excellent dye adsorption and light-scattering, as compared to the ZnO nanomaterials synthesized by a homogeneous precipitation method [75]. It was proposed that at the interface, the nucleation, and the subsequent growth, occurs because of the diffusion of water into ethanol, leading to a recession of the perimeter of the water droplets, thereby forming hierarchically structured ZnO aggregates.

Figure 14

Schematic representation of ZnO nanocrystalline aggregates formation through a homogeneous method and b interface precipitation methods. Reprinted with permission from [75], © 2012 Elsevier Ltd

The crystal size of the aggregates is an important factor because it not only affects the extent of dye adsorption sensitizer but also controls the electron transmission and recombination of photogenerated electrons [76]. Guo et al. [76] adopted a two-step hydrothermal method to synthesize ZnO aggregates with diverse particle sizes controlled by small amounts of additive water and growth temperatures. ZnO aggregates of diameter 90, 100, 180, and 500 nm were, respectively, obtained by the addition of 0, 0.1, 0.3, and 0.8 mL water at 75 °C. Zheng et al. [78] reported in situ hydrothermal growth of hierarchical nanourchin ZnO aggregates. These aggregates consist of nanorod arrays on the surface of primary nanocrystallite aggregates. These features are supposed to enhance charge transport and light-scattering effects.

ZnO nanoparticles and nanocrystals

ZnO materials in which all the dimensions are within the range of nanoscale, i.e., less than 100 nm, are called zero-dimensional nanomaterials. The most common representation of zero-dimensional nanomaterials is nanoparticles and nanocrystals [141]. Due to their small size, these nanostructures provide a sufficiently large surface area for dye adsorption. Reports in the literature demonstrate the use of ZnO nanoparticles and nanocrystals for the fabrication of efficient DSSCs [79,80,81,82,83,84,85,86].

Many chelating agents, such as monoethanolamine, diethanolamine, triethanolamine [79, 142], and PEG-400 [81], are used to prevent the agglomeration and intercalation of ZnO nanomaterials. Liu et al. [73] synthesized CTAB-assisted ZnO nanomaterials of about 10–20 nm diameter using Zn(NO₃)₂·6H₂O and the chelating agent diethanolamine. Aging of the materials was carried out at 80 °C for 24 h in a thermostat.

In the absence of chelating agents, intercalation between the nanomaterials was clearly observed. Bu et al. [80] applied a two-step growth process in a chelation agent-free environment for the synthesis of intercalated ZnO nanostructures. Nanoparticles initially synthesized from zinc acetate dihydrate were further treated with Zn(NO₃)₂·6H₂O and HMTA, which preferentially absorbs onto the radial surfaces, thereby facilitating ZnO growth along the c-axis. It has been reported that these larger intercalated ZnO nanoparticles are found to increase light-scattering efficiency and conduction pathways, whereas smaller particle size increases dye-loading, which is another key factor in achieving high cell efficiency [142]. Thus, proper engineering for intercalation and particle size is required to simultaneously achieve maximum conduction pathways and high dye-loading for ZnO nanomaterials. Figure 15a represents ZnO nanoparticles synthesized using of chelating agents PEG-400, whereas Fig. 15b depicts ZnO nanomaterials synthesized without using a chelating agent [80, 81]. A strong intercalation is observed between the nanomaterials synthesized without the use of a chelating agent. Another factor, in addition to the chelating agent, is the pH of the growth solution which plays a major role in controlling the size of the ZnO nanoparticles. Figure 15c shows the variations of crystallite size with the pH of the growth solution. A maximum crystallite size of 14.5 nm has been observed [83]. It can be postulated that the OH⁻:Zn²⁺ ratio in the growth solution controls the concentration of $[ {\text{Zn(OH)}}_{ 4} ]^{ - 2}$ growth units during the growth process.

Figure 15

a TEM image of ZnO nanoparticles synthesized in the presence of PEG-400. Reprinted with permission from [81], © 2010 Elsevier Ltd. b SEM image of ZnO nanoparticles synthesized in a chelate-free environment (The inset shows the corresponding TEM image). Reprinted with permission from [80], © 2012 Elsevier B.V. c Variations of crystallite size of ZnO nanomaterials with pH of growth solution. Reprinted with permission from [83], © 2008 Elsevier B.V

In a similar approach, Al-Kahlout et al. [84] synthesized ZnO nanoparticles with a diameter of 20 nm at a still higher pH of 12. Devabharathi et al. [85] prepared ZnO nanocrystals through a simple precipitation method at different pH values of the growth solution. Crystals of larger sizes were obtained at a higher pH of 9, compared to pH 5, at a post-annealing temperature of 400 °C. However, the crystal size was lower at pH 9 compared to pH 5 at higher annealing temperatures of 500 and 600 °C. Figure 16 shows the X-ray diffraction pattern for ZnO nanocrystals with hexagonal structure and grain sizes ranging from 30 to 60 nm. The crystallinity of the nanocrystals was found to improve with increasing pH values as well as with annealing temperature.

Figure 16

X-ray diffraction patterns of ZnO nanocrystals synthesized at a pH = 5 and b pH = 9 for different annealing temperatures. Reprinted with permission from [85], © 2014 Elsevier Ltd

Dunkel et al. [89] deposited ZnO nanoparticles through electrochemical deposition method on indium tin oxide (ITO) and antimony tin oxide (ATO) nanofibers at 70 °C using an O₂-saturated ZnCl₂ aqueous solution. Figure 17 shows a deposition of a ZnO layer over ITO, as well as ATO nanofibers. Nanoporous ZnO filled the larger gaps between the fibers below, as well as from the underlying FTO glass substrate. A ZnO coating of thickness 100–150 nm and a high density of zinc on conductive ITO nanofibers exhibit greater diameter, approximately 2 μm as compared to 1.5 μm in the case of ATO. The latter images clearly show the locations of the nanofibers.

Figure 17

SEM images of a ITO and b ATO nanofibers coated with ZnO nanoparticles. Reprinted with permission from [89], © 2012 Elsevier Ltd

ZnO thin films and sheets

Two-dimensional ZnO-based nanostructured porous layers, thin films, and coatings deposited over conducting substrates with comparatively larger surface-to-volume ratios also are one of the important morphologies studied for DSSCs applications. The porosity of ZnO-nanostructured thin films directly influences the dye adsorption rate and hence photocurrent conversion efficiencies. Many organic dyes also have been found to be useful for controlling the surface morphology and porosity of ZnO thin films. Loaded dyes can be extracted from the deposited films by treating them with dilute base solutions. A number of physical, chemical, chemical spray pyrolysis, and electrochemical applications for ZnO thin film deposition for NO₂ gas sensing applications have been reported in the literature [34, 90,91,92,93,94,95].

