Literature Overview
The material system of Cu(In,Ga)Se 2 quaternary
CuInxGa1-xSe2 (CIGSe) quaternary is a p-type semiconductor material which is a member of the I-III-VI 2 chalcopyrites family [1] It has been known as an effective light-absorbing layer for thin film solar cell industry, and CIGSe thin film solar cells were first studied at the Bell laboratories in 1975 Much efforts have been done on CICSe solar cells, therefore the efficiency of CIGSe solar cells have reached up to 23.35 %, recently [2]
1.1.1 Introduction of the structural property
The basis for the structure of CIGSe quaternary is the structure of CISe ternary Some indium atoms in unit cell of CISe ternary are replaced by gallium atoms to form unit cell of CIGSe [1]
Fig 1.1: The unit cell of the chalcopyrite lattice structure a) CuInSe2, b) Cu(In,Ga)Se2[3].
CIGSe crystals have the tetragonal chalcopyrite lattice structure which drived from the cubic structure of the group IV semiconductor The unit cell is corresponds to two stacked zincblende unit cells, where copper, indium and gallium atoms are regularly ordered Each group
I (Cu) and group III (In or Ga) atom is tetrahedrally bonded to four group VI atoms (Se), and
2 each Se atom is coordinated is tetrahedrally coordinated to 2 group I atoms and 2 group III atoms [1] The lattice parameters of CuIn x Ga 1-x Se 2 are a=b= 5.696, value of c depend on value x composition of Ga element [4]
CIGSe semiconductors possess high optical absorbtion coefficient α > 10 5 cm -1 , and relatively low direct band gap energy value in the range 1.0-1.7 eV, resulting from the hybridization of Cu d-orbitals and Se p-orbitals in the valence bands [5,6]
The band gap energy of CuIn 1-x Ga x Se 2 compounds depend on the Ga/Ga+In concentration ratio, but also by the Cu content The optical band gap energy of CuInSe 2 is about 1.0 eV when In atoms are gradually substituted by gallium atoms, the band gap can be increase to 1.68 eV, which is the optical band gap energy of CuGaSe 2 [1] The band gap energies of the CuIn1-xGaxSe2 quaternary with different compositions from x= 0 to 1 are calculated based on an empirical ―bowing equation‖ [7,8]:
Eg= 1.010 +0.626x – 0.167x(1-x) (1) Where the so-called optical bowing coefficient is 0.167
The CuIn1-xGaxSe2 absorber layer is a p-type semiconductor with hole concentration about 10 16 - 10 18 cm -3 at room temperature The hole mobilities have been obtained values in the range of 15-150 cm 2 /Vs with single crystals and values of 5-50 cm 2 /Vs with polycrystalline films [9,10].
Overview of solar cell
1.2.1 The story of solar cell
In 1839, a French physicist-Alexander Edmond Becquerel discovered the technology directly producing electricity from the solar energy [11] This was the beginning of the solar cell technology-which open a new gate for the renewable energy technology
Solar cells are often divided into three main caregories which are called generations up to recent years:
- The first generation: Monocrystalline silicon (mono c-Si), polycrystalline silicon (poly c-Si) and armophous silicon cells
- The second generation: amorphous silicon (a-Si), CdTe and CuInS/CIGS
- The third generation: DSSCs, Perovskite and organic The first generation
The first silicon solar cell with an efficiency of 6% was studied at Bell laboratories and reported by Chapin, Fuller and Pearson in 1954 [12] One year later, the first commercial solar cells with 2% efficient were produced by Hoffman Electronics corporation using wafers of silicon (monocrystalline silicon) The following success of this technology was mostly founded on the further development of the Czochalski method in the 1940s enabling the production of high purity silicon [13], leading to the efficiency of these solar cells currently reaching to 26.7% [14] However, the cost of fabricating monocrystalline silicon solar cells are high because of the purification process of bulk In 1958, T Mandelkorn created n-on-p silicon solar cells which were better suited for space with more resistance to radiation damage at U.S Signal Corps Laboratories, and these cells were applied to a space satellite ― Vanguard 1‖ – a first solar powered satellite launched at this year with a 0.1W, 100 cm 2 solar panel [15]
4 Different from monocrystalline silicon solar cells, the solar cells made from polycrystalline silicon or amorphous silicon are much cheaper because they are much less pure than monocrystalline silicon, therefore the way slicon made is easier and simpler However, the highest efficient solar cell made from polycrystalline silicon is recorded of 22.3% [14], lower monocrystalline solar cell
Fig 1.2 shows structure of a silicon solar cell A silicon solar cell typically is a sandwich of two layers : the first layer is a positive layer ( p-type silicon ) which usually made by doping silicon with boron to form extra holes in the lattice The second layer is a negative layer ( n-type silicon ) which usually made by doping silicon with phosphorus to gain extra electrons in the lattice
Fig 1.2: The working of a silicon solar cell[16].
Some limitations of this generation : Silicon possesses an indirect band gap with a low absorption coefficient, so it is not a ideal material for a light conversion Besides, the silicon absorber layers require a high purity and a thickness up to around 200 àm, therefore the cost of fabrication is high and the product is less visual aesthetic
5 The second generation solar cells were studied with the purpose of reducing limitations of the first generation solar cells mentioned above The advantage of this generation is that Semiconductor materials used in this generation possess direct band gap
The second generation solar cells consist of absorber layers only 1 to 4 àm thick, therefore they are called as thin film solar cells These layers are much more thinner than silicon solar cells, so they have more visual aesthetic Absorber layers are deposited onto a large, inexpensive substrate such as glass, polymer or metal
The second generation solar cells have a lower fabricating cost than the first generation solar cell, however, the efficiencies of cells are lower, i.g amorphous Si (a-Si) solar cell (9.5 % ), CdTe/CdS solar cell (19.6%) and CIGSe solar cell (23 %) [14] Besides, these generation solar cells also exist another drawbacks such as using semiconductor materials which are either becoming rare or highly toxic
The structure of the second generation solar cell still relies on a p-n junction design as the first generation solar cells Therefore, it works as the first generation solar cells
The third generation solar cells emerged with efforts aiming to minimize the fabricating costs of first generation solar cells and improve toxicity materials of second generation solar cells These solar cells aren’t designed relying on the p-n junction as the previous generation solar cells
Dye sensitized solar cells (DSSCs)
6 Dye sensitized solar cells (DSSCs) are also called Gratzel cells which discoveried by Michael Gratzel and Brian O’regan in 1991 [17] These are low- cost solar cells, however, the conversion efficiencies are still quite low, around 11.9 % [14]
The basic structure of DSSC is shown in Fig 1.3 It consists of anode (photoelectrode), cathode (counter electrode), a dye sensitizer and a redox electrolyte The counter electrode is created by coating a thin layer of carbon, platinum or graphite on a conducting plate ( the conducting plate was made of glass plate on which a conducting layer of indium or fluoride doped tin oxide is deposited) The anode is also prepared by coating a thin layer of TiO2 on conducting plate This plate is then immersed in a dye sensitizer The anode and cathode are then joined and sealed with a redox electrolyte inside
Fig 1.3: The structure of a dye sensitized solar cells[18].
