The climate change is indisputably the main problem of this century. Therefore, courageous policies of energy supply renovation towards green and renewable sources must be acted worldwide, as soon as possible. Particularly, concerning the photovoltaic technology, some serious challenges are exponentially arising, related to the availability and costs of raw materials and, at the same time, to the increased request of clean energy. For these reasons, an expanding number of scientific publications is nowadays devoted to find alternative, non-toxic and earth-abundant materials, applicable in the pre-existing technologies. In the field of thin-film solar cells (TFSC) technologies, the search of new solutions for each constituting layer is needed to significantly improve performances. The objective of this work is to explore possible materials and deposition processes modifications for the chalcogenide-based TFSC. The typical TFSC architecture is composed by a supporting material such as glass or polymers; a metallic or transparent conductive oxide (TCO) substrate; the main absorber layer made of complex chalcogenides stacked under a CdS buffer layer to create the p-n junction; two contact layers made of undoped and Al doped ZnO. Three main possible alternatives were investigated: [1] Engineering back contact substrates to improve performances and enlarging active areas for bifacial solar cells (BFSC) application; [2] Study of the RF-sputtering deposition parameters for inducing an oriented growth of the alternative absorber layer Sb2S3, made of abundant and available elements; [3] Development of low-cost, green and scalable deposition and processing methodologies for the realization of innovative solar devices based on CuInS2 (CIS) varnishes/inks, obtained via high-energy ball milling. [1] A BFSC, thanks to its two transparent contacts, can collect both direct and albedo light. They represent a smart solution to improve PV performances, limiting at the same time the amount of material employed to produce a single cell. BFSCs are generally less efficient than classical single-sided TFSC cells. This is due to the characteristics of the back contacts, usually TCOs, which are less performing than metals in the photocurrent extraction on mean and large area devices. For this reason, we worked on defining new engineered substrates which could possibly merge the optical properties of a TCO with the electrical properties of a metal. Specifically, Mo grid embedded in a TCO layer, were developed through a tricky photolithographic process followed by a selective chemical etching. After that, the selected TCO was deposited through RF-sputtering. The overall process was developed either with positive or negative photolithography, starting from both commercial and auto-produced substrates. The mixed layer granted an improvement of the cell active areas thanks to cooperative action of TCO and metal: the former provided good bifacial behaviour, the latter an effective collection of the carriers. In particular, Mo:FTO mixed layers presented improved performances as compared with a not engineered bifacial device. Nevertheless, some intrinsic limitations, affecting the Mo:FTO processing, make this engineered layer not suitable for the final application. Contrarily, Mo:ITO substrates resulted to be more proficient due to their quasi-ideal diode behaviour. For Mo:ITO back contacts overall active areas are greatly improved, over 2 cm2, while very low sheet and high shunt resistances are retained. Further studies, devoted to refining the processability of these embedded grids mixed contacts, are ongoing: the aim is to obtain the smoother surface to properly fabricate CuIn0.5Ga0.5Se2-based BFSC deposited via Low Temperature Pulsed Electron Deposition (LT-PED). [2] One main topics in the scientific research on TFSC is to find alternative solutions to abandon expensive and/or toxic elements like In, Ga, and Cd. One promising candidate is Sb2Se3. It is high absorbing p-type semiconductor which can in principle reach efficiencies of over 28%, according to the Shockley–Queisser theory. Thanks to its quasi-mono-dimensional ribbon-like covalent structure, the photogenerated carriers can almost freely flow along covalent bonded ribbons while they are hindered to flow perpendicularly, due to the presence of Van der Waals forces involving the neighbouring ribbon chains and leading to an unfavoured hopping-driven conduction mechanism. Therefore, in such a system, a key role is played by the crystallographic orientation of the AS layer: particularly, (hk0) directions, which identify ribbons completely lied on substrate, have to be avoided. Contrarily, (hk1), corresponding to slanted ribbons orientation, and (00l), characterized by ribbons normal to substrate, should be obtained in order to maximize the current extraction. Therefore, a theoretical study was done in order to evaluate which substrates can possibly influence a correct Sb2Se3 growth. Reasoning