Parthiban et al. [90] employed a chemical spray pyrolysis technique for the deposition of ZnO thin film on ITO substrate. For this purpose, 0.05 M of Zn(CH₃COO)₂·2H₂O salt dissolved in DI water was used as a precursor. With the aid of compressed air, the precursor solution was sprayed as a fine mist onto the surface of the ITO substrate placed at a distance of 50 cm on a hot plate at 250 °C. As-deposited ZnO thin film (Fig. 18a) sensitized with Evans blue was used as a photosensitized anode for DSSCs applications. In a similar approach, Kushwaha et al. [92] deposited thin ZnO films of FTO substrate using a presynthesized nanostructured ZnO colloidal solution (Fig. 18b). The ZnO solution was loaded with the aid of compressed air onto the substrate, followed by annealing in air at 450 °C for 1.5 h. An aqueous solution of Zn(NO₃)₃·6H₂O and SDS was used as precursor solutions by Lin et al. [91] to deposit nanostructured ZnO thin films through cathodic electrochemical deposition method. For this, a ZnO foil as a counter electrode and conducting ITO-coated glass substrate as working electrode were used. The effect of variations of the concentration of the Zn(NO₃)₃·6H₂O aqueous solution was studied. The deposition process continued for 4 h with a deposition current of 0.9 mA. Zi et al. [93] also deposited ZnO thin films and sheets as a function of a concentration of Zn(NO₃)₃·6H₂O aqueous solution (Fig. 18c, d) via three-electrode electrochemical deposition on FTO glass substrate, used as a working electrode. A saturated Ag/AgCl electrode was taken as a reference electrode. Deposited samples were annealed at 500 °C for 30 min.

Figure 18

a FESEM of ZnO thin film deposited by chemical spray pyrolysis. Reprinted with permission from [90], © 2014 Elsevier B.V. b ZnO film deposited using presynthesized nanostructured ZnO colloidal solution. Reprinted with permission from [92],© 2011 Hydrogen Energy Publications, LLC. c SEM image thin film and d nanosheets synthesized via electrochemical deposition. Reprinted with permission from [93], © 2014 Published by Elsevier Ltd and Techna Group S.r.l

Chemical methods also have been reported for the synthesis of two-dimensional (2D) nanosheet structure using chemical processes [94, 95]. Baviskar et al. [96] initially deposited ZnO thin films through successive ionic layer adsorption and reaction (SILAR), followed by chemical bath deposition (CBD) onto fluorine-doped tin oxide (FTO)-coated glass substrate. As-deposited ZnO thin films were annealed at 200 °C for 30 min. These films consist of cactus-like morphologies having interconnected ZnO nanoparticles with grain size, ranging from 50 to 100 nm. It was supposed that these interconnections in the cactus morphology lead to high surface area and porosity, suitable for greater dye adsorption. Another advantage of these interconnections is the enhanced electron transport to the conduction layer of DSSCs.

ZnO nanostructures of various other morphologies, including nano-/microflowers [87, 88], nanoflakes and nanobullets [99], and nanocombs [100] also have been used for DSSC applications (Table 2). Hydrothermal, solution and thermal evaporation methods have been used for the synthesis of these nanomaterials.

The unseeded growth of ZnO microflowers was carried out by Xu et al. [87] through a simple solution deposition method and subsequently was utilized as a light-scattering layer for DSSCs. Zn(NO₃)₂·6H₂O was used as a source of Zn²⁺ ions. The presence of KOH in reaction solutions favors the formation of anionic complex $[{\text{Zn}}({\text{OH}})_{4} ]^{ - 2}$, which enhances growth along the positively charged polar [0001] direction oriented along the c-axis. Figure 19a represents the typical schematic growth process in which a glass slide is used as the substrate. A piece of zinc foil in the reactions vessel also serves as the ZnO source. A growth temperature of 60 °C for 1 h results in the deposition of a white powder that consists of ZnO microflowers on the surface of the glass slide. Figure 19b, c shows the FESEM images at low-magnification and high-magnification powers for as-synthesized ZnO microflowers with a diameter of 5 µm. It can be clearly seen that each microflower consists of a pencil-type nanorod with hexagonal cross sections with a diameter of ~300 nm and a length of ~3 µm (insets of Fig. 19c) [87].

Figure 19

a Schematic diagram of a synthetic process for ZnO microflowers through solution deposition method; b and c low-resolution and high-resolution SEM images of ZnO microflowers. Reprinted with permission from [87], © 2014 Elsevier Ltd

Umar et al. [88] synthesized flower-shaped ZnO nanostructures using Zn(NO₃)₂·6H₂O, HMTA, and NaOH at a low growth temperature of 145 °C and pH 10 for 3–7 h. The typical diameter of a single flower is 350 ± 50 nm. Petal thicknesses were 70 ± 10 and 150 ± 20 nm at the tips and bases, respectively. Figure 20a, b and c, d, respectively, represents the low-magnification and high-magnification FESEM images of these flower-shaped ZnO nanomaterials.

Figure 20

Typical (a, b) low-magnification and (c, d) high-resolution FESEM images of flower-shaped ZnO nanostructures. Reprinted with permission from [88], © 2012 Elsevier Ltd

Mou et al. [99] reported ZnO nanobullets/nanoflakes synthesized by a simple hydro-/solvothermal method using a water/ethylene glycol solvent and Zn(CH₃COO)₂·2H₂O as a source of Zn²⁺ ions. A mixture of water/ethylene glycol and pure ethylene glycol resulted in the formation of ZnO nanobullets (Fig. 21a) and nanoflakes (Fig. 21b) during the hydrothermal growth process.

Figure 21

High-magnification SEM image of a nanobullets, and b nanoflakes (The inset of part (a) to corresponding EDS spectrum). Reprinted with permission from [99], © 2010 Elsevier B.V

Nanobullets are obtained as a result of preferential growth along [0001] direction with lowest surface energy as compared to radial [101] directions. The growth process in the absence of water lacks hydroxyl OH⁻ ions, resulting in very short pillars or flakes, whereas in an aqueous medium, because of the presence of OH⁻ ions $[ {\text{Zn(OH)}}_{ 4} ]^{ - 2}$, growth units are generated. These negatively charged growth units are preferably adsorbed on the positive polar along [0001] direction by Coulomb interactions.

ZnO nanocombs also have been grown using a horizontal quartz tube furnace onto the FTO substrate and subsequently utilized as photoanode material in DSSCs by Umar et al. [100]. In this thermal evaporation, highly pure metallic zinc powder was used as source material. The temperature during the deposition process was maintained at 570 °C, and the reaction lasted for 60 min. Figure 22a, b shows the typical low-magnification and high-magnification FESEM images, whereas Fig. 22c represents the typical high-magnification TEM image of the as-synthesized ZnO nanocombs.

Figure 22

Typical a low-magnification and b high-magnification FESEM images of nanocombs, and c typical high-magnification TEM image of nanocombs. Reprinted with permission from [100], © To the authors, 2009

Basic DSSC components

A working DSSC consists of four main components: (1) a ZnO nanostructure coated with a transparent conducting electrode (a photoanode), (2) a counter conducting electrode, (3) photosensitizer dye molecules, and (4) an electrolyte-containing iodide–triiodide (${\text{I}}^{ - } / {\text{I}}_{ 3}^{ - }$) redox couple.