The working principle of DSSCs is shown in Fig 1.4
When sunlight shines on the cell, photosentizer absorbs photon and electrons are excited from ground level D to excited state D*
7 These excited electrons are moved to conduction band of TiO2 by oxidizing the dye sensitizer D + :
The electrons in conduction band of TiO2 are transported between TiO 2 particles and finally reach the cathode through the circuit
The oxidized dye sensitizer regains electron from electrolyte solution and oxidizes I - to I 3- :
The oxidized I 3- ion gets electron from carbon or platinum at cathode and reduce to I - ion:
Fig 1.4: Mechanism of dye sensitized solar cell[19].
8 The limitation of these solar cells are using liquid electrolytes, which cause short- term stability because of organic solvent evaporation and leakage, difficulty in sealing device and electrode corrosion.
Organic or polymer solar cells
These solar cells use thin films ( ~100 nm) of organic semiconductors including polymers such as copper phthalocyanin, polyphenlylene vinylene, carbon fullerene derivatives The basic structure of a organic or polymer cell is described in Fig 1.5The active layer include electron acceptor and electron donor materials, they are sandwiched between two metallic conductors When the semiconductor material absorbs a photon, a electron-hole pair is formed The charge trends to remain bound in the form of a exciton and is separated when the exciton diffuses to the donor-acceptor interface The advantages of these solar cells : absorber layers are materials having the quite high optical absorption coefficient Besides, the organic/polymer solar cells are inexpensive, light weight and flexible The disadvantages of organic/polymer solar cells : the efficiencies of these cells are relatively low ( around 11% ) [14] and they have low stability
Fig 1.5: Basic structure or Organic Solar Cell[20]
Quantum dot solar cells includes Schottky CQD solar cells and bulk heterojunction QD solar cells
- Schottky CQD solar cells: the quantum dot films could act as both absorbers and as the charge transport medium These solar cells work relying on illumination through a transparent ohmic contact to a p-type CQD film which formed rectifying junction with a Shallow work function metal and conversion efficiencies of cells reached over 3 %
- Bulk heterojunction QD solar cells: Different to Schottky CQD solar cells, these cells use a monolayer of molecular absorbers on a wide band gap semiconductor matrix, the depleted heterojunction architecture applies a highly doped n-type metal oxide in a p-n heterojunction with a p-type CQD film The CQD films have 50-400 nm thick
Perovskite solar cells are a new family of solar cells, belongs to the third generation The general formular of perovskite is ABX3, where A is monovalent, B is divalent ions and X is either O, C, CN or a halogen CH 3 NH 3 PbI 3 , CH 3 NH 3 PbBr 3 are the most common perovskites applied for solar cells
Thin film deposition techniques
A wide variety of deposition techniques have been applied to make CIGSe thin film These could be sorted by two main methods: The vacuum method and the non-vacuum method
The simpler, vacuum evaporation method, is shown in the figure 1.7 The system operates under a high vacuum (10 -6 torr or better) Material from the evaporation source is converted into the gaseous phase by heating or electron bombardment, and gaseous material then deposits on the substrate and its surroundings as a film Various substrate materials are used, depending on the subsequent application of the film that is to be deposited
Fig 1.7: Schematic of evaporation system
Backing Pump Cold Trap Diffusion Pump Vent Gas Crucible Charge Wafer
12 This is the simplest process for preparation of CIGSe absorber layer All elements are deposited simultaneously from separated sources at constant evaporation rates and substrate temperature (Fig 1.8) That is a reason lead to the film is always Cu poor which grown from this process possessed columnal grains typically less than 1 àm wide and the grains sizes trend to be smaller near the back contact
The two-stage process has been originally studied by Mickelsen and Chen [22,23]
In this deposition process, the bulks of the film were grown from a Cu-rich precursor layer at a substrate temperature in the range of 400-450 o C (stage 1), followed by a Cu- poor precursor layer deposited at 500-550 o C (stage 2) until the overall composition is copper deficient [24]
Films grown by the two-stage process have larger grain sizes in comparirion with the single stage process It is explained that a Cu-rich precursor layer deposited in the first stage formed CIGSe quaternary and Cu x Se binary impurities, which were assumed to be liquid at high temperature (~ 500 oC) Therefore, the film containing CIGSe and
CuxSe would then grow via a vapour-liquid-solid mechanism, which favours to increase grain size in film
This is a sequential process first proposed by Kessler et al [25] The CIGSe solar cells reaching the highest efficiency to date were fabricated with absorber layers grown with the three-stage process This process was studied by Kessler consisting of three stages:
- Stage 1: An (In,Ga)xSey precursor layer is deposited at low substrate temperature of about 330 o C
- Stage 2: Copper and Selenium elements are co-evaporated at a higher substrate temperature of around 500-560 o C until a Cu/In+Ga ratio of 1.15 at the endpoint to extend the grain size The increased substrate temperature load to an alloying of the element and the formation of the chalcopyrite phase
- Stage 3: a small amount of gallium, indium and selenium are co-evaporated at the same substrate temperature of stage 2 until the required overall copper deficiency of the film ( Cu/In+Ga ~ 0.84-0.88) is obtained
The absorber layers prepared with this process have a smooth surface and more crystallinity
Fig 1 8: Schematic illustration of different coevaporation recipes[26]
Sputtering is a common physical vapor deposition used by manufacturers of semiconductors This technique is possible to produce alloys of precise composition Sputtered
14 films exhibit excellent uniformity, density, purity and adhesion Sputtering method was first applied to prepare CuInSe 2 thin film by Chu et al [27] However, only the metals were deposited by sputtering while selenium elements were incorporated in the second step Therefore, this process is of them called the two-step process
