The steep increase of the computing power offered by the downscaling of complementary metal oxide semiconductor (CMOS) electronics is reaching physical limitations. In this framework, the development of efficient spintronic devices could represent a paradigm shift for information technology. Indeed, pure spin currents are dissipation-less and exploitable for low power electronics. At the same time, a semiconducting material allows for the realization of information processing in transistors. Unfortunately, the generation of spin currents in semiconductors requires ferromagnetic elements, whose injection efficiency is typically rather low due to conductivity mismatch. The emerging field of spin-orbitronics proposes to overcome this limitation exploiting spin-orbit coupling in non-magnetic materials for the electrical generation and detection of spin through spin to charge interconversion. In this context, our approach to spin-orbitronics is based on ferroelectric Rashba semiconductors (FERSC), in which multiple interconnected physical properties coexist. The thesis activity is part of a research project of Dott. Christian Rinaldi, working in the NaBis group (Nanomagnetism for Biology and Spintronics, Department of Physics, Politecnico di Milano). In particular, the work focused on the investigation of group IV tellurides GeTe and SnTe, both expected to belong to the class of FERSC. The father compound of FERSC is germanium telluride (GeTe), a CMOS-compatible ferroelectric semiconductor showing a giant Rashba effect at room temperature. By switching its ferroelectric polarization, the spin circulation in the Rashba bands can be inverted, paving the way to the electrical control of spin current generation and spin transport in all-in-one spintronic device. While the ferroelectric control of the Rashba effect in GeTe was already demonstrated, nothing was known about the ferroelectric control of spin-to-charge interconversion. To this aim, the thesis was largely devoted the growth of suitable heterostructures for spin pumping experiments in GeTe. Fe/GeTe heterostructures were grown in UHV by molecular beam epitaxy, and finely characterized by in-situ electron diffraction and photoelectron spectroscopy. The magnetic properties of Fe were measured by vibrating sample magnetometry and the ferroelectric switching of GeTe by electrical gating was studied and optimized. Iron was used as pure spin current injector in spin pumping experiments, but also as electrical gate. Measurements showed efficient spin-to-charge conversion in Fe/GeTe heterostructures, comparable to that of reference heavy metals like Pt. Noteworthy, such conversion is controlled in magnitude and sign by the ferroelectric polarization. First principles calculations suggest spin Hall effect as the main physical mechanism responsible for such interconversion. Some attention was also devoted to InGeTe, where the inclusion of indium was used to compensate for the high intrinsic p-type doping of GeTe, and the limit of In solubility was found. Finally, the FERSC nature of SnTe was investigated. It has been predicted to have long-living spin helix states associate to its Rashba bands in ultrathin films, which could be interesting for efficient spin generation and transport in semiconductors. SnTe(111) thin films were epitaxially grown on both BaF2 and GeTe and characterized by photoelectron spectroscopy and diffraction. SARPES experiments were performed at the beamline APE of ELETTRA synchrotron, revealing giant Rashba splitting as well as significant spin polarization of the valence band. These preliminary results demonstrate the potential of SnTe for spin-orbitronics. In conclusion, multifunctional FERSC can be exploited for spin logic-in-memory, offering the possibility to store information in the ferroelectric polarization state and to process it by exploiting ferroelectrically driven spin to charge interconversion.
L’incremento esponenziale che la potenza di calcolo ha avuto negli ultimi decenni è stato reso possibile da una continua miniaturizzazione dei suoi principali componenti, i transistori metallo-ossido-semiconduttore (CMOS), che stanno però raggiungendo il loro limite fisico. Un cambio di paradigma rispetto alle tecnologie attuali potrebbe essere rappresentato dalla messa a punto di dispositivi spintronici efficienti in cui le correnti di spin, trasportando informazione senza dissipare energia, permettano lo sviluppo di dispositivi a basso consumo di potenza. Le correnti di spin vengono però generate attraverso iniettori ferromagnetici, la cui efficienza di iniezione è tipicamente bassa. Per superare questo ostacolo, il campo emergente della spin-orbitronica propone l'utilizzo dell'interazione spin-orbita in materiali non magnetici per produrre e misurare correnti di spin, tramite conversione spin-carica e carica-spin. Il lavoro presentato in questa tesi si concentra sullo studio dei materiali semiconduttori Rashba ferroelettrici (FERSC), nei quali coesistono diverse proprietà fisiche interconnesse. L'attività è stata svolta nel contesto del progetto di ricerca del Dott. Christian Rinaldi, del gruppo NaBiS (Nanomagnetism for Biology and Spintronics del Dipartimento di Fisica, Politecnico di Milano). In particolare, il lavoro si è focalizzato sui tellururi del gruppo IV GeTe e SnTe, entrambi predetti come appartenenti alla classe dei FERSC. Il germanio tellurio è il capostipite di questa classe di materiali emergenti: è un semiconduttore ferroelettrico (compatibile con la tecnologia CMOS) e presenta un effetto Rashba gigante a temperatura ambiente. In precedenti lavori del gruppo è stato dimostrato che, invertendo la direzione della polarizzazione ferroelettrica, la circolazione degli spin nelle bande Rashba viene a sua volta invertita. Diversamente, il controllo della conversione spin-carica tramite l'inversione della polarizzazione ferroelettrica era rimasto ancora inesplorato, e verrà dimostrato in questo lavoro. La tesi tratta la crescita di eterostrutture Fe/GeTe per esperimenti di pompaggio di spin. Queste sono state cresciute in UHV tramite epitassia da fasci molecolari e caratterizzate con diffrazione elettronica e spettroscopia fotoelettronica in-situ. Le proprietà magnetiche del Fe sono state misurate con un magnetometro a campione