Supplementary Figure 1: Reduction of Peak Broadening in the AGR Sample Compared to The

Supplementary Figure 1: Reduction of peak broadening in the AGR sample compared to the PP sample by XRD

Supplementary Figure 2: Sample gel electrophoresis of PP in D5W and water and AGL PP in saline.

Supplementary Figure 3: TGA measurements of PP and AGR in water and AGR in D5W

Supplementary Information

Testing Reversibility of AGL PP and AGR Formation

Based upon the characterization data presented in this paper for the AGL PP, the individual PP did not appear to undergo morphological changes, yet the AGL PP were still functioning as the larger moieties in several tests. The ease of AGL reversibility partially depends on the type of NP as well as the environment to which that the particle is exposed. Some NP AGL may easily revert back to the individual PP simply by agitating the sample, while other NP are more strongly associated, although not irreversibly fused together. Therefore, the goal behind this portion of the study was to determine if any method would revert the AGL PP back to individual PP, thus indicating that the structures were formed by inter-particle interactions other than covalent interactions. A variety of different solvents and conditions (including heating and sonicating) were tested for the ability to revert the AGL PP back to PP. Several methods indicated that the PP could be recovered, but vortexing the AGL PP in DMSO produced the most dramatic SPR color change. A 20 µL sample of 0.85 g/L AGL PP in saline or 0.85 g/L AGR in autoclaved D5W was diluted in 200 µL of DMSO. The new solutions were measured by DLS, TEM, and analytical ultracentrifugation following the procedures described in the paper. Each procedure was repeated three times.

Upon vortexing the AGL PP in DMSO, the purple solution became pink, and DLS measurements listed the majority of the particles as ~ 10 nm by number weighting. In addition, the majority of the un-agglomerated particles did not move by density ultracentrifugation whereas the agglomerated particles traveled the entire length of the centrifuge tube. However, the particles rested in the 15% glycerol layer instead of the 5% layer of the PP in water most likely because the DMSO was denser than the 5% glycerol. This movement occurred prior to centrifugation. In addition, TEM showed that after being suspended in DMSO, the majority of the AGL PP did revert back to PP. However, some of the individual PP did have a distorted shape, and some particles were either smaller or larger than the typical 5 – 8 nm PP. This result is possibly due to the gold NP interacting with the DMSO, ripening in the solution, or beginning to form AGR as the saline solution aged. In contrast, the AGR sample did not show a color change when DMSO was added. The DLS data had a peak ~200 nm by number weighting, and multiple peaks measuring 100 – 6000 nm by intensity weighting, similar to the AGR in D5W. Prior to the centrifugation, the AGR moved to the 15% glycerol layer due to the density of DMSO. Following centrifugation, the AGR sample did not move, which would be expected since the 15% glycerol layer is where the AGR in water are normally found. TEM showed that the AGR individual NP did not regain a spherical shape when diluted in DMSO. In addition, the UV-Vis spectrum had changes that showed the AGL PP were reverting to PP. There was a significant broadening of the maximum peak so that it included the PP SPR peak of 520 nm. The maximum absorbance shifted from 600 – 744 nm to 550 – 580 nm. These data imply that for the AGL PP sample, the individual particles interact in a reversible manner when they self associate as opposed to the AGR, where the individual particles are irreversibly fused.