The operation: The apparatus used for sputtering is outlined in the figure Basically, it consists of a bell jar which contains a reduced pressure-10 -1 to 10 -2 torr-of an inert gas, argon or xenon This gas is subjected to a potential drop of several kilovolts creating a glow discharge from which positive ions are accelerated towards the cathode(target) These high energy ions remove material from the cathode which then condenses on the surroundings, including the substrates to be coated, which are placed in a suitable position relative to the cathode The mechanism of sputtering or removal of material from the cathode, involves the transfer of momentum from the gaseous ions to the cathode in such a way that atoms or ions are ejected from the cathode
Fig 1.9: Schematic of sputtering system[28]
15 Chemical vapor deposition is a chemical process used to produce high quality, high performance, solid materials The process is often used in the semiconductor industry to produce thin films
The CVD method is developing into an extremely important way of making high purity thin films and coatings for industrial applications, especially in electronics, as well as for fundamental scientific research Conceptually, it is simple; precursor molecules containing the elements of interest are decomposed in the gas phase and the products deposit as thin films on every available object in the vicinity
There are various acronyms used to describe variations on the CVD technique; a commonly-used one is MOCVD which refers to the metal-organic nature of the precursors In many applications, such as the fabrication of multilayer semiconductor devices, it is necessary for the deposited films to have the correct structural orientation and to be coherent with the underlying layer Hence vapor phase epitaxy is essential in the growth mechanism
Fig 1.10: Schematic of chemical vapor deposition system
Fig 1.11: Schematic of chemical vapor deposition system
The characteristic of this method is that the films are made from the nanoparticles containing ink These nanoparticles were synthesized in a isolated step and dispersed then in a suitable solvent to form ink This ink is coated onto substrate by various methods such as spin-coating, dip casting, doctor blade, chemical spray, drop casting…
Some main method to synthesize nanoparticles as following:
- Hot inject and heat up methods: The both methods use organic ligands to control the nucleation and growth kinetics and they act as capping agents to provide colloidal stability However, In the hot inject method, the purpose is to separate nucleation from growth Many articles reported about synthesis of nanoparticles by this method such as: Tang et al [29] used oleylamine solution of selenium element at 80 o C to inject into other precursors Ahmadi et al.[30] used hexadecyleamine as coordinating solvent to synthesize CIGSe nanoparticles Li et al [31] synthesized CISe nanoparticles using dedecylthiol solvent as organic ligand and trioctylphosphine (TOP) as the solvent to dissolve selenium element nuclei of CuInSe2 appeared immediately after selenium solution injected, etc In the heat up method, the gradual heating of all precursor results in simultaneous nucleation, growth leading to form more polydisperse nanoparticles Similar to hot inject method, a variety of nanoparticles were synthesized by this approach: Chiang et al.[32] used oleylamine as a capping agent to synthesize CIGSe nanoparticles at the gram scale In 2014, Singh et al [33] report a large scale synthesis of CIGSe nanocrystals by heat up method using dodecylamine as a coordinating solvent
- Solvothermal: This method associates with the heating of solution containing precursors at high temperature and pressure during which the equilibrium is varied with temperature The solution containing precursors is heated in a autoclave ( a sealed vessel) in which the pressure and temperature exceed the atmostphere pressure and the boiling point of the solvent Some reports relate to this method such as: Chun et al [34] synthesized CIGS nanoparticles at a reaction temperature of 280 oC for 36 h Li et al [35] heated a solution consisting of ethylendiamine and precursors ( Se, CuCl2, InCl3) in autoclave at 180 oC for 15 h to obtain CISe nanocrystals Gu et al.[36] conducted experiment at a reaction temperature of 230 oC for 24 h through a solvothermal method to synthesized CIGSe nanoparticles, etc
- Hydrothermal method: this method is similar to sovothermal method However, the hydrothermal method uses aqueous solution containing precursors to synthesize nanocrystals Wu et al.[37] synthesized CuInSe2 in
2011 at 180oC for 1 h Shim et al synthesized CuInSe2 at 200oC for 12 h using acetic acid as a mineralizer [38] Ramkumar et al synthesized CuInSe2 at 150oC for 2 h using ethylendiamine as a capping agent [39]
- Polyol method: this approach was studied in the 1980s by Fievet et al [40] for preparation of micron and submicron size metal particles This method involves the preparation of metallic powders by reduction of inorganic compounds in liquid polyols ( ethylene glycol, diethylene glycol, triethylene glycol…) The role of these polyols: at first, it acts as a solvent to dissolve the solid precursors due to its rather high dielectric constant, then reduce the
Physical methods for characterizing solids
There are many physical methods for investigating the structures of solids In this dissertation, some methods are described:
19 X-ray diffraction is able to determine the precise atomic positions and therefore the bond lengths and angles of molecules within a single crystal are identified X-ray diffraction has its limitations and only gives an overall, average picture of a structure; it cannot usually identify localized defects or define the positions of small quantities of dopants However, when it is possible to use X-ray diffraction, it is extremely powerful; the results are very accurate, giving bond lengths to a few tens of picometres
In 1985, the German physicist Wilhelm Rửntgen discovered X-rays, and he was awarded the first Nobel Prize in Physics about this invention in 1901
The generation of X-ray is described as shown as Fig 1.11 A tungsten filament is electrically heated, emitting electrons which are accelerated by a high potential difference (20-50 kV) to strike a anode (metal target) which is water cooled The electrons from the inermost K shell of the metal target are knocked out by these bombarding electrons creating vacancies which are filled by electrons descending from the shells above The decrease in energy leads to appear radiation Therefore, the anode