Photoanode electrode and counter electrode

To create a ZnO nanostructure-based photoanode, a slurry of ZnO nanomaterials is made by mixing with an aqueous polyethylene glycol [81, 83, 87, 98] solution using a mortar and pestle grinder. The slurry was coated onto a conducting substrate, such as FTO glass, ITO-coated glass, Ti foil, fiber-type stainless steel, Zn foil, or glass (Tables 1, 2), usually by doctor blade method [31, 42, 52, 75, 76, 80, 81, 89, 99], or spin coating [33]. Chemical bath deposition [32], electrochemical deposition [96], and chemical vapor deposition [72, 136] techniques are also reported in the literature for the direct growth of the ZnO nanostructures on conducting substrates. Ultrathin graphene films can be used as a solid substrate for DSSCs instead of commonly used ITO and FTO substrates. CoS nanoparticles deposited electrochemically on the flexible ITO/PEN film function similarly to Pt as a triiodide reduction catalyst in DSSCs as well as in counter electrodes [143]. Drying the coated substrates is generally followed by further calcination or annealing at high temperature, 400–500 °C, to increase the crystallinity and porosity of the coated ZnO nanomaterials thin layers. This is required for better percolation of the electrons through the layers, and to reach the electron conduction layer for charge collection and extraction. Direct growth of the ZnO nanomaterials on the conducting substrate is another method to create the photoanode. The use of semiconductor ZnO quantum dots (QDs) in place of nano- or micromaterials could enhance photoconversion efficiency. QDs can generate a large number of electron–hole pairs per photon, which helps photovoltaic devices attain high photoconversion efficiencies. Moreover, QDs have a high specific surface area for better dye adsorption. It is important to mention that the thickness of the ZnO-nanostructured thin layer is one of the key factors that determine the efficiency of the DSSCs [97]. These QDs can meet the requirements of thin films. A very thin layer would certainly help the diffusion of the photoelectrons to the conducting surface of the photoanode. According to Hod et al. [144] and Tian et al. [145], QDs provide a possible presence of grain boundaries between them. Energy level positions are altered on interaction with the electrolyte affecting the ratio between electron injection to the oxide and to the electrolyte and hence enhancing the cell performance. Thus, smaller QDs exhibit a higher rate of electron injection than their nanosized counterparts. It is also important to note that the transparency of the conducting electrode is reduced by coating or growing thin layers of ZnO nanomaterials over the entire visible region, in addition to near-IR regions, of the solar spectrum. The counter electrode consists of a catalyzing layer of either graphite, platinum, or cobalt sulfide (Co₉O₈) coated over FTO glass substrates to facilitate the electron transfer mechanism to the electrolyte [30,31,32, 42, 146]. Another choice is the use of highly porous graphite and graphene as a counter electrode in place of costly Pt metal.

Photosensitizer dyes

The photosensitizer dyes used in ZnO-based DSSCs to harness solar light may be either metal-based complex dyes (Fig. 23) or metal-free organic dyes (Fig. 24) (Table 3). In spite of short and low-cost synthetic routes, easy dye structural modifications, significantly higher extinction coefficients, and quite narrow absorption bands of organic dyes, metal-based dyes, particularly N719, have been reported to reach maximum extent [30,31,32,33, 41,42,43,44,45,46,47,48,49, 51,52,53,54,55,56] (Table 3). The high magnitude of extinction coefficients in metal-free organic dyes also reduces the amount of dye required for efficient light-harvesting. For the adsorption of these dyes, ZnO nanomaterial-coated conducting substrates are immersed in an alcohol solution of the dye for a few hours at room temperature [147]. The extent of the dye adsorbed depends on the morphology and specific surface area of the ZnO nanomaterials, dye concentration, and soaking time. The unadsorbed dye is rinsed from the photoanode with ethanol, followed by drying in an inert atmosphere at room temperature [30,31,32,33, 42,43,44, 56,57,58,59,60]. Because the efficiency of ZnO nanomaterial-based DSSCs depends on the amount of light absorbed by the dye and subsequent excitation, the molecular designing of novel dyes can be helpful. Figures 23 and 24 represent the metal-based complex dyes and metal-free organic dyes, respectively.

Figure 23

Some ruthenium metal-based complex dyes used as photosensitizers in DSSCs

Figure 24

Some metal-free organic dyes used as photosensitizers in DSSCs

Table 3 ZnO morphologies, DSSC features, growth conditions, and various photovoltaic performances

This discussion makes it clear that absorbing visible light to the maximum extent is the key factor to improve photo-to-current conversion efficiency [50, 51]. Structural modification of dye molecules that can create a high molar extinction coefficient can increase the extent of absorbed photons with energy that matches the dyes’ excitation energy. Another approach for better solar light absorption is the use of a combination of dyes that complement each other in their spectral properties [51].

Although high-performing synthetic ruthenium-free dyes have been used extensively, they are very expensive to purchase and tedious to manufacture. It is also important to test the dye toxicity. These problems could be solved by using less expensive, non-toxic, natural dyes in ZnO-based DSSCs. These dyes can be easily and safely extracted from plants. Engineering the anchoring groups in these dyes can be further beneficial. A sufficiently large extent of dye adsorbed on the ZnO nanomaterials may reduce transparency, leading to reduced photoconversion efficiencies. One suggested an approach to this may be the in situ dye adsorption during ZnO nanostructure fabrication.

Recently, Singh et al. [148] synthesized monomeric N,N-dimethyl- N4-((pyridine-4yl)methylene) propaneamine and polymeric polyamine-4-pyridyl Schiff base dyes through a single-step condensation reaction between 4-pyridinecarboxaldehyde, N,N-dimethylpropylamine and polyamine using dry methanol as solvent. Other organic dye such as (E)-2-Cyano-3-(50-{8-[4-(diphenylamino)phenyl]-2,3-diphenylquinoxalin-5-yl}-[2,20-bithiophen]-5-yl) acrylic acid (DJ-104), (E)-2-Cyano-3-(5-(8-(4-(diphenylamino)phenyl)-2,3-diphenylquinoxalin-5yl)Thiophen-2-yl) acrylic acid (CR-147), 5,15-bis(2,6-dioctoxyphenyl)-10-(bis(4-hexylphenyl) amino-20-4-carboxyphenyl ethynyl) porphyrinato]zinc(II) (YD2-o-C8), (E)-2-(5-((7-(4-(bis(4-(5-(4-(hexyloxy)phenyl)thiophen-2-yl)phenyl)amino)phenyl)-2,3-dihydrothieno[3,4-b][1, 4]dioxin-5-yl)methylene)-4-oxo-2-thioxothiazolidin-3-yl)acetic acid (B18), (E)-2-(5-((6-(4-(bis(4-(5-(4-(hexyloxy)phenyl) thiophen-2-yl)phenyl)amino)phenyl)-4,4-didodecyl-4H-cyclopenta [1,2-b:5,4 b’] dithiophen-2-yl) methylene)-4-oxo-2-thioxotetrahydrothiophen-3-yl) acetic acid (CPTD-R) and (E)-2-(5-((5-(7-(4-(bis(4-(5-(4-(hexyloxy) phenyl)thiophen-2-yl) phenyl)amino) phenyl)benzo[c] [1, 2, 5]thiadiazol-4-yl) thiophen-2-yl)methylene)-4-oxo-2-thioxotetrahydrothiophen-3-yl)acetic acid (BTD R)are also reported in the literature to harvest the photoenergy [149,150,151,152].

