Environment may also affect outer disk structure. Younger et al (2007) showed that prograde minor mergers can drive mass inward and outward, creating a Type III profile. Borlaff et al (2014) also suggested that Type III S0 galaxies can result from a merger. This is consistent with observations in Erwin et al (2012) that S0 galaxies in Virgo have proportionally more Types I and III, suggesting that interactions or mergers have been important. Erwin et al. also found that bars have little effect on the proportion of exponential types. Athanassoula et al (2016) simulated a gas-rich major merger and showed that it formed an exponential disk in the final system.
On the other hand, Maltby et al (2012) measured the V-band radial profiles of 330 galaxies observed with Hubble Space Telescope over a half-degree field surrounding a galaxy supercluster at redshift 0.165. They found no dependence on environment, cluster versus field, for the ratio of the outer to the inner disk scale length or the outer scale length itself. Head et al (2015) got a similar result looking at S0 galaxies in the Coma cluster; using a profile decomposition algorithm to remove the bulge, they found that bars are important for disk structure, correlating with Types II and III (contrast this with the Erwin et al (2012) result above), but that location in the cluster is not important. According to Head et al (2015), the relative proportion of Types I, II, and III is the same in the core, at intermediate radii, and in the outskirts of Coma (in fact, most of the S0 galaxies were Type I).
Some of the appearance of Type III could be from a bright halo or extended bulge and not from stars in the disk (Erwin et al 2005). Maltby et al (2015) suggested that half of the S0 Type III structures in various environments come from extended bulge light, although this fraction is only 15% in later Hubble type spirals. This implies that disk fading can make an S0 from a spiral, preserving the scale length. Simulations by Cooper et al (2013) also found that the outer stellar structure can be in a halo and not a disk, as a result of mergers.
Bars and spirals seem to be important in determining the break radius for down-bending (Type II) exponentials. Muñoz-Mateos et al (2013) suggested that the break radius for Type II's is either at the outer Lindblad resonance (OLR) of a bar or the OLR of a spiral that is outside of a bar. The spiral and bar are assumed to have their pattern speeds in a resonance with the inner 4:1 resonance of the spiral at the corotation radius of the bar (see Pohlen and Trujillo 2006 and Erwin et al 2008). Laine et al (2014) also found that bars and spirals are important: 94% of Type II breaks are associated with some type of feature; 48% are in early type galaxies with an outer ring or pseudoring; 8% are with a lens, assumed to be the OLR of a bar, and if there is no outer ring, then the breaks are at 2 times the radius of an inner ring (this being the ratio of radii for outer and inner ring resonances); 14% are in late type galaxies associated with an end to strong star formation, and 24% are at the radius where the spiral arms end. For Type III breaks studied by Laine et al (2014), 30% are associated with inner or outer lenses or outer rings.