20 emits a continous spectrum of white X-radiation, in which the intense X-ray peaks are
Kα, Kβ lines (Fig 1.12) Electrons descend from the L shell giving the Kα line, and electrons from the M shell giving the Kβ line The frequencies of the Kα and Kβ lines are at 154.18 pm and 71.07 pm, respectively
Fig 1.11: Schematic illustration of X-ray generation system[46]
Fig 1.12: (a) Section through an X-ray tube (b) An X-ray emission spectrum[46]
In 1912, Maxvon Laue used a crystal of copper sulfate as the diffraction grating which brang him the Nobel Prize in Physics in 1914 This discovery has been taken note by W.H and W.L Bragg They conducted experiments on using X-ray crystal diffraction as a way to determine structure of crystals, and the crystal structure of NaCl was first determined in 1913 W.L Bragg noted that X-ray diffraction behaves like a ―reflection‖ from the planes of the atoms within the crystal and that only at specific orientations of the crystal with respect to the source and detector are X-rays ―reflected‖ from the planes
There are three X-ray diffraction methods:
transmission Laue method: X-ray pass through the crystal and scattering
Film for record the data set behind of the crystal Some scattering beam transmit the crystal
Back-reflection Laue method: Film that have tiny hole set in front of the crystal X-ray pass through tiny hole on the film and diffracted behind of the crystal It came back to film and film record the data
22 Rotating-crystal method: Setting the crystal vertical with X-ray, Place the cylindrical film around the crystal and rotate it with setting axis Setting the axis of crystal and film become same
Powder method: Make minute size of sample and radiate X-ray It is same as single crystal state that rotating every axis
XPS is known as electron spectroscopy for chemical analysis (ESCA) This technique is widely used for analyzing the surface chemistry of materials It can
23 determine the elemental composition , empirical formular, chemical state and electronic state of the elements within a material
XPS spectra are recorded by irradiating a material with a beam of X-rays while simultaneously measuring the kinetic energy and number of electrons that emitted from the top 1-10 nm of material being analyzed A photoelectron spectrum is recorded by counting ejected electrons over a range of electron kinetic energies Peaks appear in the spectrum from atoms emitting electrons of a particular characteristic energy The energies and intensities of the photoelectron peaks enable identification and quantification of all surface elements ( except Hydro)
Fig 1.13: Basic components of a monochromatic XPS system [47]
Raman spectroscopy is widely applied to determine vibrational modes of modecules, therefore moleculars can be identified This technique based on inelastic scattering of monochromatic light, usually from a laser source
24 Operating principle: When a beam of light (laser source) hits a sample, photons are absorbed by the sample and scattered The majority of the scattered light has the frequency as the incident light and are called as Rayleigh scattering (elastic scattering) However, a tiny of the light scatters at the frequency different from the incident light and are known as Raman scattering ( inelastic scattering ) The shift in frequency between the incident light and Raman scattered light is equal to the frequency of a vibration of the scattering molecular, and hence it gives the information about the vibrational structure of the substance Therefore, Raman scattered lights are collected and then recorded by a detector In the contrary, The Rayleigh scattered lights are filtered out by either a notch filter or a band pass filter
Fig 1.14: Schematic illustration of a raman spectroscopy system[48]
25 Electron microscopy has been widely applied for characterizing solids to study morphology, structure and size, to examine defects and to determine the distribution of elements
The basic principle of the electron microscope is that a beam of accelerated electrons is used instead of visible light Because of the wave particle duality of electrons, they behave like electromagnetic radiation and, at high energies, have very short wavelength (λ)
The electron beam is produced by heating a tungsten filament, a lanthanum hexaboride (LnB 6 ) crystal or from a field emission gun (FEG), which uses a cathode of either tungsten or zirconium oxide The beam is focused by magnetic coil magnets in a high vacuum to a fine spot Detection can be by scintillation counter, film or CCD
Here, I introduce 2 techniques: Scanning electron microscopy and transmission electron microscopy
This technique is applied to describe a map of surface topography of materials such as catalysts, minerals and polymer It is useful for looking at particle size, crystal morphology, magnetic domains, and surface defects
SEM work as shematic diagram Fig 1.15 Electrons are produced at high voltage by a electron gun in vacuum chamber, accelerated down passed through a combination of lenses and aperture to generate a finely focused beam of electrons which impact on the atoms of the sample’s surface As these electrons interact with the sample, they produce back scattered electrons (the reflection of electrons by elastic scattering), the secondary electrons from inelastic scattering, characteristic X-
26 ray (the production of secondary electrons) An SEM image is mainly recorded from the secondary electrons which collected by detector The number of electrons determines the brightness of the image The samples should be coated with platinium or similar substance, electrical thing to prevent charge building up on the surface
Fig 1.15: Schematic illustration of scanning electron microscopy system [49]
Transmission electron microscopy is a powerful tool for material science, it is applied to observe the features such as structure, size and morphology of very small specimens It was first developed by German Scientists Max Knoll and Ernst Ruska in 1931 and has evolved over the years to become a common technique that is used
27 globally in science and engineering to look at micro and nanoparticles This technique can be used to observe samples at a much higher magnification and resolution than can be achieved with a light microscope because wavelength of an electrons is much smaller than that of light It also provides higher resolution images than a scanning electron microscopy due to a very high energy beam of electrons passed through a very thin specimen