Electrolytes containing the redox couple iodide/triiodide (${\text{I}}^{ - } /{\text{I}}_{3}^{ - }$) and Co (II)/Co(III)

In ZnO-based DSSCs, regeneration of the oxidized dye, and hence the completion of the circuit, occurs with the help of the electrolyte containing the ${\text{I}}^{ - } /{\text{I}}_{3}^{ - }$ redox couple. The viscosity of the electrolyte greatly affects electron mobility and hence conductivity, so a less viscous electrolyte is preferred. In the redox couple, the source of the iodide ion may be LiI, NaI, KI, tetra-butylammonium iodide, or a mixture of these iodides (Table 3). Counter ions of these iodides do not significantly affect the conductivity of the electrolytes. These iodides, along with molecular iodine I₂, usually are dissolved in non-protonic solvents such as acetonitrile, propylene carbonate, or methoxypropionitrile to make the electrolyte (Table 3). The major fabrication shortcoming of DSSCs is leakage of the filled liquid state redox electrolyte, leading to reduced photoconversion efficiency and stability. For better mechanical stability, a solid-state electrolyte can be used as an alternative to the liquid electrolyte. Replacement of the liquid electrolyte with either inorganic p-type semiconductors or organic hole transporting materials can reduce evaporation as well as leakage of liquid electrolyte. Gel electrolytes are another attractive substitute because gelled ionic liquids have high ionic mobility in spite of their high viscosity. The presence of the 4-tert-butylpyridine (4-TBP) and polymer gel forms quasi-solidified electrolytes, avoiding the leakage of the electrolyte from the DSSC and resulting in better cell performance (Fig. 25; Table 3).

Figure 25

Structure of 4-tert-butylpyridine (4-TBP)

The most commonly used additive in the redox electrolyte is 4-TBP, which is a weak base because of the presence of a lone pair of electrons on the N-atom. Additionally, this additive is reported to increase V _OC due to its function as a charge recombination inhibitor with no marked influence on the photocurrent [153].

It has been noticed that redox couple ${\text{I}}^{ - } /{\text{I}}_{3}^{ - }$ is highly corrosive in nature and corrodes the commonly used sealants and metal interconnect (such as silver, copper, aluminum, and gold) with time and hence reduces the magnitude of the open-circuit voltage of the DSSC [154]. To reduce the corrosiveness of the redox couple ${\text{\rm I}}^{ - } /{\text{\rm I}}_{3}^{ - }$ and to obtain high V _OC, alternative materials with higher redox potential, such as cobalt polypyridine complexes, should be introduced to DSSCs [155]. Fan et al. [156] observed superior performance for [Co(bpy)₃]^(2+/3+)-based ZnO DSSCs as compared to ${\text{\rm I}}^{ - } /{\text{\rm I}}_{3}^{ - }$-based DSSCs. V _OC enhancement was attributed to better correlation between the [Co(bpy)₃]^(2+/3+) redox potential and the oxidation potential of the photosensitizer dye. In addition to bipyridyl complexes, some hexadentate polypyridyl ligands, such as [6,6′-bis(1,1-di(pyridin-2-yl)ethyl)-2,2′-bipyridine, bpyPY₄] and [(2,6-bis(1,1-bis(2-pyridyl)ethyl)pyridine, PY5Me₂)], are used to form a cobalt redox couple, i.e., [Co(bpyPY₄)]^2+/3+. Such redox couples are reported to be better and even more stable than [Co(bpy)₃]^(2+/3+) redox couples. Higher stability in these polydentate–cobalt complexes can be easily explained by the greater stability constant that results from chelation [157].

DSSC fabrication

Figure 26 represents the scheme for the steps in DSSC fabrication. The surface of the conducting substrate is coated with the slurry of the ZnO nanomaterials prepared in different solvents such as polyethylene glycol [30, 32, 33, 42, 50, 56], ethanol [31], ethyl cellulose and terpineol [76], and even water [99]. The doctor blade method is most widely reported for coating this slurry onto the substrate [31, 42, 52, 75, 76, 80, 81, 89, 99]. However, direct synthesis and deposition of ZnO nanomaterials onto the surface of the substrate also are reported [32, 33, 71, 96, 136].

Figure 26

Schematic representation of the DSSC fabrication process

Coated or deposited ZnO thin films are dried at room temperature and subsequently are calcined at a specific temperature. The coated substrate is immersed in the dye solution in ethanol at room temperature, followed by rinsing with ethanol. This is the most important step of the device fabrication because the extent of dye adsorption plays a crucial role in photon trapping. The dye-adsorbed ZnO thin films are dried under N₂ at room temperature. Another sheet of Pt-coated conducting substrate serves as a counter electrode. This counter electrode is mounted over the dye-adsorbed ZnO photoanode. The edges of the DSSC are sealed with a sealing sheet or glue. For the introduction of the liquid electrolyte into the cell, one or two small holes are drilled in the counter electrode. After the introduction of the electrolyte, these holes are sealed with a sealing sheet or glue and a cover glass.

Basic parameters for evaluating the performance of DSSCs

To evaluate the performance of DSSCs based on ZnO nanomaterials, current–voltage measurements are taken into account. Standard conditions, such as a constant cell temperature of 278 K, a power density of 1000 W/m², and simulated ideal sunlight characterized as AM 1.5G (the air mass coefficient), which corresponds to incident sunlight at an angle of 48° as measured from the zenith, are used globally to compare photoconversion efficiencies (η). The various basic parameters used for comparative analysis are discussed in this section.

Short-circuit current density (J _SC)

The short-circuit current density (J _SC) is defined as the ratio of the short-circuit photocurrent to the active cell area. The short-circuit photocurrent (I _SC) is the cell photocurrent measured at zero voltage. The value of J _SC depends on the amount of sunlight harnessed from the visible region of the solar spectrum by the photosensitizer dye molecules [158].

Open-circuit voltage (V _OC)

The open-circuit photovoltage (V _OC) is the voltage generated by the cell when no external load is present, that is, when the current within the cell is equal to zero. It is related to the work function of the electrode material and provides information for charge separation [159, 160]. The value of V _OC is greatly affected by the potential energy difference between the quasi-Fermi level of electrons in the nanosized semiconductor materials, and the chemical potential of the redox mediator in the electrolyte [148, 159,160,161,162,163].

Fill factor (FF)

The fill factor (FF) is the ratio of the maximum power output to the product of the short-circuit current density (J _SC) and open-circuit voltage (V _OC) (Eq. 15).

$$ FF = \frac{{J_{\text{MP}} . V_{\text{MP}} }}{{J_{\text{SC}} V_{\text{OC}} }} $$

(15)

where J _MP and V _MP are the corresponding current density and voltage, respectively, of the maximum power point.

Photoconversion efficiency (η)

Photoconversion efficiency (η) for a dye sensitizer cell can be expressed in the form of Eq. 16.

$$ \eta = J_{\text{SC}} \cdot V_{\text{OC}} \cdot FF $$

(16)

where J _SC represents the short-circuit current density, V _OC is the open-circuit photovoltage, and FF is the fill factor of the solar cell.

In some cases, the efficiency of the DSSC is expressed in terms of incident photon-to-current conversion efficiency (IPCE), which depends on short-circuit current density (J _SC), and is defined using photoresponse and light intensity as given in Eq. 17.

$$ {\text{IPCE}}(\lambda ) = \frac{{1240\,({\text{eV}}\,{\text{nm}}) \times J_{\text{SC}} (\upmu{\text{A}}/{\text{cm}}^{2} )}}{{\uplambda({\text{nm}}) \times I(\upmu{\text{W}}/{\text{cm}}^{2} )}} $$

(17)

where λ is the wavelength of the absorbed visible light and I is the light intensity corresponding to λ [164]. DSSCs usually are illuminated with a light intensity of 1000 W/m² (Air Mass 1.5 Global; AM 1.5G).