vibrante, mentre l'inversione ferroelettrica del GeTe è stata studiata ed ottimizzata tramite un gate elettrico. Lo strato di ferro è stato utilizzato sia come iniettore di spin negli esperimenti di spin pumping, sia come gate elettrico. Le misure sulle eterostrutture di Fe/GeTe hanno dimostrato un'efficiente conversione spin-carica, comparabile a quella di materiali pesanti di riferimento come il Pt. É importante notare che l'intensità e il segno della conversione possono essere controllati tramite la polarizzazione ferroelettrica. Simulazioni sulle eterostrutture in esame suggeriscono l'effetto spin Hall come principale responsabile della conversione spin-carica di Fe/GeTe. Parte del lavoro è stato dedicato anche all'analisi del composto InGeTe, dove l'inserimento dell'indio nel GeTe è stato usato per compensare il suo alto drogaggio intrinseco di tipo p, ed è stato trovato il limite di solubilità dell'indio. Infine, sono state analizzate le proprietà FERSC di SnTe. Per film ultrasottili di SnTe, è stata predetta l'esistenza di eliche di spin a lunga durata, associate ai suoi stati Rashba, interessanti per la generazione ed il trasporto efficiente di spin. Film sottili di SnTe(111) sono stati cresciuti tramite epitassia da fascio molecolare su substrati di BaF2 e GeTe, e caratterizzati tramite spettroscopia e diffrazione fotoelettronica. Esperimenti SARPES, svolti presso la linea APE del sincrotrone ELETTRA, hanno dimostrato la presenza di un effetto Rashba gigante e una polarizzazione in spin della banda di valenza. Questi risultati preliminari dimostrano il potenziale di questo materiale per la spin-orbitronica. In conclusione, la multi-funzionalità dei FERSC può essere sfruttata per concentrare l’elemento di logica e di memoria di un dispositivo nello stesso spazio fisico. L'informazione può essere registrata nella polarizzazione ferroelettrica, mentre la computazione avverrebbe tramite la conversione spin carica, controllata dalla ferroelettricità.
Group IV tellurides for all-in-one spin devices
NOVATI, ALESSANDRO
2019/2020
Abstract
The steep increase of the computing power offered by the downscaling of complementary metal oxide semiconductor (CMOS) electronics is reaching physical limitations. In this framework, the development of efficient spintronic devices could represent a paradigm shift for information technology. Indeed, pure spin currents are dissipation-less and exploitable for low power electronics. At the same time, a semiconducting material allows for the realization of information processing in transistors. Unfortunately, the generation of spin currents in semiconductors requires ferromagnetic elements, whose injection efficiency is typically rather low due to conductivity mismatch. The emerging field of spin-orbitronics proposes to overcome this limitation exploiting spin-orbit coupling in non-magnetic materials for the electrical generation and detection of spin through spin to charge interconversion. In this context, our approach to spin-orbitronics is based on ferroelectric Rashba semiconductors (FERSC), in which multiple interconnected physical properties coexist. The thesis activity is part of a research project of Dott. Christian Rinaldi, working in the NaBis group (Nanomagnetism for Biology and Spintronics, Department of Physics, Politecnico di Milano). In particular, the work focused on the investigation of group IV tellurides GeTe and SnTe, both expected to belong to the class of FERSC. The father compound of FERSC is germanium telluride (GeTe), a CMOS-compatible ferroelectric semiconductor showing a giant Rashba effect at room temperature. By switching its ferroelectric polarization, the spin circulation in the Rashba bands can be inverted, paving the way to the electrical control of spin current generation and spin transport in all-in-one spintronic device. While the ferroelectric control of the Rashba effect in GeTe was already demonstrated, nothing was known about the ferroelectric control of spin-to-charge interconversion. To this aim, the thesis was largely devoted the growth of suitable heterostructures for spin pumping experiments in GeTe. Fe/GeTe heterostructures were grown in UHV by molecular beam epitaxy, and finely characterized by in-situ electron diffraction and photoelectron spectroscopy. The magnetic properties of Fe were measured by vibrating sample magnetometry and the ferroelectric switching of GeTe by electrical gating was studied and optimized. Iron was used as pure spin current injector in spin pumping experiments, but also as electrical gate. Measurements showed efficient spin-to-charge conversion in Fe/GeTe heterostructures, comparable to that of reference heavy metals like Pt. Noteworthy, such conversion is controlled in magnitude and sign by the ferroelectric polarization. First principles calculations suggest spin Hall effect as the main physical mechanism responsible for such interconversion. Some attention was also devoted to InGeTe, where the inclusion of indium was used to compensate for the high intrinsic p-type doping of GeTe, and the limit of In solubility was found. Finally, the FERSC nature of SnTe was investigated. It has been predicted to have long-living spin helix states associate to its Rashba bands in ultrathin films, which could be interesting for efficient spin generation and transport in semiconductors. SnTe(111) thin films were epitaxially grown on both BaF2 and GeTe and characterized by photoelectron spectroscopy and diffraction. SARPES experiments were performed at the beamline APE of ELETTRA synchrotron, revealing giant Rashba splitting as well as significant spin polarization of the valence band. These preliminary results demonstrate the potential of SnTe for spin-orbitronics. In conclusion, multifunctional FERSC can be exploited for spin logic-in-memory, offering the possibility to store information in the ferroelectric polarization state and to process it by exploiting ferroelectrically driven spin to charge interconversion.| File | Dimensione | Formato | |
|---|---|---|---|
|
Tesi_Alessandro_Novati.pdf
Open Access dal 08/04/2021
Descrizione: Testo della tesi
Dimensione
33.15 MB
Formato
Adobe PDF
|
33.15 MB | Adobe PDF | Visualizza/Apri |
I documenti in POLITesi sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/10589/153105