on epitaxial mismatches and lattice parameters, a wide selection of materials to be employed as back contact was put in comparison. Structural data were retrieved through Vesta software and ICSD simulations. The results were controversial: all tested metals, ZnO and FTO likely favour (hk0) reflections; moreover, ITO can produce random orientations. Despite this, an interesting fact was observed: RF-sputtering depositions, if prolonged in time, seem to allow Sb2Se3 self-orientation. Preliminary tests on this showed great improvements in the promotion of (hk1) and (00l) reflections. Lastly, a new engineered back contact was ideated, and novel studies are ongoing to verify its real effectiveness. Specifically, a distribution of ZnO nanowires will be prepared oriented normal to the substrate. Such layer could influence a growth of ribbons concordant with the ZnO nanowire axes, thus favouring (00l)-type reflections. First attempts seem to confirm the above statement. [3] A very comprehensive work concerned the development of innovative CIS-based solar devices, obtained through ultralow-cost liquid phase deposition. Homogeneous and sub-micrometrical CIS powders were obtained by mechanochemical reaction performed through high-energy planetary ball milling experiment. After that, CIS varnishes were prepared by means of liquid assisted refinement, adding 2-propanol as solvent and mixing a tiny amount of NaF (0.2% in mass) as dopant. These varnishes were deposited by simple techniques, like manual brush painting or drop-casting. Through a standardised process, composed by four steps, the powder distribution was then transformed into a film and completed with CdS buffer layer . However, about 2% of S was systematically lost, probably during the annealing step. Morphological analyses, carried out by AFM and SEM, confirmed that after this articulated process the films were flatter than as deposited samples, with highly reproducible thicknesses of about 2 µm. Remarkably, resistivity of 200 Ω·cm, average mobility of 10 cm2/(V∙s) and intrinsic carriers’ concentration of 1015 cm were achieved; surprisingly, these results were comparable with those obtained on CIS deposited by common vacuum techniques such as LT-PED. Two windows layers (ZnO and Al:ZnO) were then deposited via RF-sputtering, completing solar cell. 0.22% of efficiency, low short circuit currents and low open circuit voltages were achieved on Mo-based devices. Consequently, high series resistance were detected too, depending on the CIS morphology: the presence of very fine CIS powders leads to tremendous surface area vs. Volume ratio. Low doping diffusion rates due to a mild use of temperature during the film processing, do not helped to passivate grain boundaries and achieve better characteristics. For these reasons, two independent but simultaneous paths were undertaken: a. Stabilisation and diffusion of the NaF dopant prepared as a varnish; b. Study of a new process for the absorber layer re-crystallization. (a) A new NaF varnish was stabilised in 2-propanol by ball milling treatment and deposited through drop-casting in order to create a homogeneous distribution of separated crystalline units on the substrate. Different operations were discriminated depending on the chosen substrate (FTO or Mo). Afterwards, CIS was manually deposited, then treated as reported above; finally, the cell was completed by RF-sputtering. A slight diffusion of the alkaline dopant from the bottom of the solar absorber layer helped to triple the extraction currents; 0.60% and 0.70% efficiencies were achieved in both Mo and FTO based devices. Despite this, open circuit voltage as well as fill factors resulted far to be optimal. Such very poor performances can be possibly ascribed to very low absorber crystallinity and dopant diffusion. (b) Open-air annealing step of recrystallization has been modified introducing a close-system thermal treatment which can allow to work in selected atmosphere enhancing internal temperatures. Particularly, it was chosen to work in overpressure of sulphur being careful of calibrating the amount of sulphur as a function of the working temperature: indeed, too high sulphur partial pressure can poison CIS layer, too low pressures can lead to sulphur sublimation from the film and so CIS decomposition. The balance was found by working with 150 mg of sulphur at 600°C for 30 minutes. As a result, CIS crystallinity was greatly improved. Despite this, after some electrical tests, the semiconductor resulted totally compensated and an insulator-like behaviour with series resistances over 1 MΩ*cm was measured. A possible explanation can be given considering a detrimental action of 2-propanol: the solvent, which possibly remains trapped inside the film, during the annealing at high temperatures, can decompose, causing CIS layer carbon poisoning and so the loss of its PV properties. In order to solve these issues, a new water-based varnish is under investigation.