Green and low-cost synthesis of CIGSe nanoparticles using ethanol as a solvent by a
Introduction
CIGSe compounds are effective light-absorbing materials for thin film solar cells, possessing a direct band gap in the range of 1.0-1.7 eV [5,6] Thin films of CIGSe compounds have been prepared on surfaces of various substrates for applications to solar cells [53-61] For example solar cells prepared from CIGSe thin films deposited by thermal evaporation showed high energy conversion efficiency of around 23.35 % [2] Although the CIGSe solar cells fabricated by thermal evaporation show high conversion efficiency, the deposition of these elements is very expensive due to the requirement of sophisticated equipment In addition, the fabricated solar cells are affected by the impurities from the substrates due to the high temperature demand of vacuum-based techniques To minimize these problems, efforts have been made to develop low- cost and environment-friendly synthesis of CIGSe nanoparticles (NPs)
Recently, CIGSe nanoparticles synthesized by solution processes have been reported [34,42,62-64] In such approaches, Cu, In, Ga, Se elements [34,63].or their salts [59,63] are dissolved in organic solvents and react with each other to form precipitates of CIGSe These solution processes typically involve toxic organic solvents, such as hydrazine, ethylenediamine, and polyetheramine, as well as reaction temperatures of up to 280 o C Many of these methods also rely on complex reaction set-ups and procedures incorporating autoclaves or glove-boxes The drawbacks of the reported solution methods are apparent, which can be eliminated by applying ultrasonic irradiation to the methods Acoustic cavitation, i.e the formation, growth, and implosive collapse of a vast amount of vapor cavities in a liquid under ultrasonic irradiation leads to the formation of localized hot spots, where the reaction could take place
32 The sonochemical method is considered suitable for the preparation of nanoparticles with a range of functional characteristics [65] A CIGSe nanoparticle synthesis by sonochemistry has been reported with a reaction time of 4 hr [43], in which a mixture of ethylene glycol and hydrazine (very toxic) need to be applied to completely dissolve the precursors at 110 0 C The as-synthesized NPs showed a chalcopyrite structure only after annealing at 500 0 C From a sonochemistry point of view, ethylene glycol is unsuitable solvent because of its relatively high viscosity (~ 16.1 mPa at 20 o C), which does not favor the development of vapor cavities to their full sizes [66,67] This was probably the reason why the reaction temperature in [43] needs to be set to 110 o C Compared to ethylene glycol, ethanol would be a more suitable solvent in sonochemical processes because of its substantially lower viscosity (~1.2 mPa at 20 o C), so that the number of full-size-developed cavities would be sufficient to facilitate reactions at moderate temperatures In addition, ethanol is a readily available low-cost, environment-friendly solvent, because it could be generated from renewable sources Sodium borohydride (NaBH4) helps dissolve and reduce Se powders in ethanol with minimal impact on the environment
In this work, a slightly modified sonochemical procedure for CIGSe nanoparticle synthesis using ethanol as the only solvent was investigated to develop a cheaper and eco-friendlier nanoparticle synthesis process for commercial applications Although a reduction agent (NaBH 4 ) must be added to the solvent, this approach is substantially less toxic than the previously reported methods
From the literature survey [55], it is evident that the CIGSe thin film solar cell exhibits the best efficiency when the atomic ratio of Ga/ In+Ga is around 0.3 Therefore, in the present work, we targeted to synthesize the CIGSe nanoparticles with a composition of CuIn 0.7 Ga 0.3 Se 2
Experiment details
33 All chemicals were used as-received Copper (I) chloride (CuCl, Aldrich 99.995%), indium (III) chloride (InCl 3 , Aldrich 98%), gallium (III) nitrate hydrate (Ga(NO 3 ) 3 xH 2 O, Aldrich
99.999%), selenium (Se, Aldrich 99.99%), ethyl alcohol (C2H5OH 99.9%, Fisher Scientific), and sodium borohydride (NaBH 4 , Aldrich 99.99% ) were used as chemicals for the synthesis, and a VCX 750 Sonics & Materials ultrasonic processor was incorporated
The elemental selenium was first dissolved in two beakers containing solutions of NaBH4 in ethanol until these solutions became transparent (it took about 15 minutes) Subsequently, InCl 3 and Ga(NO3)3.xH2O were added to each beaker separately The transparent solutions changed to turbid colloidal suspensions with yellow precipitates These two suspensions were mixed with each other, and finally CuCl was added, the yellow precipitates changed quickly to deep black precipitates The mixture was ultrasonicated (f = 20 kHz, P = 300 W) at room temperature for different time intervals of 2 hr, 3.5 hr, 4.5 hr, and 5 hr, respectively The precipitates were filtered, washed twice with ethanol and once with deionized water, and vacuum evaporated at 40 oC for 8 hr
The as-synthesized samples were characterized using a range of techniques X-ray photoelectron spectroscopy (XPS, Thermo Scientific K-Alpha) was performed using Al Kα X- rays as the excitation source The relevant core levels (Cu 2p, In 3d, Ga 2p, and Se 3d) were used to quantify the elemental atomic composition using the sensitivity factors from the database The crystal structure of the as-synthesized products was characterized by powder X-ray diffraction (XRD, PANalytical, X’Pert-PRO MPD) and Raman spectroscopy (XploRA Plus Horiba) The XRD patterns of the powder samples were measured using Cu/Kα radiation The Raman spectra were obtained from 100 cm -1 to 600 cm -1 at room temperature The chemical compositions of the
34 CIGSe nanoparticles were measured by inductively coupled plasma emission spectroscopy (ICPS, ICPS 8100 Shimadzu system) with standard HCl/HNO 3 digestion
The size and morphology of the nanoparticles were characterized by transmission electron microscopy (TEM, Philips CM-200), and scanning electron microscopy (SEM, Hitachi S-4800) The optical properties of as-synthesized CIGSe nanoparticles were measured by ultraviolet- visible-near infrared (UV-Vis-NIR, Cary 5000, Varian) over the range 300–1800 nm.
