684 research outputs found
PFAS: a new threat to human health Multi-analyte method development and background levels monitoring
No full text“PFAS are defined as fluorinated substances that contain at least one fully fluorinated methyl
or methylene carbon atom (without any H/Cl/Br/I atom attached to it), i.e., with a few noted
exceptions, any chemical with at least a perfluorinated methyl group (−CF3) or a perfluorinated methylene group (−CF2−) is a PFAS”. The “noted exceptions” refer to a carbon atom
with a H/Cl/Br/I atom attached to it (Organisation for Economic Co-operation and Development
(OECD), 2021; Wang et al., 2021).
The family of compounds denoted by the acronym PFAS encompass:
• perfluoroalkyl substances, which are defined as aliphatic substances for which all of the H
atoms attached to C atoms in the non-fluorinated substance from which they are notionally
derived have been replaced by F atoms, except those H atoms whose substitution would
modify the nature of any functional groups present;
• polyfluoroalkyl substances, defined here as aliphatic substances for which all H atoms attached to at least one (but not all) C atoms have been replaced by F atoms, in such a manner
that they contain the perfluoroalkyl moiety CnF2n+1− (Buck et al., 2011).
The PFAS molecular structure (R−X) consists of a hydrophobic alkyl chain, R, of varying
length (typically C4−C16) and a hydrophilic end group, X (Figure 4) (EFSA, 2020). The hydrophobic part may be fully [R = F(CF2)n−] or partially fluorinated. The hydrophilic end
group can be neutral, positively, or negatively charged. The resulting substances are non-ionic, cationic, or anionic surface-active agents due to their amphiphilic character (Table 1) (EFSA,
2020).The properties of PFAS depend on the characteristics of the fluorine atom and the nature of the
carbon-fluorine bond. Fluorine belongs to the group of halogens which possess seven valence
electrons. This means they only need one electron to complete the octet. They have high electron affinity, ionization energy, and electronegativity. Fluorine is the element of the periodic
table with the highest electronegativity which gives it a great ability to attract bonding electrons.
For this reason, the bond with carbon has a decidedly more polar character than that of the
carbon-hydrogen bond and has an electric dipole inversion. Fluorine forms the strongest single
bond with carbon due to the optimal overlap of their orbitals during bond formation. This results
from the similarity in energy levels between fluorine and carbon orbitals, especially since they
belong to the same period in the periodic table. Each carbon-fluorine bond increases the strength
of other C−F bonds related to the same carbon atom. Furthermore, being very small, fluorine
has an occupational space like that of hydrogen and guarantees a small steric encumbrance. It
has low polarizability, and this is reflected in the fact that fluorocarbons have weak intermolecular attractive forces, providing these compounds with a low surface energy, lower than that of
the respective hydrocarbons.
PFAS are characterized by high thermal stability and low reactivity, they can be both hydrophobic and lipophobic and their chemical- physical properties vary with the length of the carbon
chain and the functional groups. These characteristics make them of great interest for industrial
applications (Buck et al., 2011; Wang et al., 2017).
In 2018, the "Global PFC Group" coordinated by the Organisation for Economic Co-operation
and Development (OECD) and the United Nations Environment Programme (UNEP), released
a comprehensive list of more than 4700 PFAS compounds. These compounds were identified
as containing either a −CnF2n− (n ≥ 3) or −CnF2nOCmF2m− (n and m ≥ 1) moiety and were confirmed or presumed to have been present in the global market (Figure 5) (Organisation for Economic Co-operation and Development (OECD), 2018, 2021)
Lettera di Alessandra
No full textUn ritratto critico dell'opera di Alessandra Carnaroli, autrice fra le più apprezzate delle ultime generazioni della poesia di ricerca. La sezione a lei dedicata, nel numero della rivista, contiene inoltre saggi di Cecilia Bello Minciacchi, Andrea Cortellessa, e Ivan Schiavone; e vari inediti dell'autrice. Il saggio è pubblicato con lo pseudonimo di Tommaso Ottonieri.A critical portrait of the work of Alessandra Carnaroli, author of the most appreciated in the latest generations of italian research poetry. Published under the pseudonym Tommaso Ottonieri
Editorial: Dysmetabolism, obesity, and inflammation: Three prominent actors in the drama of major neuropsychiatric disorders
No full text
Una facezia di Poggio nell'«Epirota» di Tommaso de Mezzo
No full textPrinted in 1483, Tommaso de Mezzo’s «Epirota» belongs to the second time of the
latin humanistic comedy of the Quattrocento, characterized by a strictly imitation of
Plaute (and also of Terence), more strong than in the texts of the first half of the century.