Mechanism of photocharge generation for ZnO-based DSSCs

The two basic approaches used in the characterization of photoconversion efficiency of DSSCs are a current–voltage curve measurement to determine the photoconversion efficiency and electrochemical impedance spectroscopy for the study of charge transfer resistances.

Current–voltage curve measurements

The current–voltage (I–V) curve represents the most important characterization method for DSSCs. These characteristics of a solar cell are described by an electrical circuit as shown in Fig. 27.

Figure 27

Equivalent electrical circuit for a dye-sensitized solar cell

During forwarding voltage biasing, the dark diode current I _dark flow is opposite to the photogenerated current I _gen. The electric current measured in the circuit is given as (Eqs. 18, 19).

$$ I_{\text{ph}} = I_{\text{gen}} - I_{\text{dark}} - I_{\text{shunt}} $$

(18)

where

$$ I_{\text{dark}} = I_{\text{sat}} \left( {e^{{{\raise0.7ex\hbox{${q(V + RsI}$} \!\mathord{\left/ {\vphantom {{q(V + RsI} {nkT}}}\right.\kern-0pt} \!\lower0.7ex\hbox{${nkT}$}}}} - 1} \right)\quad {\text{and}}\quad I_{\text{shunt}} = \frac{{V + IR_{s} }}{{R_{\text{Shunt}} }} $$

(19)

I _sat is the reverse saturation current, V is the bias voltage, and n is the ideality factor. The charge recombination in the photoactive layer induces shunting current I _shunt. The shunt resistance can be minimized by proper design of the cell, but the series resistance has a significant impact on the characteristics. The origin of the dark current in DSSCs is attributed to the loss of the injected electron from nanostructured semiconductor materials. The reduction in dark current increases the open-circuit voltage of the DSSC (Eq. 20, 21) [165].

$$ {\text{For}}\,R_{\text{Shunt}} \to \infty ,\quad {\text{and}}\,R_{S} \to 0,\quad I_{\text{ph}} = I_{\text{gen}} \,{\text{for}}\,V = 0 $$

(20)

$$ {\text{and}}\quad V_{\text{OC}} = \frac{nkT}{q}\ln \left( {\frac{{I_{\text{gen}} }}{{I_{\text{sat}} }} + 1} \right)\quad {\text{for}}\,I_{\text{ph}} = 0 $$

(21)

In the case of nanomaterial-based DSSCs, the surface area exposed to irradiation is important; hence, in such cases, the actual curves are expressed in terms of short-circuit current density (J _SC) in place of current. The open-circuit photovoltage (V _OC), short-circuit current density (J _SC), and shape of the J–V curve determine the efficiency of DSSCs under any given light condition. The solar conversion efficiency is given according to the relation in Eq. 16 modified to Eq. 22.

$$ \eta = \frac{{J_{\text{SC}} \times V_{\text{OC}} \times FF}}{{P_{\text{in}} }} $$

(22)

where FF is the fill factor, which depends on the diode quality factor, and $P_{\text{in}}$ is the solar energy incident on the DSSC. The fill factor is defined according to Eq. 15. The value of FF is reduced by the resistance of the cell, including the sheet resistances of the conducting substrate and counter electrode, in addition to resistance from the photoanode [165]. A typical J–V curve of a solar cell with respect to a standard reference simulated illumination with a light intensity of 100 mW/cm² (Air Mass 1.5 Global; AM 1.5G) is shown in Fig. 28.

Figure 28

Typical current–voltage (J–V) characteristics of a ZnO photoanode-based DSSC. Reprinted with permission from [88], © 2012 Elsevier Ltd

Electrochemical impedance spectroscopy

Electrochemical impedance spectroscopy (EIS) is an experimental technique used to analyze electrochemical systems working on the basic principle of Faraday’s laws. Recently, this technique has been widely applied in the field of characterization of electrochemical properties of semiconductor nanomaterials, including characterization of thin film coatings, storage batteries, and fuel cells. It is used extensively as a method for investigating mechanisms in electrodeposition, passivity, the diffusion of ions, and in the study of semiconductor interfaces. The system observes the current and impedance in the angular frequency domain upon applying low AC voltage. The impedance profile is obtained using the relation given in Eq. 23.

$$ Z\left( \omega \right) = \frac{V\left( \omega \right)}{I\left( \omega \right)} $$

(23)

EIS data for electrochemical cells generally are represented in the form of Nyquist plots, which depict imaginary impedance against the real impedance of the cell. Imaginary impedance is helpful in explaining the capacitive and inductive character of DSSCs. These curves provide an insight into the possible mechanism of DSSCs.

EIS spectrum typically shows three semicircles in the Nyquist plot. In the Nyquist plot, three characteristic semicircles represent characteristic charge dynamics in the electrolyte/counter electrode interface (cathode) (R _ct1), the ZnO/electrolyte interface (charge transfer) (R _ct2), and the diffusion process of ${\text{I}}^{ - } / {\text{I}}_{ 3}^{ - }$ in the electrolyte called Warburg diffusion (R _diff) in the frequency range of few MHz to kHz. But in the case of ZnO-based DSSCs, the conventional Warburg diffusion resistance (R _diff) of the redox couple virtually overlaps with R _ct2 due to the very thin ZnO films and very short length of I ⁻ ion diffusion [44]. One such Nyquist plot for DSSCs with photoanodes composed of ZnO nanoparticles and ZnO nanorods is shown in Fig. 29. The equivalent circuit for the DSSC is shown in the inset of Fig. 29. The results represented in Fig. 29 are theoretical plots based upon the electrical circuit shown in the inset. R _S represents the ohmic serial resistance of the equivalent circuit.

Figure 29

EIS spectra of ZnO nanoparticles and ZnO nanorod composite-based DSSCs measured at 100 mW/cm⁻². Reprinted with permission from [44], © 2012 Elsevier Ltd

At the electrode interface, an electrical double layer is formed due to the ion attraction to the surface charge [166]. Charges can be stored and accumulated in the electrical double layer, which can be described as a capacitor. In addition, the process describing the charge moving toward the semiconductor/electrolyte interface can be represented by a charge transfer resistance. Meanwhile, the electrolyte/counter electrode interface also possesses characteristic capacitance and charge transfer resistance that are affected by the interfacial properties. The impedance at the semiconductor/electrolyte and electrolyte/counter electrode interfaces can be represented by a parallel combination of chemical capacitance and charge recombination resistance. Between the electrodes, ion diffusion through the electrolyte also can be considered Warburg impedance, with resistive and capacitive components. In an ${\text{I}}^{ - } / {\text{I}}_{ 3}^{ - }$ electrolyte, the Warburg impedance (diffusion impedance) is governed by the diffusion length of triiodide ion in the electrolyte [167, 168]. In some cases, the capacitive components in equivalent circuit mode are replaced by constant phase elements (CPEs) because of the roughness at the interface [169]. In the presence of surface roughness, the current density along the surface is not homogeneous and the capacitance dispersion is due to surface geometry and adsorption effects [170].