Results and discussion
The experimental procedures described above were designed (and later verified by the results obtained) based on the following tentatively proposed reaction pathways :
The first reaction clearly shows that the role of NaBH4 is to completely reduce Se to HSe - , which is soluble in ethanol The second and third reactions took place parallel in separate beakers to form Ga2Se3 and In2Se3 Finally, the fourth and fifth reactions took place simultaneously when these two beakers were mixed together and CuCl was added, so that the freshly forming Cu 2 Se could not be aged to its stable structure Although details about the effects of ultrasonification on these two reactions are unclear, they are supported by the XRD results In fact, the XRD results
Cu 2 Se + xIn 2 Se 3 + (1-x)Ga 2 Se 3 2CuIn x Ga 1-x Se2 (5) ultrasonication ultrasonication
35 shown in the next section revealed that trace amounts of Cu2Se compound as impurities in the products with 2-4.5 hr ultrasonic irradiation, which seem to disappear prolonging the ultrasonic irradiation to 5 hr (see Fig 2.1) These results suggest that the ultrasonic effect, beside other effects, favors the decomposition of Cu 2 Se to promote reaction 5
2.3.2 Effect of the ultrasound time interval
Fig 2.1 presents XRD patterns of the as-synthesized products with ultrasonification times of
2 hr, 3.5 hr, 4.5 hr, and 5 hr, respectively We could easily observe that the ultrasonification time of 2 hr was not enough to form the typical crystal structure of the Cu(InxGa1-x)Se2 compound Prolonging the ultrasonification time to 3.5 hr and 4.5 hr resulted in much better crystallinity with the expected structure, but small amounts of Cu2Se crystals also appeared Prolonging the ultrasonification time further to 5 hr yielded single-phase tetragonal CuIn0.7Ga0.3Se2 nanocrystals with five main index planes of (112), (220), (312), (400) and (332) without any signs of Cu 2 Se Therefore, under the conditions investigated in this work, an ultrasonification time of 5 hr is the most suitable time interval to form the Cu(In x Ga 1-x )Se 2 compounds with the expected crystalline structure
Fig 2.1: XRD patterns of the CIGSe nanoparticles synthesized with difference ultrasonification time
It is worthwhile to note that SEM images did not indicate any effects of the ultrasonification time intervals on the spherical shape of the as-synthesized nanocrystals between 2 and 5 hr (see Fig 2.2) Similarly, this ultrasonification time range had an insignificant effect on the reaction temperatures, which were between 61–64 o C
Fig 2.2: SEM images of CIGSe NPs with ultrasonification time of 2 hr (a), 3.5 hr (b), 4.5 hr (c) and 5 hr (d)
The as-synthesized products with 5 hr ultrasonification were characterized using additional methods Fig 2.3 presents a typical Raman spectrum of CuIn 0.7 Ga 0.3 Se 2 nanocrystals The dominant CIGSe mode was observed clearly at 177-178 cm -1 , which was somewhat higher than the 173-174 cm -1 for pure CuInSe 2 , due to the presence of gallium This is in accordance with the literature, which indicates that the CIGSe mode frequency increases linearly with increasing Ga content [68,69] Beside the dominant CIGSe mode, no additional modes appeared at around 260 cm -1 , indicating that Cu2Se is not an impurity in that product
Fig 2.3: Raman spectra of as-prepared CuIn0.7Ga0.3Se2 NPs with ultrasonification time of 5 hr
The elemental composition of the as-synthesized products (5 hr ultrasonification) was analyzed by the ICPS and calculated from the XPS results, showing only small deviations from each other Table 1 shows that the composition of the synthesized product is indeed CuIn0.7Ga0.3Se2.
Table 1: Chemical compositions of the as-synthesized CIGSe compounds
Composition measured by XPS (atomic ratio, Cu/In/Ga/Se )
(atomic ratio, Cu/In/Ga/Se )
The composition was further confirmed using the Vegard’s law [70,71] to evaluate the XRD results For Cu(InxGa1-x)Se2, Vegard’s law could be expressed as follows:
39 a CuIn Ga Se x a CuInSe x a CuGaSe x x (1 )
The unit cell parameters of CuInxGa1-xSe2 with x = 0, x = 1, x = 0.7 taken from the Joint Committee on Powder Diffraction Standards (JCPDS) and those calculated from XRD patterns of the as-synthesized products with a 5 hr ultrasonification time are listed in Table 2
Table 2: Lattice parameters of the Cu(InxGa1-x)Se2 with x = 0, x = 1, x = 0.7 and as- synthesized CIGSe compounds
Comparing the data listed in Table 2 and applying the equation (6) for the unit cell parameters, the value (x) of 0.68 was obtained for the as-synthesized product, i.e., CuIn0.68Ga0.32Se2, which was in accordance with the ICPS and XPS results
The valence states of the elements in the as-synthesized product can be determined by evaluating the observed binding energies presented in Fig 2.4 The binding energies were corrected for specimen charging by referencing the C 1s to 284.6 eV The calculated binding energies for Cu (2p3/2), In (3d5/2), Ga (2p3/2) and Se (3d5/2) core levels are 932.26 eV, 444.9 eV, 1118.25 eV, and 54.63 eV, respectively, which are in good agreement with the values reported in the literature [72-74] Fig 4b illustrates the Cu 2p core level spectrum with two intense peaks at
40 932.26 eV and 951.68 eV, corresponding to Cu 2p3/2 and Cu 2p1/2, respectively The full widths at half maximum (FWHM) calculated for Cu 2p 3/2 was 1.27 eV This value is consistent with the value reported for Cu + [75] Furthermore, the binding energy for Cu 2+ at 934.76 eV as a typical satellite peak was not detected by XPS spectrum So, we can conclude that the chemical valence of copper in the compound is +1 and only Cu +1 is present in CIGSe compound [74]
Fig 2.4: a) XPS survey spectrum and Cu 2p (b), In 3d (c), Ga 2p (d) , and Se 3d (e ) core-level spectrum of as-synthesized CuIn0.7Ga0.3Se2 nanoparticles
The morphology of the as-synthesized product was examined by SEM and TEM The images revealed uniformly quasi-spherical particles with an average size of approximately 8 nm as in Fig 2.2 and Fig 2.5a The agglomeration of the as-synthesized CIGSe nanoparticles took
41 place because of the absence of organic capping agents, which is commonly observed in inorganic nanoparticles synthesized by sonochemical methods The TEM image in Fig 2.5b showed that after being annealed at 500 o C under nitrogen and selenium atmosphere for 1 hr, the sizes of as-prepared CIGSe nanoparticles increased slightly to 17–30 nm It is worthwhile to note that for CIGSe thin film solar cell manufacture, a grain size in the range 1-2 μm is expected, i.e an annealing step might be favorable for thin films However, the annealing conditions should be studied in details
Fig 2.5: TEM images of CIGSe nanoparticles with ultrasonification time of 5 hr before (a) and after (b) annealed at 500 o C
The XRD pattern of annealed products in Fig 2.6 shows that its crystallinity was enhanced substantially compared to the original one Moreover, some new peaks appear indicating the presence of other index planes of CIGSe nanocrystal in addition to the five planes observed in the as-synthesized products (without annealing)