Il cambiamento climatico è di certo il maggior problema del secolo. Politiche coraggiose di rinnovamento energetico verso fonti rinnovabili devono essere attuate a livello mondiale il più velocemente possibile. In particolare, per quanto riguarda la tecnologia fotovoltaica, sorgono alcune gravi sfide, legate alla disponibilità e ai costi delle materie prime e, allo stesso tempo, alla crescente richiesta di energia pulita. Per questi motivi, un grande numero di pubblicazioni scientifiche è oggi dedicato a trovare materiali alternativi applicabili nelle tecnologie preesistenti. Tra queste, le celle solari a film sottile (CSFS) richiedono uno studio accurato su nuove soluzioni per ogni strato costitutivo al fine di migliorare significativamente le prestazioni. L’obiettivo di questo lavoro è quello di esplorare possibili materiali e processi di deposizione modifiche a CEFS a base di calcogenuri. L’architettura di partenza è composta da un materiale di supporto come vetro; un substrato metallico o composto da ossidi conduttivi (OC); la giunzione p-n (parte principale della cella) composta da strato assorbitore e buffer di CdS e due strati di contatto in ZnO non drogato e drogato con Al. Tre approcci principali sono stati investigati: [1] Ingegnerizzazione di substrati di back contact per migliorare le prestazioni e ampliare le aree attive per l’applicazione delle celle solari bifacciali; [2] Studio dei parametri di deposizione mediante tecnica di RF-sputtering per la crescita orientata di uno strato assorbente alternativo (Sb2S3); [3] Sviluppo di metodologie di deposizione e lavorazione a basso costo, ecologiche e scalabili per la realizzazione di dispositivi solari innovativi basati su vernici/inchiostri CuInS2 (CIS) ottenuti tramite macinazione a sfere ad alta energia. [1] Una cella bifacciale, grazie ai suoi due contatti trasparenti, può raccogliere sia l’albedo che la luce diretta. Rappresenta una soluzione intelligente per migliorare le prestazioni del fotovoltaico, limitando allo stesso tempo la quantità di materiale impiegato per produrre una singola cella. Le celle bifacciali sono generalmente meno efficienti delle celle classiche. Ciò è dovuto alle caratteristiche del contatto posteriore, solitamente gli OC sono meno performanti dei metalli nell’estrazione della fotocorrente nei dispositivi a grande area. Per questo motivo, abbiamo lavorato alla definizione di nuovi substrati ingegnerizzati che potrebbero fondere le proprietà ottiche di un TCO con le proprietà elettriche di un metallo. In particolare, una griglia di Mo viene incorporata in uno strato di OC, attraverso un processo fotolitografico seguito da un etching chimico selettivo. Subito dopo l’OC selezionato è depositato attraverso RF-sputtering. Il processo complessivo è stato sviluppato con processi fotolitografici positivi o negativi, a partire da substrati commerciali o auto-prodotti. Lo strato misto garantisce un miglioramento delle aree attive: l’OC fornisce un comportamento bifacciale, il Mo una efficace raccolta dei portatori. In particolare, gli strati misti Mo:FTO (F:SnO2) presentano prestazioni migliori rispetto ad un dispositivo bifacciale non ingegnerizzato. Tuttavia, alcune limitazioni del processo di preparazione di Mo:FTO, rendono questo strato ingegnerizzato non adatto per l’applicazione finale. Al contrario, i substrati Mo:ITO (In2-xSnxO3) risultarono essere i migliori grazie al comportamento quasi ideale dei diodi, al raggiungimento di aree attive superiori ai 2 cm2, alle resistenze di strato