Area Scienze chimiche, fisiche, matematichehttp://elea.unisa.it/xmlui/handle/10556/602024-09-14T13:54:15Z2024-09-14T13:54:15ZMolecular Properties Via Induced Current DensitiesSumma, Francesco F.http://elea.unisa.it/xmlui/handle/10556/73322024-09-10T09:15:42Z2024-02-22T00:00:00ZMolecular Properties Via Induced Current Densities
Summa, Francesco F.
The PhD thesis written by Francesco F. Summa on Molecular Properties via
Induced Current Densities is a careful report of a highly qualified and innovative research. Sixteen
peer-reviewed publications on renowned journals in the chemical physics and theoretical chemistry
areas have originated from the work reported in the thesis, besides other six publications, also co-
authored by F. F. Summa on different subjects. The number, excellent quality and originality of the
16 papers mentioned above is far beyond any usual standard for a PhD student, even when the
worldwide reckoned scientific excellence of the research group(s) within which the work has been
performed is properly taken into account.
The work done in the thesis consists of three well identified parts: theory development, SW
implementation of the developed theory into the freely available SYSMOIC package and test
applications of the newly developed theory and SW on a good number of carefully selected cases.
All three mentioned steps of the PhD work are reported in the thesis with great clarity, perfect
progression and concatenation. The writing is terse and quite accurate. After a brief and clear
introduction, the general quantum mechanical framework for the treatment of a molecular system
interacting with electric and magnetic time dependent fields is elegantly recapped and exposed.
Chapter 3 introduces the equations defining the total current density induced by a time-independent
magnetic field and illustrates in detail the procedures aimed at solving the problem of the origin
dependence of the computed current density vector field. Chapter 4 deals with time-dependent
perturbations and how they do result into time-dependent charge and current densities. Chapter 5
reports a number of very interesting applications, both in the static and in the dynamic regimes, of
the theoretical and SW developments illustrated in the Chapter 3 and 4. The work performed in the
thesis and the writing thereof both show a very high command of symmetry analysis of systems in a
magnetic field and of the topological analysis and classification of current densities, two subjects
which are not at all trivial. The provided list of acronyms, in the initial pages of the thesis, is highly
appreciated. Overall, this thesis provides an extremely recommendable reading for those who are
interested or want to be acquainted with the most recent progresses on the subject of Molecular
Properties via Induced Current Densities. [Report by Carlo Gatti on the PhD thesis of Francesco F. Summa]
2021 - 2022
2024-02-22T00:00:00ZHidden Structures and Conflicting Items in Combinatorial Optimization ProblemsSerra, Domenicohttp://elea.unisa.it/xmlui/handle/10556/73302024-09-10T09:03:11Z2023-02-14T00:00:00ZHidden Structures and Conflicting Items in Combinatorial Optimization Problems
Serra, Domenico
Combinatorial optimization-based methods are widely used to solve complex real life problems.
In this thesis, we use some of these methods for addressing several emerging combinatorial opti-
mization problems on graphs that can be classified in two macro areas: i) graph substructures
identification problems; ii) combinatorial optimization problems with conflict constraints.
In graph theory, graphs are defined as mathematical structures that describe entities of inter-
est (as nodes) and their relationships (as edges). Many real-world problems can be described
through the use of a graph. For example, graph theory finds application in: i) the context of
social network analysis, where graphs are used to represent the interactions (edges) between users
(nodes); and in ii) the study of biological networks, such as protein-protein interaction, where
the nodes are proteins, and the edges represent their physical interactions. In graph analysis one
common application is the identification of clusters (or communities) of nodes that are tightly
connected. In social networks, a community could represent a set of users sharing the same
interest, while in the protein-protein interaction networks it could represent a set of very similar
proteins forming a protein complex.
In this thesis, three problems related to identifying graph substructures have been tackled.
The first problem addresses the 2-Edge-Connected Minimum Branch Vertices, that finds ap-
plication in the design of optical networks. A graph is 2-Edge-Connected if by removing one
edge, the graph is still connected. The problem looks for a spanning 2-edge connected subgraph
having the minimum number of branch vertices that is vertices with degree strictly greater than
two. In these networks, branch vertices are associated with switch devices that split the light
signals and send them to the adjacent vertices. For this NP-complete problem we developed a
genetic algorithm using ad-hoc designed operators.