Factors affecting the performance of ZnO-based DSSCs

Morphology of ZnO nanomaterials and extent of dye adsorption

The tetrahedral and non-centrosymmetric crystal structure of ZnO with polar symmetry along the hexagonal axis results in anisotropic crystal growth. As a result, a variety of morphologies for ZnO nanostructures is formed. The active surface area of these nanostructures, in turn, depends on the morphology of the ZnO nanomaterials. It is a well-known fact that the greater the surface area, the greater the ability to harness the visible light to the maximum extent, which is the key factor for enhancing the photo-to-current conversion efficiency of a DSSC. Therefore, the role and extent of photosensitizer dye-loading by ZnO nanomaterials directly influences the efficiency of a DSSC. In general, greater the extent of dye-loading, greater is the assurance that incident photons are trapped to the maximum extent, which leads to a larger short-circuit current density [41, 56, 74]. The promising properties of the 1D nanostructures favor the transport and minimization of recombination effects. Further, due to the existence of a grain boundary-free direct pathway toward the external circuit, electron collection is also improved for 1D nanostructures [171]. Fang et al. [41] observed that the optical absorption and hence photoconversion efficiency of DSSCs based on branched ZnO nanorod arrays was better than corresponding primary nanorod arrays. This suggests that branching increases the surface area and dye-loading. Additional branching increases the surface area and thereby improves light-harvesting properties (Table 3). In a similar study, Xiao et al. [56] observed that ZnO nanoflower films with a greater surface area had better dye-loading than ZnO nanorod films. The conversion efficiency of a ZnO nanoflower-based DSSC was 1.37% higher than a ZnO nanorod-based DSSC. Zhu et al. [74] also observed improvement in energy conversion efficiencies with the surface area by improving the N-719 dye-loading for ZnO nanograss-based DSSCs.

The pH of the growth solution has a remarkable effect on the morphology of the ZnO nanomaterials. Lizama-Tzec et al. synthesized ZnO nanomaterials with different morphologies and textures, through sonochemical synthesis method at different pH values of 5.5, 8, 10, and 12. As-synthesized nanomaterials were deposited on FTO via screen printing technique for the fabrication of photoanode. It was suggested that the influence of nanoparticles size and morphology, on the photoconversion efficiency, was affected by the total surface area of the film coated and inter-particle electrical connectivity [172].

From Fig. 30, it can be concluded that the short-circuit current density and hence the photoconversion efficiency correlates with the dye adsorption extent related to different morphologies and textures of the different ZnO materials, rather than with total surface area [172].

Figure 30

a Efficiency, b short-circuit current density, c total surface area, and d dye concentration as a function of the synthesis pH corresponding to different ZnO morphologies and textures. Reprinted with permission from Ref. [172], © The Royal Society of Chemistry 2015

Hierarchical ZnO nanocrystalline aggregates with a larger surface area, better dye-loading, and improved light-scattering exhibited better photoconversion efficiency as compared to DSSC based on ZnO nanoparticles synthesized through a precipitation method as studied by Gao et al. [75]. Jia et al. [77] synthesized submicrometer-scale ZnO composite aggregates consisting of nanorods and nanoparticles using a wet-chemical route and studied for DSSC activities. These composite aggregates exhibited improved DSSC efficiency due to the relatively high surface area, fast electron transport, and enhanced light-scattering capabilities as compared to the monodisperse ZnO nanoaggregate-based DSSCs. In a similar study, DSSC based on ZnO nanoflakes was found to have better light-to-current conversion efficiency due to a greater dye adsorbing capacity as compared to the DSSC based on ZnO nanobullets [99]. Figure 31 represents the characteristic J–V curves of the DSSCs based on ZnO nanobullets and ZnO nanoflakes.

Figure 31

Characteristic J–V curves of the DSSCs based on a ZnO nanobullets, and b ZnO nanoflakes. Reprinted with permission from [99], © 2010 Elsevier B.V

The porosity of the ZnO nanomaterials plays a crucial role in the photoconversion efficiencies [173]. The extent of dye adsorption and hence light absorption capacity can be enhanced by increasing the porosity of the ZnO films deposited as photoanodic materials. Sutthana et al. [174, 175] reported an acid vapor texturing technique to modifying photoelectrode using nitric acid vapors. Photoconversion efficiencies and short-circuit current densities were increased from 1.83 to 2.08 and 6.06–6.30 mA/cm², respectively, when texturing time was increased from 0 to 4 min [175]. For a further increase in texturing time to 6 and 8 min, both photoconversion efficiencies and short-circuit current densities were decreased (Fig. 32).

Figure 32

Various parameters for ZnO DSSCs using photoelectrode modified at various texturing times using HNO₃ vapors. Reprinted with permission from [175], © 2016 Elsevier B.V

According to Chang et al. [176], ZnO nanorod-based DSSC exhibited shorter electron transport time (τ _d), longer electron lifetime (τ _n ) along with higher τ _n /τ _d ratio as compared to ZnO nanoparticle-based DSSCs. Recently, Tao et al. [177] compared morphologies of the ZnO nanorod synthesized through a newly designed continuous reactor and batch-type reactor. The flow rate in the reactor will not only affect the morphology of the ZnO nanoarrays but also affect the pattern distribution of nanoarray on the electrode surface. ZnO nanorods with uniform morphology and controlled length were formed by continuous reactor as compared to the batch-type reactor. Additionally, wire-shaped ZnO-type photoanode fabricated from the continuous reactor showed excellent photoconversion performances than that of the batch-type reactor.

To conclude, nanoarchitecture engineering for the ZnO nanomaterials with greater surface area and porosity can lead to large dye-loading, strong light-scattering, and direct electron transport networks, which results in enhanced dye-sensitized solar cell performance [78].

Effect of film thickness and additives

Another important factor that affects photoconversion efficiency of ZnO-based DSSCs is the thickness of the ZnO thin layer deposited on the conducting substrate. Giannouli et al. [97] observed that the proper optimization of the thickness of the ZnO layer may increase efficiency. The current density of the DSSC was found to increase with increasing film thickness. This was attributed to improved dye-loading of the ZnO films. ZnO thin films with thickness greater than 10 μm showed, however, low current density. It was suggested that the very thick ZnO films do not adhere effectively to the substrate. High dark current values further attribute to the lower photoconversion efficiencies for DSSCs with thick ZnO films. High dark current values resulted from the penetration of electrolyte molecules into the inner layers of ZnO films with high thickness. Figure 33 shows the variations of photo-to-current conversion efficiencies for Rose Bengal and Rhodamine B sensitizer dyes, with film thickness. It was further observed that the fill factor of these DSSCs was found to increase with the film thickness to an optimum value.

Figure 33

Variations of cell efficiency (%) with film thickness (mm) for films sensitized with Rose Bengal and with Rhodamine B. Reprinted with permission from [97], © 2011 Elsevier Ltd

Chen et al. [178] synthesized olive-shaped ZnO nanocrystalline aggregates with submicron-sized hierarchical structures via a facile and economical aqueous solution method. The aggregates were composed of a large number of ZnO nanocrystals with 20 nm diameter. The overall dimensions of the submicron-sized hierarchical aggregates were approximately 150 nm × 300 nm. A bilayered photoanode fabricated using commercial ZnO nanoparticles with average diameter of 20 nm (underlayer) and submicron-sized aggregates as the light-scattering surface (overlayer) of variable thicknesses. Photoconversion efficiencies for the photoanodes with a underlayer thickness of 20 μm and overlayer thicknesses of 5, 10, 20, and 25 μm were found to be 3.65, 3.96, 4.43, and 3.93%, respectively. These photoconversion efficiencies were higher as compared to unilayered photoanodes constituting only either nanoparticles or submicron-sized ZnO aggregates.