basse ed elevate resistenze di corto circuito. Ulteriori studi sono in corso per raffinare i processi di preparazione di questi contatti misti e per fabbricare correttamente una cella fotovoltaica a base di CuIn0.5Ga0.5Se2 depositata tramite Low Temperature Pulsed Electron Deposition (LT-PED). [2] Una delle soluzioni alternative dei film sottili è trovare materiali attivi che non contengano materiali costosi o tossici come In, Ga e Cd. Un candidato promettente è Sb2Se3. È un semiconduttore di tipo p ad alto assorbimento che può, teoricamente, raggiungere efficienze superiori al 28%, in accordo con la teoria Shockley–Queisser. Grazie alla sua struttura covalente quasi monodimensionale a nastro, le coppie fotogenerate possono muoversi quasi liberamente lungo i nastri covalenti mentre in altre direzioni sono ostacolati da conduzioni a hopping a causa di forze di Van der Waals. Pertanto, in un tale sistema, un ruolo chiave è svolto dall’orientamento di crescita cristallografica dello strato: in particolare, devono essere evitate le riflessioni (hk0), che identificano i nastri completamente adagiati sul substrato. Al contrario, sia l’orientamento dei nastri (hk1), cioè inclinato, sia quello (00l), caratterizzato da nastri normali al substrato, dovrebbero essere ottenuti per massimizzare l’estrazione di corrente. Pertanto, è stato effettuato uno studio teorico per valutare quali substrati possono eventualmente influenzare una corretta crescita di Sb2Se3. Ragionando sulla compatibilità epitassiale e i parametri reticolari, è stata messa a confronto un’ampia selezione di materiali per il contatto posteriore. I dati strutturali sono stati recuperati tramite il software Vesta e le simulazioni ICSD. I risultati sono stati controversi: tutti i metalli testati, ZnO e FTO favoriscono le riflessioni (hk0), inoltre, ITO può produrre orientamenti casuali. Nonostante ciò, è stato osservato un fatto interessante: le deposizioni RF-sputtering, se prolungate nel tempo, sembrano consentire l’auto-orientamento di Sb2Se3. I test preliminari hanno mostrato grandi miglioramenti nella promozione delle riflessioni (hk1) e (00l). Infine, è stato ideato un nuovo back contact ingegnerizzato, e studi in corso ne stanno verificando l’efficacia. Verrà preparata una distribuzione di nanofili di ZnO orientati normali al substrato. Tale strato potrebbe influenzare una crescita di nastri concordanti con gli assi dei nanofili di ZnO, favorendo così riflessioni di tipo (00l). I primi tentativi sembrano confermare l’affermazione di cui sopra. [3] Uno studio molto completo ha riguardato lo sviluppo di dispositivi solari innovativi basati su CIS ottenuti mediante deposizione in fase liquida a bassissimo costo. Polveri omogenee e sub micrometriche sono state ottenute mediante reazione meccanochimica eseguita mediante esperimenti di macinazione planetaria a sfere ad alta energia. Successivamente, le vernici CIS sono state preparate mediante raffinamento assistito da un solvente (2-propanolo) e miscelando una piccola quantità di NaF (0,2% in massa) come drogante. Queste vernici sono state depositate con tecniche semplici, come la pittura manuale a pennello o goccia a goccia. Attraverso un processo standardizzato, composto da quattro passaggi, la distribuzione della polvere è stata poi trasformata in un film e completata con uno strato di CdS per creare una giunzione p-n. Tuttavia, circa il 2% di S è stato sistematicamente perso, probabilmente durante la fase di annealing. L’analisi morfologica, effettuata da