Fig 2.6: XRD pattern of CuIn0.7Ga0.3Se2 NPs before and after annealed at 500 oC
The UV-Vis-NIR absorption spectrum of annealed CIGSe nanoparticles in Fig 2.7a revealed an absorption peak centered at approximately 490 nm The band gap, Eg, of the CIGSe compound was calculated from transmittance spectra using the following equation [76]:
Where α is the absorption coefficient, A is a constant, h is Plank’s constant, ν is the light frequency, n = ẵ for a direct allowed transition, i.e., in this case, (h) 2 is proportional to the photon energy h In Fig 2.7b, (h) 2 of the annealed product was plotted as a function of the corresponding photon energy, h , resulting in a band gap of Eg ~ 1.2-1.3 eV, which similar to the band gaps of Cu(In x Ga 1-x )Se 2 quaternary compounds reported in the literature
Fig 2.7: Absorption spectrum (a) and plot of (αhν)2 versus photon energy hν (b) of the annealed
The formation of CIGSe thin film by a nanocrystals ink based spin-coating method
Introduction
The chalcopyrite CuIn x Ga 1-x Se 2 (CIGSe) compounds have been known as semiconducting materials, which possess direct band gap in the range of 1.0 – 1.7 eV and high optical absorption coefficient α > 10 5 cm -1 [5,6] Therefore, much efforts have been made to prepare CIGSe thin film as an effective light-absorbing layer for thin film solar cell industry, and the efficiency of
CIGSe solar cell has reached up to 23.35 % recently [14] However, the cost of fabrication is very high and energy consumption is significant, because these thin films were prepared by vacuum-based deposition methods such as sputtering and co-evaporation [61,77-86] which require sophisticated equipment On the other hand, the impurities from the substrates also significantly affect the performance of thin film solar cells, which results from high temperature employed in vacuum-based techniques
In order to minimize these issues, many studies have been done on non-vacuum methods such as printing [33,64,87,88] , dip coating [89], drop casting [90], supersonic kinetic spraying [91,92], spin coating [93,94], electro-deposition [95,96], doctor blading [97,98], solvolthermal [99], milling process [100] and hot injection [101] to develop low-cost manufacturing method of CIGSe thin film However, these methods are solution-based processes typically involving toxic organic solvents such as trioctylphosphine, tributylphosphine, polyetheramine, ethylenediamine and hydrazine (very toxic if inhaled), which are harmful to the environment as well as people’s health, or the solvents having a high boiling point such as oleyamine, octadecane, polyetheramine resulting in high reaction temperatures up to 280 o C and producing carbon residues in the resulting thin films
In this paper, we present a new methodology to prepare thin films from CIGSe nanocrystals by spin coating method Herein, CIGSe nanocrystals were synthesized using ―green‖ solvents – ethanol by sonochemistry The fabrication of a thin film by spin coating is challenged by the need to find out a suitable but less toxic solvent compared to hydrazine as previous reports, into which the as-synthesized CIGSe nanocrystals could be well dispersed resulting in stable
"ink" solutions [43,102] Furthermore, in manufacture of CIGSe thin film solar cell, a grain size is expected in the range 1 – 2 àm Thus, we also attempted to optimize the annealing step to improve CIGSe grain size of resulting thin film It is worthwhile to note that no long hydrocarbon chains containing ligands were applied in any of the process involved, so this is a surfactant-free procedure
According to the literature survey [83,103], the composition of CuInxGa1-xSe2 compound should be CuIn 0.7 Ga 0.3 Se 2 to give the best efficiency in CIGSe thin film solar cells Therefore, in the present work, we targeted to prepare thin film from the as-synthesized CuIn0.7Ga0.3Se2 nanocrystals.
Experiment detail
All chemicals CuCl (99.99%), InCl3 (99.999%), Ga(NO3)3 xH2O (99.9%), Se (99.99%) sodium borohydride (NaBH 4 ), 2-propanol (C 3 H 8 O, 99.5%), ethyl alcohol (C 2 H 5 OH, 99.5 %) and 2-methoxyethanol (C3H8O2, 99.3%) were products of Sigma aldrich and be used as-received The applied equipments included a VCX 750 Sonic & Materials ultrasonic processor (P= 300 W, f= 20 kHz), a furnace machine (Lenton Thermal Designs) and a spin coater ACE-200
The thin film was fabricated by ink solution based coating method In order to prepare an ink solution, we studied dispersing the as-synthesized nanocrystals in three different solvents: 2-
46 propanol, 2-methoxyethanol and solvent mixture of 2-propanol and 2-methoxyethanol (2:1 v/v ratio) Then, these mixtures were constantly stirred for 24 h Subsequently, these colloidal solutions were kept stable for one month to observe the stability of ink solution This step aims to find out the suitable solvent as a medium of dispersion for CIGSe ink solution
The as-prepared ink solution was coated on Mo-coated soda lime glass by spin coater (6000 rpm, 40 s) to form thin film The spin coated films were then annealed under nitrogen and selenium atmosphere with various temperature programs to optimize grain size and density of the CIGSe thin film The furnace machine is type of two zone furnace Selenium powder was put in first zone and CIGSe film was in second zone Temperature of first zone was always controlled at around 240 o C in all various annealing mode to selenium powder forming selenium vapor supplying to CIGSe film In the second zone, temperature program was changed in each annealing mode as follows:
The products were analyzed by various techniques The Raman spectroscopy (XploRA plus Horiba), powder X-ray diffraction (XRD, PANalytical, X’Pert-PRO MPD) and thin film X-ray reflectometry (PANalytical, X’Pert-PRO) were used to characterize the crystal structure of the as-synthesized product The XRD patterns were recorded using Cu/Kα radiation (λ= 1.056) The Raman spectra were measured at room temperatureusing 532 nm Laser as an excitation light source X-ray photoelectron spectroscopy (XPS) measurements were carried out on Thermo scientific K-alpha spectrometer using Al Kα monochromatized radiation to investigate the valence states of the elements in the products Inductively coupled plasma atomic emission spectroscopy (ICPS, ICPS 8100 Shimadzu system) was employed to analyze composition of elements in the as-synthesized products The transmission electron microscopy (TEM, Philips
48 CM-200) and scanning electron microscopy (SEM, Hitachi S-4800) were used to characterize the size and morphology of nanoparticles and thin film The transmittance spectra were measured on a Cary 5000 UV-vis-NIR spectrophotometer (Varian) to calculate the band gap energy (Eg) of the as-synthesized CIGSe thin film Hall effect measurement was carried out on Nanometrics hall measurement (HL5500) to characterize the electrical properties of the as-synthesized product.
