We describe and analyze in detail the shapes of Fe islands encapsulated under the top graphene layers in graphite. Shapes are interrogated using scanning tunneling microscopy. The main outputs of the shape analysis are the slope of the graphene membrane around the perimeter of the island, and the aspect ratio of the central metal cluster. Modeling primarily uses a continuum elasticity (CE) model. As input to the CE model, we use density functional theory to calculate the surface energy of Fe, and the adhesion energies between Fe and graphene or graphite. We use the shaft-loaded blister test (SLBT) model to provide independent stretching and bending strain energies in the graphene membrane. We also introduce a model for the elastic strain in which stretching and bending are treated simultaneously. Measured side slopes agree very well with the CE model, both qualitatively and quantitatively. The fit is optimal for a graphene membrane consisting of 2-3 graphene monolayers, in agreement with experiment. Analysis of contributions to total energy shows that the side slope depends only on the properties of graphene/graphite. This reflects delamination of the graphene membrane from the underlying graphite, caused by upward pressure from the growing metal cluster. This insight leads us to evaluate the delamination geometry in the context of two related, classic models that give analytic results for the slope of a delaminated membrane. One of these, the point-loaded circular blister test model, reasonably predicts the delamination geometry at the edge of an Fe island. The aspect ratio also agrees well with the CE model in the limit of large island size, but not for small islands. Previously, we had speculated that this discrepancy was due to lack of coupling between bending and stretching in the SLBT model, but the new modeling shows that this explanation is not viable. metal. This is challenging, since 2D materials have intrinsically low surface energies, so metals tend to grow on top of them as 3D clusters [7]. It is therefore attractive to consider synthesis strategies wherein metal morphologies are kinetically-limited [8], or the metal morphology is constrained (and stabilized) by intercalation, to promote the 2D morphology.Elsewhere, we have reported that metal nanoclusters can be synthesized at the surface of the 3D van der Waals material, graphite, in an intercalated form if two conditions are met [9][10][11][12]. First, the graphite surface must be ion bombarded to introduce defects that can act as entry portals for the metal atoms. Second, the graphite must be held at relatively high temperature while the metal is deposited, so that portals do not become blocked by growing metal clusters. Under these two conditions, we observe stable metal nanoclusters that are encapsulated beneath the graphite surface. Depending on the metal, they are a few atomic layers to hundreds of atomic layers tall, and about ten to hundreds of nm wide. We have reported the growth conditions and characteristics of such clusters in detail fo...