AFM e SEM, ha confermato che, dopo questo articolato processo, i film erano più lisci rispetto ai campioni depositati, con spessori altamente riproducibili di circa 2 µm. Sono state raggiunte una resistività notevole di 200 Ω·cm, una mobilità media di 10 cm2/(V∙s) e una concentrazione dei portatori intrinseci di 1015 cm. Sorprendentemente, questi risultati erano comparabili con quelli ottenuti su CIS depositati con le comuni tecniche sottovuoto come LT-PED. Due strati finestra (ZnO e Al:ZnO) sono stati quindi depositati tramite RF-sputtering, completando la cella solare. È stato raggiunto solo lo 0,22% di efficienza, oltre a basse correnti e tensioni sui dispositivi a base di Mo. Inoltre, sono state individuate anche elevate resistenze di serie dovute alla presenza di polveri CIS molto fini porta a un enorme rapporto superficie/volume. Il basso tasso di drogaggio e probabilmente una cattiva diffusione dovuta ad un blando utilizzo della temperatura durante la lavorazione del film, non hanno aiutato a passivare i bordi di grano e ad ottenere caratteristiche migliori. Per questo motivo due strade indipendenti ma simultanee sono state indagate: a) La stabilizzazione e diffusione di una vernice di dopante a base di NaF; b) Lo studio di nuovi processi di ricristallizzazione dello strato assorbitore. (a) Una nuova vernice NaF è stata stabilizzata in 2-propanolo mediante trattamento di macinazione a sfere. Una goccia è stata depositata sul substrato scelto (FTO o Mo) e diverse operazioni, differenti per ogni materiale, hanno portato alla preparazione di una distribuzione omogenea di unità cristalline separate. Successivamente, il CIS è stato depositato manualmente, quindi trattato come sopra riportato; infine, la cella è stata completata mediante RF-sputtering. Una leggera diffusione del drogante alcalino dal fondo dello strato assorbente solare ha contribuito a raddoppiare le correnti di estrazione. Le efficienze dello 0,60% e dello 0,70% sono state raggiunte sia nei dispositivi basati su Mo che FTO. Nonostante ciò, la tensione a circuito aperto e i fattori di riempimento non sono risultati ottimali. Tali prestazioni molto scarse possono essere eventualmente attribuite alla bassissima cristallinità dell’assorbitore e alla diffusione del drogante. (b) La fase di annealing in sistema aperto è stata modificata introducendo un trattamento termico a sistema chiuso che potesse consentire di lavorare in atmosfera selezionata aumentando le temperature interne. In particolare, si è scelto di lavorare in sovrapressione di zolfo avendo cura di calibrarne la quantità in funzione della temperatura di lavoro: una pressione parziale di troppo alta può infatti avvelenare lo strato CIS, una pressione troppo bassa può portare alla sublimazione dello zolfo o alla decomposizione del CIS. L’equilibrio è stato trovato lavorando con 150 mg di zolfo a 600°C per 30 minuti. La cristallinità CIS è stata notevolmente migliorata ma, nonostante ciò, dopo alcune prove elettriche, il semiconduttore è risultato totalmente compensato ed è stato misurato un comportamento di tipo isolante con resistenze in serie superiori a 1 MΩ*cm. Una possibile spiegazione può essere data considerando un’azione dannosa del 2-propanolo: il solvente rimasto intrappolato all’interno del film, durante l’annealing ad alte temperature, può decomporsi, causando avvelenamento da carbonio dello strato CIS e quindi perdita delle proprietà fotovoltaiche. Per risolvere questi problemi, è in fase di studio una nuova vernice a base d’acqua.