Results and discussion
The images in Fig 3.1 revealed the level of stability of each solution (after being kept stable in one month) when the as–synthesized CIGSe nanocrystals were dispersed in three different solvents: 2-propanol, 2-methoxyethanol and solvent mixture of 2-propanol and 2- methoxyethanol (2:1 v/v ratio)
Fig 3.1: The images of colloidal solutions with different dispersion mediums: 2-propanol (a), 2- methoxyethanol (b) and solvent mixture of 2-propanol and 2-methoxyethanol (c) after being kept in 1 month
Fig 3.1a shows that there is a slight agglomeration of colloids in 2-methoxyethanol solvent
On the contrary, these colloids seem to be dispersed well in 2-propanol solvent and solvent
49 mixture of 2-propanol : 2-methoxyethanol, hence, the ink solutions were stable as shown in Fig 3.1b&c This might be because the value of surface tension of 2-methoxyethanol (42.8) is higher than 2-propanol (23.3), thus it is more difficult to disperse CIGSe nanocrystals in 2- methoxyethanol solvent However, the spin coated film from the 2-propanol solvent based ink solution was not uniform (Fig 3.2a), and a little bit of solvent still remained on film even if we increased the speed to 8000 rpm, it is perhaps because the value of viscosity of 2-propanol is high (2.04 cP) Indeed, for solvent mixture of 2-propanol and 2-methoxyethanol (the viscosity is 1.72 cP), the film formed really smooth and no solvent remained after spin coating (Fig 3.2b)
Because of that, this solvent mixture was chosen as dispersion medium to form CIGSe ink solution and all the thin films were deposited by spin coating from this ink solution used for the annealing step
Fig 3 2: The images of the spin coated films by 2-propanol solvent based ink solution (a), and solvent mixture based ink solution (b)
Three different annealing modes were conducted to optimize grain size as well as density of the CIGSe thin film For the thin film of the first annealing mode, we could easily observe that
50 the size of the as-synthesized CIGSe nanoparticles in thin film increased significantly, but the density of thin film was not good, many voids appeared in film (Fig 3.3) Maybe it is due to the rapid evaporator of the solvent at 500 o C leaving a lot of holes in thin film
Fig 3 3: Cross-section SEM image of CIGSe thin film with the first annealing mode
However, in the second annealing mode, the image in Fig 3.4 showed that the density of film was better than the first annealing mode This may be explained as follows: temperature was hold at 90 o C in first 10 min helping the evaporator of solvent occurring slowly, beside, temperature was kept at 215 o C in next 10 min favoring the melting of selenium in CIGSe compound gradually taking place, and thus had the time to stick nanoparticles together
51 However, temperature of 450 o C in the last 40 min was not enough to expand the size of CIGSe nanoparticles as the first annealing mode Indeed, for the thin film of the third annealing mode, when temperature of the last 40 min was changed into 500 o C, the size of CIGSe nanoparticles increased much more (Fig 3.5) Fig 3.5 also showed the thickness of thin film be about 1 àm, this is in range of thickness of CIGSe absorber layer for fabricating CIGSe solar cell
Fig 3 4: Cross-section SEM image of CIGSe thin film with the second annealing mode
52 Fig 3.6 presents surface morphology of thin film of the third annealing mode, the image shows grain sizes of the CIGSe compound in the range of 200-500 nm, significantly increasing in comparison with the as-synthesized CIGSe nanoparticles As such, the third annealing mode is most suitable among the three annealing modes tested, even though some voids still exist in thin film However, this is the nature of method, which prepared the ink solution based thin film
Fig 3 5: Cross-section SEM image of CIGSe thin film with the third annealing mode
53 The thin film annealed with the third annealing mode was characterized with XRD to confirm the structure of CIGSe compound Fig 3.7 shows that the XRD pattern of CIGSe compound was obtained as expected Beside, some others peaks also appeared indicating the presence of Mo substrate and MoSe2 phase forming in annealing process However, the formation of a MoSe 2 layer at the back interface between CIGSe thin film and Mo substrate under a selenium atmosphere has been reported in literatures [104-106] The presence of MoSe2 layer is beneficial for an increase in Ohmic contact at the CIGSe/Mo back interface, and it also supports the adhesion of the CIGSe layer on the underlying back contact [107-110]
Fig 3 6: Surface SEM image of CIGSe thin film with the third annealing mode
Fig 3 7: XRD pattern of CIGSe thin film after being annealed with the third annealing mode
Hall effect measurement was conducted with magnetic field intensity of 0.32 T and the current of 0.0021 mA The result showed that the as-prepared CIGSe absorber layer exhibited p- type conductivity with hole concentration (n p ) of 7.876 ˟ 10 17 cm -3 , the carrier mobility (à) of 22.2 cm 2 /V-s and resistivity (ρ) of 0.3574 Ohm-cm
The optical property of the annealed CIGSe thin film was analyzed by UV-Vis-NIR spectroscopy The bandgap energy (Eg) value was evaluated from the measured transmittance spectra data by using Tauc’s relation [111]:
55 Where α is the absorption coefficient, hν is the photon energy In Fig 3.8, the band gap energy of the film was estimated about ~ 1.3-1.4 eV by extrapolating (h) 2 vs h plot This value is close to the CuIn 0.7 Ga 0.3 Se 2 compound in the literatures
Fig 3 8: Plot of (αhν)2 versus photon energy hν of CIGSe thin film after being annealed with the third annealing mode
In summary, CuIn0.7Ga0.3Se2 nanoparticles were synthesized sonochemically using a substantially greener procedure at room temperature with ethanol as the only solvent, NaBH 4 as a reducing agent, and facilitated by 5 hr ultrasonic irradiation The as-synthesized products possess typical chalcopyrite crystalline structure The reaction scheme proposed in this paper briefly outlined the reaction pathways, revealing the role of NaBH4 as a reducing agent and ultrasonic irradiation in decomposing the intermediate Cu2Se in the products The non-toxic solvent used in the synthesis with no additional heating makes the developed method cheaper and ―greener‖ than the previously reported methods
The CIGSe thin films were prepared from the as-synthesized nanoparticles by spin- coating method using solvent mixture of 2-propanol and 2-methoxyethanol for the stable ink solution as a dispersion medium The films annealed with temperature program: 90 o C in the first
10 min, 215 o C in the next 10 min and 500 o C in the last 40 min were realized that the grain size and density improved dramatically The results obtained from characterizations by X-ray diffraction, SEM, Hall Effect and UV-Vis revealed that the thin films possess typical tetragonal crystalline structure and similar band gap to those reported for CIGSe thin film Besides, the thickness of the film was around 1 àm, which is in range of thickness of absorber layer for fabricating CIGSe thin film solar cells
[1] J.W L Shay, Ternary Chalcopyrite Semiconductors, Pergamon, Oxford, (1975)