The first section of this paper presents Tommaso de Mezzo’s life and works, and plot,
types and subjects of «Epirota». The second part of the article offers the analysis of some
scenes of the comedy, in which the author clearly employs one of Bracciolini’s «Facetiae»
Analog resistive switching devices optimized for deep learning applications
No full textIn today’s rapidly evolving landscape, Artificial Intelligence (AI) stands at the forefront of transformative technologies, revolutionizing operations across various sectors. This evolution has been facilitated by the optimization of digital hardware accelerators, crucial for enhancing AI capabilities. However, conventional digital accelerators, following the Von-Neumann architecture, face inherent limitations due to the separation of memory and logic units, leading to energy inefficiencies and operational latencies. In response to these challenges, crossbar arrays of resistive switching devices, or Resistive Processing Units (RPUs), have emerged as promising alternatives to traditional digital accelerators. Integrated within the Back-End-Of-the-Line (BEOL) of chips, these arrays offer high-density storage of Neural Networks’ (NNs) parameters directly on the chip. Besides their non-volatile memory functionality, RPUs serve as analog signal processors, executing Vector-Matrix Multiplications (VMM) fully in parallel with O(1) time complexity, significantly enhancing the efficiency of arithmetic computations in AI applications by mitigating memory access. While RPUs have shown competitive performance in NN inference tasks, their utilization for training NNs has been hindered by various non-idealities. To be compatible with both inference and training applications, the technology inside the RPU must meet stringent programming criteria, such as a fast, low-power, linear, and symmetric resistive-switching capability. Filamentary Resistive-switching Random Access Memory (ReRAM) devices, particularly those based on a Conductive-Metal-Oxide (CMO)/HfO2 material structure, have emerged as promising candidates for addressing these challenges. Recent optimizations in material properties have mitigated non-idealities, making them viable for both inference and training tasks. This dissertation investigates ReRAM devices based on the CMO/HfO2 material structure, focusing on key material properties essential for fast, analog resistive switching. Through material and processing optimizations, improvements in unit cell scaling and operational power reduction during programming/reading measurements were achieved. These enhanced devices were integrated into small-scale crossbar array architectures, demonstrating parallel weight updates for NN training acceleration, thus contributing to bridging the gap between research on inference and training accelerators
Development and characterization of RRAM crosspoint arrays for neuromorphic computing
No full textLAUREA MAGISTRALECon il termine "Neuromorphic computing" si intende una branca mista
dell’informatica e dell’elettronica, che si e’ sviluppata al fine di imitare le
interazioni ad alta efficientza energetica che occorrono nelle reti neurali degli
esseri umani. L’elemento alla base dei circuiti neuromorfici e’ il memristore,
cioe’ un dispositivo elettronico caratterizzato da una curva I-V "pinzata", i cui
stati resistivi possono essere dunque modulati al variare di alcuni parametri
relativi alle sue operazioni: nel caso delle STT-RAM si tratta della polarizzazione
di spin della corrente, nelle Fe-RAM della polarizzazione dello strato
di materiale ferroelettrico, nelle memorie a resistenza "switching" dell’impulso
elettrico comandato ai suoi terminali. Tra le tecnologie appena menzionate,
l’ultima e’ la piu’ adatta ad essere integrata in architetture neuromorfiche
emergenti, come il crosspoint array, cioe’ una matrice di elementi memristivi,
le cui righe e colonne definiscono rispettivamente le wordline e le bitline della
memoria. I dispositivi a resistenza "switching", sia le PCM-RAM che le RRAM,
possono essere facilmente implementati in questo tipo di architetture, perche’
hanno solo due terminali di accesso e, piu’ in particolare, le RRAM, sono il piu’
semplice tipo di memristore da fabbricare, in quanto consiste in uno strato di
materiale ossido incapsulato tra un metallo inerte ed uno ossidabile. L’architettura crosspoint puo’ essere disegnata per applicazioni di salvataggio
dati, propenendosi come uno schema per ottenere al contempo proprieta’ di velocita’ e di densita’ di integrazione intermedie tra le soluzioni ad alta densita’
