To compare the efficacy of two surgical techniques used to remove silicone oil (SiO) emulsion tamponade after pars plana vitrectomy: triple air-fluid exchange (AFX) and balanced salt solution lavage (BSSL). X-ray photoemission spectroscopy measured silicon content of the dry residue of fluid samples taken during AFX and BSSL. Ten patients underwent AFX and five BSSL. Three fluid samples were taken per patient, and the dry residue of 10 drops per sample were analyzed. A fluid sample from a patient who never received SiO tamponade was also analyzed to set a "blank" reference sample. Patients' demographics showed no significant difference. Sample 1 of the two groups contained comparable silicon content while samples 2 and 3 of the AFX group contained significantly more silicon than that of the BSSL group (15.0 ± 0.1 and 12.0 ± 0.9 for the AFX group vs. 10.7 ± 1.4 and 5.2 ± 0.6 for the BSSL group, respectively; P < 0.05). The cumulative amount of silicon in the three successive samples was also significantly higher for the AFX group (42.3 ± 1.6 vs. 32 ± 2; P < 0.0001). The average silicon content ratio of consecutive samples was significantly higher for the AFX group compared to the BSSL group (0.90 ± 0.01 vs. 0.58 ± 0.06; P = 0.006). Triple AFX removed more silicon than triple lavage. The eye wall actively interacts with silicon emulsion retaining silicon content rather than behaving as a neutral container. Triple air-fluid exchange removed more silicon than BSS lavage. Neither technique behaved as a well-mixed box dilution, suggesting the eye walls actively retain emulsion and a dynamic equilibrium is established between silicon dispersion and the eye wall surface.