Sviluppo di celle fotovoltaiche bifacciali o a film sottile basate su calcogenuri complessi con deposizioni fisiche da vuoto e da fase liquida
Andrea, Sala
2022
Abstract
The climate change is indisputably the main problem of this century. Therefore, courageous policies of energy supply renovation towards green and renewable sources must be acted worldwide, as soon as possible. Particularly, concerning the photovoltaic technology, some serious challenges are exponentially arising, related to the availability and costs of raw materials and, at the same time, to the increased request of clean energy. For these reasons, an expanding number of scientific publications is nowadays devoted to find alternative, non-toxic and earth-abundant materials, applicable in the pre-existing technologies. In the field of thin-film solar cells (TFSC) technologies, the search of new solutions for each constituting layer is needed to significantly improve performances. The objective of this work is to explore possible materials and deposition processes modifications for the chalcogenide-based TFSC. The typical TFSC architecture is composed by a supporting material such as glass or polymers; a metallic or transparent conductive oxide (TCO) substrate; the main absorber layer made of complex chalcogenides stacked under a CdS buffer layer to create the p-n junction; two contact layers made of undoped and Al doped ZnO. Three main possible alternatives were investigated: [1] Engineering back contact substrates to improve performances and enlarging active areas for bifacial solar cells (BFSC) application; [2] Study of the RF-sputtering deposition parameters for inducing an oriented growth of the alternative absorber layer Sb2S3, made of abundant and available elements; [3] Development of low-cost, green and scalable deposition and processing methodologies for the realization of innovative solar devices based on CuInS2 (CIS) varnishes/inks, obtained via high-energy ball milling. [1] A BFSC, thanks to its two transparent contacts, can collect both direct and albedo light. They represent a smart solution to improve PV performances, limiting at the same time the amount of material employed to produce a single cell. BFSCs are generally less efficient than classical single-sided TFSC cells. This is due to the characteristics of the back contacts, usually TCOs, which are less performing than metals in the photocurrent extraction on mean and large area devices. For this reason, we worked on defining new engineered substrates which could possibly merge the optical properties of a TCO with the electrical properties of a metal. Specifically, Mo grid embedded in a TCO layer, were developed through a tricky photolithographic process followed by a selective chemical etching. After that, the selected TCO was deposited through RF-sputtering. The overall process was developed either with positive or negative photolithography, starting from both commercial and auto-produced substrates. The mixed layer granted an improvement of the cell active areas thanks to cooperative action of TCO and metal: the former provided good bifacial behaviour, the latter an effective collection of the carriers. In particular, Mo:FTO mixed layers presented improved performances as compared with a not engineered bifacial device. Nevertheless, some intrinsic limitations, affecting the Mo:FTO processing, make this engineered layer not suitable for the final application. Contrarily, Mo:ITO substrates resulted to be more proficient due to their quasi-ideal diode behaviour. For Mo:ITO back contacts overall active areas are greatly improved, over 2 cm2, while very low sheet and high shunt resistances are retained. Further studies, devoted to refining the processability of these embedded grids mixed contacts, are ongoing: the aim is to obtain the smoother surface to properly fabricate CuIn0.5Ga0.5Se2-based BFSC deposited via Low Temperature Pulsed Electron Deposition (LT-PED). [2] One main topics in the scientific research on TFSC is to find alternative solutions to abandon expensive and/or toxic elements like In, Ga, and Cd. One promising candidate is Sb2Se3. It is high absorbing p-type semiconductor which can in principle reach efficiencies of over 28%, according to the Shockley–Queisser theory. Thanks to its quasi-mono-dimensional ribbon-like covalent structure, the photogenerated carriers can almost freely flow along covalent bonded ribbons while they are hindered to flow perpendicularly, due to the presence of Van der Waals forces involving the neighbouring ribbon chains and leading to an unfavoured hopping-driven conduction mechanism. Therefore, in such a system, a key role is played by the crystallographic orientation of the AS layer: particularly, (hk0) directions, which identify ribbons completely lied on substrate, have to be avoided. Contrarily, (hk1), corresponding to slanted ribbons orientation, and (00l), characterized by ribbons normal to substrate, should be obtained in order to maximize the current extraction. Therefore, a theoretical study was done in order to evaluate which substrates can possibly influence a correct Sb2Se3 growth. Reasoning on epitaxial mismatches and lattice parameters, a wide selection