piu’ comuni, come le HDD o le flash, e le memorie a semiconduttore piu’ veloci
ma costose, come le DRAM, definendo cosi una nuova classe di sistemi di memorie
chiamato SCM, "Storage Class Memory". Oppure puo’ essere progettata
per applicazioni di calcolo, perche’ proprio grazie alla naturale proprieta’ per
cui la resistenza del dispositivo puo’ essere modulata in vari stati analogici,
semplicemente applicando diversi impulsi elettrici ai suoi capi, se tali dispositivi
vengono disposti in implementazioni a forma di matrice, sara’ possibile eseguire
il calcolo di alcuni algoritmi in soli due passaggi, mentre invece utilizzando
le convenzionali architetture di computazione di tipo Von Neumann, dove
memoria e unita’ aritmetica sono fisicamente separate, l’algoritmo avrebbe
richiesto l’iterazione di molti piu’ passaggi computazionali. Il primo consiste
nel programmare ciascun elemento della matrice nello stato resistivo desiderato,
mentre il secondo nell’applicazione di una tensione di lettura alle celle della
matrice. La corrente raccolta alla fine di ogni bitline si rivela essere il prodotto
tra la tensione del vettore di lettura e la mappa delle conduttanze degli elementi
della matrice; in questo modo lasciamo alla natura stessa il compito di calcolare
la moltiplicazione, che si genera di conseguenza a quanto dettato dalla legge di Ohm. L’attivita’ sperimentale si e’ focalizzata su due aspetti: la fabbricazione
e la caratterizzazione elettrica di memorie emergenti a resistenza "switching",
nell’intenzione di ottimizzare le proprieta’ elettriche delle RRAM verso una
implementazione efficace all’interno di una matrice crosspoint. Le attivita’ di
processo e di caratterizzazione topografica sono state condotte nella cleanroom
"PoliFab", il centro di micro e nano-tecnologia del Politecnico di Milano, mentre
le misure elettriche sono state eseguite nel laboratorio di dispositivi elettronici, al Dipartimento di Elettronica, Informazione e Bioingegneria.
Il capitolo 1 introduce i problemi relativi al ridimensionamento delle componenti
elettroniche nei circuiti, specificando come questa situazione abbia
provocato, negli anni piu’ recenti, una svolta verso lo sviluppo di schemi
computazionali alternativi alla convenzionale architettura di Von Neumann.
Prima di discutere delle tecnologie SCM piu’ mature e piu’ promettenti per
raggiungere performance in velocita’ e costi intermedie tra i dischi rigidi e le
DRAM, viene presentata la gerarchia di memoria odierna, indicando i principi
di funzionamento di queste memorie.
Il capitolo 2 descrive i principi teorici sulle tecniche di fabbricazione impiegate
durante l’attivita’ sperimentale condotta nella cleanroom PoliFab, includendo
i passaggi per la pulizia, la litografia EBL e ottica, il RIE e l’etching chimico,
e le soluzioni per la deposizione di film sottili, come la CVD e l’evaporazione.
Il capitolo 3 presenta gli strumenti impiegati per la valutazione della qualita’
di processo, sia in termini di caratterizzazione fisica, e cioe’ lo studio topografico
delle matrici di dispositivi disegnate, sia in termini di caratterizzazione elettrica.
In questo capitolo viene anche presentato il setup impostato per eseguire le
misure elettriche.
Al capitolo 4 comincia la relazione sull’attivita’ di tesi, in particolare
specificando le ricette adottate per la fabbricazione di RRAM disegnate da EBL
e indicando i risultati ottenuti dalla caratterizzazione quasi-statica. Alcune
misure effettuate su un array crossbar riveleranno che c’era bisogno di ulteriore
ricerca per migliorare rapidamente le performance dei dispositivi, nell’ottica di disporli all’interno di matrici di tipo crosspoint. Il capitolo 5 descrive i processi di fabbricazione delle RRAM disegnate da
litografia ottica, evidenziando i grandi vantaggi di tempistica permessi da questa
tecnica litografica. E’ stata eseguita una piu’ approfondita caratterizzazione elettrica per questi dispositivi, che include analisi statistiche piu’ dettagliate.
Sono state studiate sia tecnologie a base di un ossido di tipo SiOx, sia HfOx.
Il capitolo 6 presenta dispositivi RRAM Volatili, fabbricati in PoliFab
con Ag al top electrode, a partire dai processi e dalle strutture di array gia’
sviluppate in precedenza per la fabbricazione di dispositivi Non Volatili. I
risultati delle misure quasi-statiche e impulsate su questi selettori a corrente
autolimitata a base di HfOx sono stati cosi’ interessanti che vi si e’ dedicato
un paper accettato alla conferenza ICECS 2019.