The second addressed problem arise in the social network analysis and aims to study how
users influence the choices of their neighbours. In particular, we addressed the Collapsed k-Core
Problem that seeks to identify a subset of critical users in the network whose choices would
alter the cohesiveness of a community. To the best of our knowledge, this is the first attempt
to formulate this problem using mathematical programming. We implemented multiple solution
approaches and compared them on a set of benchmark instances.
The last case studied is related to network clustering, where a cluster graph is a disjoint union
of cliques. The Cluster Deletion problem is defined as the identification of the minimum number
of edges to remove from a network to produce a cluster graph. This is a well-known NP-hard
problem and we faced it using integer linear programming formulation and a heuristic approach
based on edge contraction operation. Our results show the effectiveness of our methodology both
on artificial and real-world biological networks.
4
Currently, the definition of many real-life problems, doesn’t always fully capture their com-
plexity. Indeed, classical optimization problems encountered over the years, although extensively
studied, do not always take into account additional limitations, such as incompatibility situa-
tions, encountered in real-world problems. In this context, two or more elements of the problem
cannot be chosen together to compose a feasible solution. Such incompatibilities are modelled
by introducing conflict constraints in classical combinatorial optimization problems, leading to
more realistic but often harder problems.
Finally, three different combinatorial optimization problems with conflicts constraints are
addressed.
The first one is a variant of the set cover problem where pairwise conflicts are added among
the subsets. In the formulation of this problem, two sets in conflict can belong to the same
solution, provided that a non-negative penalty is paid. We introduced two mathematical formu-
lations for the problem and offered a parallel Greedy Randomised Adaptive Search Procedure
for its solution. The performance of our algorithm was evaluated through an extensive set of
experiments. The results shown the effectiveness and efficiency of our methodology compared to
the mathematical model solutions.
The second problem is related to the Maximum Flow Problem with Conflict constraints on
the edges, for which we present a matheuristic method based on the combination of two different
approaches: Carousel Greedy and Kernel Search. The results shown that the Carousel Greedy
selection substantially improves the effectiveness of the Kernel Search.
The last problem is the Minimum Spanning Tree with Conflicts, that we solved by using a
Kernel Search method. Also in this case, the results on benchmark instances shown that our
methods identifies a better solution compared with existing methods. [edited by Author]
2021 - 2022
2023-02-14T00:00:00ZSustainable approaches for the synthesis and the chemical degradation of polyestersSantulli, Federicahttp://elea.unisa.it/xmlui/handle/10556/73292024-09-10T08:58:05Z2023-03-03T00:00:00ZSustainable approaches for the synthesis and the chemical degradation of polyesters
Santulli, Federica
Petrochemical-based plastics have a significant impact on modern society, as underlined by their high
production and countless applications. Indeed, the production of plastics has increased over the past 50 years,
reaching 390.7 million tonnes in 2021.1 Although these materials have been designed to last over time, every
year around 40% of the world's plastic production is used in the short-term packaging field and about 50 million
tons of plastic waste are retained in landfills or spread into the environment.2
Severe consequences due to the accumulation of plastic waste have encouraged current scientific research to
develop new strategies to address these problems. Some examples are the chemical recycling of traditional
plastics, which would reduce accumulation and help reuse the raw material,3
but also the introduction of new
bio-based polymers, for which an effective waste management strategy is defined beforehand. Both of these
approaches are consistent with a circular economy vision, in which the material retains its value even after its
use.4
In this context, polylactide (PLA) is considered the most promising material because it combines good
mechanical properties, biodegradability and biocompatibility. Although these characteristics design it as a
green material, there are some weaknesses in terms of sustainability regarding the production and management
of waste. The catalyst used industrially to produce PLA is tin(II) bis(2-ethylhexanoate) [Sn(Oct)2], which is
classified as a toxic compound, and the waste management strategy of PLA is still aligned with an economic
model linear. A promising end-of-life path for PLA is its chemical recycling through alcoholysis in which the
product, alkyl lactate, can be converted into lactide, thus realizing a circular economy process, or used as a
green solvent in industries.5
The aim of this PhD thesis was the development of new catalytic systems able of synthesizing PLA using non-
toxic and robust metals enough to withstand even industrially relevant reaction conditions. Specifically, new
heteroleptic and homoleptic complexes of zinc and magnesium supported by pyridyl-phenoxy-imine ligands
were synthesized.