Trilayered photoanode comprising ZnO nanowires with the average length and diameter of 3–4 μm and 200–400 nm, respectively, as the underlayer, small ZnO hierarchical microspheres of diameters 3–5 μm as the intermediate layer and large ZnO hierarchical microspheres of diameters 12–14 μm as the overlayer, was designed for DSSCs by Kang et al. [179]. The overall thickness of this trilayered film was 25–30 μm. As compared to unilayered and bilayered photoanodes, the photoconversion efficiency of 3.21% was observed for as fabricated Trilayered photoanode. Chanta et al. [180] studied the photoconversion efficiencies for DSSCs based on ZnO double-layer anti-reflection thin films prepared by depositing ZnO nanostructures of different morphologies prepared by RF-magneton sputtering (RF-ZnO) and sparking technique (SP-ZnO). Short-circuit current density (J _SC) of 5.80 mA/cm² and the highest photoconversion efficiency of 1.88% were recorded for photoanode comprising RF-ZnO as underlayer layer and SP-ZnO as overlayer. However, photoconversion efficiency was drastically reduced if the deposition order of the ZnO layers was reversed. The higher photoconversion efficiency for the SP-ZnO overlayer was attributed to reduced rate of light reflection and increased the extent of light transmission.

Recently, Lee et al. [149] reported a very high photoconversion efficiencies of cell efficiencies of 7.40 and 6.95% and short-circuit current densities of 15.84 and 15.35 mA/cm² for bilayered ZnO-based photoanodes with 10-μm-thick ZnO nanoparticles under layers and 5-μm-thick ZnO hexagonal plates and hexagonal cubes overlayers, respectively, using homemade organic dye, (E)-2-Cyano-3-(50-{8-[4-(diphenylamino)phenyl]-2,3-diphenylquinoxalin-5-yl}-[2,20-bithiophen]-5-yl)acrylic acid (DJ104). Hu et al. [181] fabricated bilayered photoanodes consisting of ZnO and ZnO/CNT nanostructure semiconductor films. As high as 6.25% photoconversion efficiency was reported for such bilayered photoanodes that was 35.57% higher than ZnO photoanode without multi-walled CNTs. The presence of the TiO₂ into ZnO thin films is also reported to enhance the adhesion and photovoltaic efficiencies of ZnO-based DSSCs [152, 182,183,184,185].

The enhancement in the photovoltaic performances was also reported for ZnO photoanodes with Au nanoparticles embedded between two layers of ZnO nanowires (AuNPs@ZnONW) synthesized through sequential hydrothermal ZnO. 20% more photoconversion was observed for AuNPs@ZnO NW DSSCs as compared to photoanodes with only ZnO nanowires [186]. Better performance of the AuNPs@ZnONW DSSCs was attributed to the formation of Schottky barrier at the interface of the Au NPs and ZnO nanowires. These Schottky barriers prevent the backtransfer of electrons from the ZnO to the dye molecules or electrolyte material [187]. Other metallic species like Ag, Pt, and Li also influence the photoconversion efficiencies of the various ZnO-based DSSCs [188,189,190,191].

Magnetic field effect on dye-sensitized ZnO nanorod-based solar cells

Cai et al. [32] analyzed the effect of external magnetic field on the photoconversion efficiencies of the ZnO-based DSSCs. Typical Nyquist plots for electrochemical impedance spectrum (EIS) and current–voltage characteristic plots for ZnO nanorod-based DSSCs sensitized by N719 dye as a function of magnetic field intensity are shown in Fig. 34.

Figure 34

a Nyquist plots with different magnetic field intensities, and b photocurrent–photovoltage characteristics for ZnO nanorod-based N719-DSSC. Reprinted with permission from [32], ©2012 Elsevier B.V

It was concluded that the J _SC and η for these cells increased even in the low magnetic field of magnitude less than 10 mT (inset of Fig. 34b), whereas parameters like V _OC and FF were not affected.

For DSSCs based on the Ru-based N719 dye as a photosensitizer, a strong spin–orbit coupling facilitates the process of intersystem crossing from heavy Ru atom of the dye molecules. This results in the additional injection of the electrons from a lower energy triplet state [32, 192, 193]. It also is quoted in the literature that the presence of an external magnetic field can affect the $e^{ - } - h^{ + }$ precession rates, resulting in the redistribution of populations among singlet and triplet states [194]. These results thus show that an external magnetic field can reduce electron transport resistance and change photocurrents by changing the ratios of the singlet and triplet states that are involved in excited charge transport processes. To conclude the presence and magnitude of applied magnetic field can potentially significantly improve the photovoltaic response of the ZnO-based DSSCs [32].

Effect of particle size

As discussed earlier, the extent of dye adsorption and electron diffusion are the key factors that affect conversion efficiency for ZnO-based DSSCs. Both of these factors are in turn dependent on the particle size of the ZnO nanomaterials [195,196,197,198,199]. Guo et al. [76] reported that ZnO aggregates with small-sized ZnO nanocrystals have a higher capability of dye-loading, while the ZnO aggregates with larger sized nanocrystals contribute to less recombination and faster electron diffusion of photogenerated electrons and redox species in the electrolyte. Thus, a proper engineering of the crystal size is compulsory for ZnO nanomaterials to be used as photoanodes in DSSCs because this can lead to higher light-to-electricity conversion efficiency [76]. In Fig. 35a, ZnO aggregates with crystal sizes 8–10, 25–30, and 45–50 nm are shown. It was observed that for ZnO aggregates with optimum crystal size 25–30 nm, maximum photocurrent efficiency of 4.54% was observed as compared to 3.78 and 4.10% of the DSSC for ZnO aggregates with a crystal size of 8–10 and 45–50 nm, respectively (Fig. 35b). There are many reports demonstrating that DSSCs based on ZnO aggregates with a crystal size of 10–20 nm are suitable for maximum photoconversion efficiency [195,196,197,198,199].

Figure 35

a Schematic representation of the dye adsorption, electron diffusion, along with the ZnO nanocrystals network, and recombination of electron and redox species in the electrolyte with different crystal sizes in the ZnO aggregate DSSCs; and b The J–V plots of ZnO aggregate-based DSSCs with different crystal sizes. Reprinted with permission from [76], © 2013 Elsevier Ltd

Nature of dye

According to Huang et al. [200], dyes with high molar extinction coefficients are more suitable than metal-based dyes for adsorption on photoanode semiconductor materials leading to better light-harnessing, and thereby resulting in more efficient solar cells. Table 3 clearly demonstrates that Ru metal-based N719 dye has been extensively used as a light sensitizer in ZnO-based DSSCs. Reports suggest that either metal-free indoline D149, D131, and D102 dyes or some organic dyes (Fig. 24) with a high molar extinction coefficient can be advantageous to anchor with chemically grown ZnO nanomaterials [96, 201]. Baviskar et al. [96] reported a high molar extinction coefficient value of 68,700 M⁻¹cm⁻¹ for D149 dye as compared to 13,900 M⁻¹cm⁻¹ for ruthenium complex dye N719. D149 and N719 sensitizer dyes were adsorbed to 5.49 × 10⁻⁸ and 1.72 × 10⁻⁸ M/cm² extent, respectively, onto the unit area of ZnO film (Fig. 36). Photoconversion efficiency for metal-free indoline D149 dye was 3.43%, whereas for N719 dye, efficiency was only 3.09%. Accordingly, a higher yield of photocurrent with metal-free D149 dye was observed, so Ru metal-based N719 dye may be due to better adsorption of D149 dye with ZnO nanoparticles [202].