Il capitolo 7 indica quali soluzioni sono state individuate per migliorare
le proprieta’ elettriche delle RRAM non volatili, verso un’ efficace implementazione
all’interno delle matrici crosspoint, focalizzandosi sui principali problemi
riscontrati: l’"overshoot" di corrente dopo l’ electroforming e le alte tensioni
a cui avviene il breakdown dell’ossido. In questo capitolo vengono riportati
i risultati dei trattamenti di ricottura per diminuire le tensioni di forming e
le caratterizzazioni elettriche effettuate su RRAM fabbricate con una linea di
C in piu’, ingegnerizzata per superare le difficolta’ riscontrare nel resettare le
celle dopo il forming.
Il capitolo 8 e’ un resoconto delle misure elettriche condotte su una matrice
crosspoint di dimensione 8x8, in cui vengono anche indicate le strategie
sviluppate per raggiungere l’obiettivo finale della mia tesi, che consiste nello
sviluppo di matrici crosspoint di celle a resistenza "switching" programmabili a
piacimento in diversi valori analogici di resistenza. Questo risultato e’ stato
dimostrato in un crosspoint ridotto, fatto di 16 dispositivi (4x4).Neuromorphic computing is a mixed branch of IT and electronics developed
to emulate the low power interactions occurring between neural networks inside
the human brain. The core electronic element at the basis of most neuromorphic
hardware is a memristor, i.e. a device characterized by a pinched I-V characteristics,
whose resistive state can be modulated by varying some parameters related
to its operations: in the case of STT-RAM it is the current spin polarization,
in Fe-RAM the ferroelectric film polarization, in resistive switching memories
the electrical pulse at its terminals. Among the just mentioned technologies,
the latter is the most suitable to be implemented in emerging neuromorphic
architectures like the crosspoint array, i.e. a matrix of memristive memory
elements, whose rows and columns respectively define the memory bitlines and
wordlines. Resistive switching device, both the PCM-RAM and the RRAM, can
be easily integrated in this kind of architectures as they are are two-terminals,
and, in particular, the RRAM is the easiest memristor to be processed, as its
structure just consists in an oxide layer sandwiched between an inert metal
and an oxidable metal. The crosspoint architecture can be designed both for storage applications,
thus targeting high integration density memory layouts to perform as a memory
technology with intermediate properties in terms of speeds and data density
between high-storage memory solutions, like HDDs or flash memories, and high-speed but expensive semicondutor memories, like DRAM, defining a new
class of memory systems called SCM; or for computing applications. Indeed,
thanks to the nature of resistive switching devices, whose resistance can be
programmed in analog states just by varying the electrical pulse applied to
their electrodes, if such devices are arranged in matrix-shaped arrays, it will
be possible to execute some calculation algorithms in just two steps, whereas
in conventional Von Neumann architectures, where memory and arithmetic
units are physically separated, they would have costed much more iteration
steps. The first one consists in a suitable programming of each array element
in the targeted resistive value, and the second, in the application of a read
voltage to the array cells. The current collected at the end of every bitline
turns out to be the product between the read voltage vector values and the
array conductance map: in this way, we are leaving nature itself to provide
calculations, as a consequence of the Ohm’s law. The experimental activity was focused on the fabrication and on the electrical
characterization of emerging resistive switching memories to optimize the
RRAM electrical properties towards an efficient implementation in a crosspoint
array. The manufacturing and topographic characterization processes were
carried out in the PoliFab clean-room (the micro and nano-technology center of
the Politecnico di Milano), while the electrical measurements were performed
at the Electron Devices Lab of the Dipartimento di Elettronica, Informazione e
Bioingegneria. Chapter 1 introduces the recent issues related to microelectronics scaling,
specifying why this situation became a breakpoint for the development of
alternative computing schemes with respect to conventional Von Neumann
architectures. Before discussing about the most mature and the most promising SCM technologies solutions for targeting intermediate specifics in terms of
speed and costs between HDDs or SSDs and DRAMs, the traditional memory
solutions defining the today’s memory hierarchy will be presented together
with the description of their basic principles. Chapter 2 describes the theoretical principles about the processing techniques
employed during the fabrication activity carried out in the PoliFab
clean-room, including the cleaning processes, EBL and optical lithography,
RIE and chemical etching, and thin-films deposition solutions like CVD and
evaporation.
Chapter 3 presents the characterization tools employed for the evaluation of
the fabrication quality, in terms of physical characterization, meaning a topographic
study of the patterned arrays of devices, and electrical characterization.
The setup installed to run the electrical measurements will be presented too in
this chapter.