The family of Zn heteroleptic complexes showed the highest activities and a good control of the polymerization
process, ranking among the most efficient catalysts reported in the literature and comparable to the industrial
tin catalyst.
Recent literature studies have shown that bimetallic complexes can show a peculiar catalytic behavior
compared to monometallic analogues, thanks to cooperation effects that can be established between the two
metal centers. The introduction of a binaphthol backbone between the two pyridyl-phenoxy-imine moieties
allowed to obtain hexacoordinate ligands, which promoted the synthesis of bimetallic complexes. These
catalytic systems, in addition to being highly active in the lactide polymerization, showed a good control of
the polymerization obtaining linear polymers with expected end groups. When the reaction is carried out at
high temperature, cyclic PLA is selectively obtained.
The same complexes were also used in the degradation reaction of PLA, by alcoholysis. The Zn heteroleptic
complexes showed the highest activities and from mechanistic studies two different pathways were highlighted
according to the degradation conditions. In THF solution, the degradation occurred through a two-step process
in which the random scission of the polymer chains leads to the formation of oligomers, which progressively
convert into alkyl lactate. Under solvent-free conditions, the degradation occurred through a progressive
erosion of the chain-end with the direct formation of alkyl lactate.
Subsequently simple and commercially available compounds, such as amides of various non-toxic metals,
have been explored in the degradation reactions of PLA and polyethylene terephthalate (PET), one of the most
commercially widespread polymers. This work has allowed us to highlight the potential of different metals in
these still underdeveloped reactions. For example, Zr amide proved to be the best catalyst in PET glycolysis
reaching activities comparable to the best reported in the literature for metal catalysts (78% BHET yield within
one hour at 180°C).
Kol’s research group of Tel Aviv University, where I carried out a visiting research period, had synthesized a
new hexacoordinate complex of Zr supported by aminotris(phenolate) ligand, which showed the highest
activity ever recorded for metal complexes in lactide polymerization under industrial conditions.6
In this
context, my work has focused on the synthesis of new Zr complexes with different substituents in the phenoxy
moieties and the study of their behavior in the stereoselective polymerization of PLA. [edited by Author]
2021 - 2022
2023-03-03T00:00:00ZAdvanced Modeling of Microlensing EventsRota, Paolohttp://elea.unisa.it/xmlui/handle/10556/73282024-09-10T08:52:26Z2023-04-17T00:00:00ZAdvanced Modeling of Microlensing Events
Rota, Paolo
The search for exoplanets is one of the most exciting challenges. From the
first exoplanets discovered around the pulsar PSR B1 257+12 back in 1992 until
today there are over 5200 exoplanets discovered. And the number will continue
to rise in the coming years with the advent of the latest generation telescopes.
There are various techniques for finding exoplanets such as transit, radial
velocity, pulsar timing, direct imaging etc. But among these, gravitational
microlensing is one of the most fascinating. Gravitational microlensing is
a particular technique to detect exoplanets otherwise unavailable with other
techniques such as transits or radial velocity. We have a microlensing event
when the light from a distant source is deflected by a lens passing through
the source and the observer. The result, in the simplest case, is a bell-shaped
peak in the light curve. Planets can be detected studying the anomalies in the
lightcurve (additional peaks or dips, longer distortions, etc.). But microlensing
is not important only for exoplanets. The fact that with this technique it is
possible to reach distances of the order of the galactic center allows us to
discover objects ranging from the disk to the bulge allowing us to have a much
more complete overview of the study of the stellar populations of our galaxy.
And in particular with the study of binary systems we can detect faint objects
that are impossible to reveal with other techniques, such as brown dwarfs,
mysterious objects of which little is known yet and which are presumed to
populate our galaxy in great abundance. Microlensing is undoubtedly the
best method to discover these objects, since having extremely low luminosities
they are difficult to see with telescopes except when they play the role of
lens. Moreover, in some cases the orbital motions of the system can also be
detected, allowing us to study their dynamics in more detail. The typology of
microlensing events is enormous and in this thesis we will analyze some of the
cases that can occur. In the first part we focus on the fundamental concepts of
microlensing theory for the exoplanets search. The second part is dedicated to
the modeling of microlensing events, which is based on the VBBinaryLensing
code. In the third part we analyze some microlensing events, each with certain
characteristics, to highlight how vast the case history of these events is. Finally
there will be the conclusions where the results achieved and future prospects
are discussed. [edited by Author]
2021 - 2022
2023-04-17T00:00:00Z