Figure 36

Variations of optical absorption (solid) and quantum efficiency (hollow) spectra as a function of wavelength for metal-free indoline D149 dye (stars) and Ru metal N719 dye (circle), respectively. Reprinted with permission from [96], © 2013 Published by Elsevier B.V

Ru–metal complex dyes also are reported to be less stable and hence are not suitable candidates for fabricating eco-friendly photovoltaic devices. In various metal-free indoline dyes, such as D131, D102, and D149, the extent of adsorption and hence the photoconversion efficiency depends on the distance between the anchoring carboxylic group and the nearest LUMO located on the benzene ring adjacent to the core of the structures [203]. Strong bonding is observed for a shorter distance between the carboxylic group and the LUMO between the semiconducting ZnO thin layers and the dyes [204]. Depending on these factors, the adsorption value for D131 was highest, and the least for D149 dye [50] (Fig. 37).

Figure 37

Dye adsorption amounts of indoline D149, D131, and D102 dyes (nmol/cm²) on ZnO nanorods deposited on stainless steel wires with different loading times (min). Reprinted with permission from [50], © 2014 Elsevier Ltd

In addition to traditional metal-based and metal-free dyes, ZnO nanoparticles capped with monomeric N,N -dimethyl- N4-((pyridine-4yl) methylene) propaneamine and polymeric polyamine-4-pyridyl Schiff base dyes also exhibited excellent photoconversion efficiencies of 1.097 and 3.25%, respectively. The corresponding V _OC = 0.67, 0.70 V, J _SC = 2.3, 6.3 mA/cm², and FF = 0.712, 0.736, respectively, were recorded for monomer and polymer dye-capped ZnO nanoparticles by Singh et al. [148]. The high photoconversion efficiency of the polymer dye-capped ZnO nanoparticles was attributed to the higher degree of photoabsorption as compared to monomer dye and fast electronic charge transfers between polymer dye and ZnO nanoparticles.

Effect of post-deposition annealing temperature

It has been observed that post-annealing of ZnO nanomaterials affects dye-loading ability, which in turn influences the DSSC activities [53]. In a study represented in Table 3, an annealing temperature of 400–450 °C was applied for many of the ZnO nanomaterial-based DSSCs. ZnO thin films pasted on conducting substrate and annealed at a low temperature usually contain imperfections and impurities of the residual organics used during the pasting of the films [84]. The presence of these defects and impurities lowers dye-loading capacity. Furthermore, recombination of the electrons is rapidly injected from the dye molecules in the defect sites. This reduces the electron density in the conduction band of the ZnO, which results in lower V _OC. At high annealing temperature, the organic material is oxidized. In addition to this, high annealing temperature also improves the charge transport due to increased interconnection between nanomaterials. Still higher annealing temperature makes the ZnO photoelectrode denser, thereby lowering dye-loading ability. Heat treatment also improves crystallinity, which influences the charge transfer to the back contact, but lowers the cell efficiency due to low fill factor resulting from the back reaction between photoexcited electrons in the ZnO nanowires and ${\text{I}}_{3}^{ - }$ ions present in the electrolyte [65]. Lu et al. [81] fabricated DSSC based on the ZnO nanoparticles annealed at 300–450 °C. It was observed that ZnO DSSC annealed at 400 °C exhibited maximum photoenergy conversion efficiency of 3.92%.

An increase in annealing temperature from 200 to 400 °C improves the photocurrent density and open-circuit voltage. Photoconversion efficiency under these conditions was found to increase from 1.07 to 3.01%. At an annealing temperature above 400 °C, grain growth caused by the interconnection between photoelectrode particles has been observed. Fill factor for these DSSCs also depends on the annealing temperature of the ZnO photoanode. A lower fill factor value at a high annealing temperature is attributed to the fact that the photoelectrons in the ZnO photoanode reduce the ${\text{I}}_{3}^{ - }$ ions in the electrolyte [84].

Liu et al. [205] however reported that the conditions of high annealing temperature can be minimized using ZnO nanoparticles with low native defect concentrations as photoanode materials. Higher photoconversion efficiency of 2.54% was obtained at a low annealing temperature of 150 °C compared to a commonly used annealing temperature of 450 °C (η = 2.07) for ZnO nanoparticles with low defect density. Choudhury et al. [182] fabricated flexible ZnO-based DSSCs with a homogeneous compact ZnO layered photoanode deposited through a new hot-compress post-treatment technique at elevated temperatures. Better surface morphologies and photoconversion performances were recorded for as-deposited photoanodes as compared to conventional post-treatments. The effect of compression on ZnO thin films on film thickness, V _OC, J _SC, and photoconversion efficiencies is shown in Fig. 34a, b. Saturation in film thickness was observed after 100 MPa compression pressure (Fig. 38a). As the compression pressure is increased from 80 to 140 MPa, J _SC as well as photoconversion efficiency is increased (Fig. 38b).

Figure 38

Effect of compression pressure on a film thickness and open-circuit voltage (V _OC) and b short-circuit current density (J _SC) and photoconversion efficiency (η). Reprinted with permission from [182], © 2016 Elsevier B.V

Conclusions and future developments

ZnO nano-/microsemiconductor materials can be effectively utilized as photoanodic materials for dye-sensitized solar cells. ZnO nanorods and other related 1D material have led over other morphologies for DSSCs in the recent past because of greater specific surface area for better adsorption of various metal-based or metal-free photosensitizer dyes. Metal-free dyes with low-cost synthetic routes, easy dye structural modification, significantly higher extinction coefficients, and quite narrow absorption bands are used less frequently than metal-based dyes. In DSSCs, the ${\text{I}}^{ - } / {\text{I}}_{ 3}^{ - }$ redox couple plays a major role in the regeneration of the oxidized dye, and hence in the completion of the circuit. The major processes that occur in a DSSC are light absorption and excitation of the dye molecules, injection of the electrons from excited dye molecules into the conduction band of the highly porous thin layer of ZnO nanomaterials, with very high specific surface area coated on a conducting electrode (CE), charge collection and extraction at the conducting electrode, generation of electrical energy, and regeneration of the dye molecules. Various factors that affect the process are the thickness of the ZnO film coated on the substrate, the particle or crystallize size of the ZnO nanomaterials, post-deposition annealing treatment, and the presence of a magnetic field in the vicinity of the DSSCs. Compared to traditional silicon-based solar cells, which are more efficient at harnessing solar energy but are more costly to manufacture, ZnO-based dye-sensitized solar cells have a very low fabrication cost, but still present major drawbacks and obstacles for commercial applications because of very low photoconversion efficiencies.

Future research should focus on synthesizing highly porous ZnO semiconductor nanomaterials with a sufficiently large surface-to-volume ratio for better dye adsorption; synthesizing semiconductor materials doped with transition metals, p-type semiconductors, n-type semiconductors, carbon nanotubes (CNTs), and most importantly, graphene, replacing commonly used ITO and FTO substrates with better substrates such as ultrathin graphene films; and improving V _OC without any dye degradation for enhancing the cell performance for economical and viable solar energy devices of commercial use.