Olivier Gagné was born in Gatineau, QC, in 1988. He graduated from High School in Gatineau, QC in 2005, winning Canada's Governor General's Bronze Academic Medal for highest graduation GPA. He graduated with a B.Sc. Honours in Chemistry from the University of Ottawa in 2009 with high honours.
Olivier moved to Winnipeg to start a M.Sc. in Geological Sciences in September 2009 with Frank Hawthorne and transferred to the Ph.D. program a year later (September 2010). He holds an NSERC PGS D3. He has significant experience in University Governance (e.g. three-term senator, one-term Senate Executive member) and leadership roles (e.g. two-term Vice-President of the University tennis club, three-term board member of the University of Manitoba Graduate Student Association). Olivier is a very active sky diver, scuba diver and tennis player.
The general failure to make reliable crystal structure predictions from chemical composition has been identified as a continuing scandal in Crystallography. While this highlights the complexity of the problem, it also reveals our limited understanding of the underlying principle affecting crystal-structure stability. To this effect, he has developed a 4-step approach that showcases the usefulness of the bond-length to bond-strength correlation of Bond-Valence Theory in addressing mineralogical and crystallographic concepts revolving around the stability of crystal structures.
In step 1, crystal structures are represented as graphs using Graph Theory. A set of ideal bond-valences (the weights of the graph) is calculated by solving the valence-sum and loop equations of the particular bonding pattern and vertex charges. Step 2 then deals with the assignment of atoms to the vertices of the graph. An atom may be assigned to a vertex if it encompasses the respective ideal bond-valences within its pre-determined bond-valence range (this range is determined from a thorough survey of the International Crystal Structure Database). Following the assignment of atoms to all vertices, the ideal bond-valences are turned into ideal bond-lengths using the bond-length to bond-strength correlation, to obtain a network of ideal inter-atomic distances. In step 3, a distance-least-squares refinement yields a structure in Cartesian space, while step 4 evaluates the structure in terms of energy.
The key concept of the approach lays in the calculated set of ideal bond-valences. Subtle variations in the parameters (e.g. connectivity) of a crystal structure yield different sets of ideal bond-valences, some more stable than others. By varying the appropriate parameters, solid solution, symmetry, order/disorder, etc., can be studied, while the inexistence of minerals or plausibility of yet-to-be-observed ones can be predicted.