In chapter 4 starts the report of the thesis activity, specifying the recipes
adopted for the fabrication of the RRAM devices patterned with EBL and
the results obtained from the quasi-static electrical characterization. A first
crossbar measurement will be presented and will reveal that a deeper research
should be undertaken to fastly improve the devices performances towards an
effective crosspoint implementation.
Chapter 5 describes the manufacturing processes for RRAM patterned with
optical lithography and introduces the great time advancements allowed by
this lithography technique. To evaluate the electrical quality of these devices, a
deeper quasi-static electrical characterization, including more detailed statistical
analysis, was carried out. Both the technologies based on a HfOx and a SiOx
switching layer were studied. Chapter 6 presents the volatile RRAM devices based on a Ag top-electrode,
fabricated in PoliFab departing from the processes and the array structures already developed for the fabrication of NV-RRAM devices. The results form
the quasi-static and the pulsed characterizations performed on these currentself-
limiting HfOx-based selectors were so interesting, that a dedicated paper
was written and accepted at the 2019 ICECS conference.
Chapter 7 indicates the solutions conceived to improve the electrical properties
of NV-RRAM devices towards an effective crosspoint implementation,
focusing on the main issues of current overshoot after electroforming and high
forming voltages. In this chapter are reported the results of the annealing
treatments carried out to reduce the forming voltages and then the electrical
characterizations performed on RRAM devices fabricated with a C line, which
was engineered to overcome the difficulties in resetting cells after electroforming.
Chapter 8 is a report of the electrical measurements performed on an 8x8
crossbar array, in which I indicated too the strategies developed to reach the
ultimate purpose of my thesis activity, which consists in the development of a
crossbar array of resistive-switching elements randomly programmable in the
desired analog resistive states. This result was demonstrated in a restricted
crosspoint made of 4x4 devices
Nel giro del sangue, su Tommaso Ottonieri
No full textIl saggio analizza l'opera di Tommaso Ottonieri "Le strade che portano al Fùcino" mettendone in evidenza l'infrazione dei generi letterari, l'originalità dei nuclei poetici e di quelli narrativi, oltre alla stratificazione plurilinguistica che è cifra stilistica e allegorica dell'opera. Il saggio, inoltre, individua i rapporti di quest'opera di Ottonieri con le immagini visive e con la cultura musicale coeva e rintraccia anche rapporti intertestuali con precedenti opere poetiche, narrative e saggistiche dell'autore.The essay analyzes the work of Tommaso Ottonieri "Le strade che portano al Fùcino" highlighting the infringement of literary genres, the originality of poetic and narrative nuclei, in addition to the plurilingual stratification which is the stylistic and allegorical cipher of the Opera. Furthermore, the essay identifies the relationships of this work by Ottonieri with visual images and contemporary musical culture and also traces intertextual relationships with previous poetic works, narratives and essays by the author
L’olio in un aneddoto dalle biografie di Cesare (Plut. Caes. 17, 9–10 e Suet. Iul. 53)
No full textThe paper discusses a debated anecdote on Julius Caesar that Plutarch and Suetonius probably derived from the writer and politician Gaius Oppius. By examining some hitherto neglected Latin agronomic and gastronomic sources, the author provides some examples of olea condita used for culinary purposes and reassesses the interpretation of the anecdote
APROXIMACIÓN AL ESTUDIO DE LA LENGUA POÉTICA DE TOMMASO STIGLIANI
No full textEn este estudio hemos querido trazar un primer acercamiento al estudio de la lengua poética del autor italiano Tommaso Stigliani y lo hemos hecho analizando su mayor obra, Il Mondo Nuovo (1628), poema épico sobre el descubrimiento de América. Hemos analizado los principales fenómenos de la ortografía, la fonética, la morfología, la sintaxis y el léxico que aparecen en el texto, teniendo en cuenta la norma lingüística del siglo XVII y las opiniones y teorías de los principales lingüistas de la época. Con este estudio, por tanto, ofrecemos un acercamiento a la concepción de lengua poética que Tommaso Stigliani, a través de su poema y de sus escritos teóricos, puso de manifiesto a lo largo de su carrera literaria.
Abstract
In this study, we have attempted to make inroads into the study of the poetic language of the Italian author, Tommaso Stigliani, analysing his principal work, Il Mondo Nuovo (1628), an epic poem on the discovery of America. We have examined the main phenomena of orthography, phonetics, morphology, syntax and lexicon that appear in the text, taking into account the linguistic norms of the 17th century, as well as the opinions and theories of the major linguistic experts of that period. Our study, therefore, gives an insight into Tommaso Stigliani’s conception of poetic language, which, through his poem and his theoretical writings, he maintained for the duration of his literary career
- …