of materials to be employed as back contact was put in comparison. Structural data were retrieved through Vesta software and ICSD simulations. The results were controversial: all tested metals, ZnO and FTO likely favour (hk0) reflections; moreover, ITO can produce random orientations. Despite this, an interesting fact was observed: RF-sputtering depositions, if prolonged in time, seem to allow Sb2Se3 self-orientation. Preliminary tests on this showed great improvements in the promotion of (hk1) and (00l) reflections. Lastly, a new engineered back contact was ideated, and novel studies are ongoing to verify its real effectiveness. Specifically, a distribution of ZnO nanowires will be prepared oriented normal to the substrate. Such layer could influence a growth of ribbons concordant with the ZnO nanowire axes, thus favouring (00l)-type reflections. First attempts seem to confirm the above statement. [3] A very comprehensive work concerned the development of innovative CIS-based solar devices, obtained through ultralow-cost liquid phase deposition. Homogeneous and sub-micrometrical CIS powders were obtained by mechanochemical reaction performed through high-energy planetary ball milling experiment. After that, CIS varnishes were prepared by means of liquid assisted refinement, adding 2-propanol as solvent and mixing a tiny amount of NaF (0.2% in mass) as dopant. These varnishes were deposited by simple techniques, like manual brush painting or drop-casting. Through a standardised process, composed by four steps, the powder distribution was then transformed into a film and completed with CdS buffer layer . However, about 2% of S was systematically lost, probably during the annealing step. Morphological analyses, carried out by AFM and SEM, confirmed that after this articulated process the films were flatter than as deposited samples, with highly reproducible thicknesses of about 2 µm. Remarkably, resistivity of 200 Ω·cm, average mobility of 10 cm2/(V∙s) and intrinsic carriers’ concentration of 1015 cm were achieved; surprisingly, these results were comparable with those obtained on CIS deposited by common vacuum techniques such as LT-PED. Two windows layers (ZnO and Al:ZnO) were then deposited via RF-sputtering, completing solar cell. 0.22% of efficiency, low short circuit currents and low open circuit voltages were achieved on Mo-based devices. Consequently, high series resistance were detected too, depending on the CIS morphology: the presence of very fine CIS powders leads to tremendous surface area vs. Volume ratio. Low doping diffusion rates due to a mild use of temperature during the film processing, do not helped to passivate grain boundaries and achieve better characteristics. For these reasons, two independent but simultaneous paths were undertaken: a. Stabilisation and diffusion of the NaF dopant prepared as a varnish; b. Study of a new process for the absorber layer re-crystallization. (a) A new NaF varnish was stabilised in 2-propanol by ball milling treatment and deposited through drop-casting in order to create a homogeneous distribution of separated crystalline units on the substrate. Different operations were discriminated depending on the chosen substrate (FTO or Mo). Afterwards, CIS was manually deposited, then treated as reported above; finally, the cell was completed by RF-sputtering. A slight diffusion of the alkaline dopant from the bottom of the solar absorber layer helped to triple the extraction currents; 0.60% and 0.70% efficiencies were achieved in both Mo and FTO based devices. Despite this, open circuit voltage as well as fill factors resulted far to be optimal. Such very poor performances can be possibly ascribed to very low absorber crystallinity and dopant diffusion. (b) Open-air annealing step of recrystallization has been modified introducing a close-system thermal treatment which can allow to work in selected atmosphere enhancing internal temperatures. Particularly, it was chosen to work in overpressure of sulphur being careful of calibrating the amount of sulphur as a function of the working temperature: indeed, too high sulphur partial pressure can poison CIS layer, too low pressures can lead to sulphur sublimation from the film and so CIS decomposition. The balance was found by working with 150 mg of sulphur at 600°C for 30 minutes. As a result, CIS crystallinity was greatly improved. Despite this, after some electrical tests, the semiconductor resulted totally compensated and an insulator-like behaviour with series resistances over 1 MΩ*cm was measured. A possible explanation can be given considering a detrimental action of 2-propanol: the solvent, which possibly remains trapped inside the film, during the annealing at high temperatures, can decompose, causing CIS layer carbon poisoning and so the loss of its PV properties. In order to solve these issues, a new water-based varnish is under investigation.File | Dimensione | Formato | |
---|---|---|---|
Tesi Sala Andrea Revisionata Definitiva.pdf
accesso aperto
Dimensione
8.56 MB
Formato
Adobe PDF
|
8.56 MB | Adobe PDF | Visualizza/Apri |
I documenti in UNITESI sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/20.500.14242/196702
URN:NBN:IT:UNIPR-196702