Abstract Particle number, size, and mass concentration; effective density; and mass-mobility coefficient are measured for a gasoline direct injection (GDI) engine fueled with E10, E30, and E50 ethanol blends and operated in three combustion modes: Stoichiometric; Lean Homogeneous; and Lean Stratified. The focus of this study is on solid particles, so all measurements are made behind a catalytic stripper. Two configurations for measuring effective density are compared: DMA-CPMA-CPC and CPMA-SMPS. Our measurements show no difference between densities measured with DMA-CPMA-CPC and CPMA-SMPS configurations within experimental uncertainty. Both combustion mode and ethanol concentration strongly influence particle emissions. Using E10: Stoichiometric operation produces the lowest particle number concentration and highest effective density; Lean Homogeneous gives the lowest particle volume and mass, smallest particles, and lowest effective density; Lean Stratified produces the highest particle concentrations and effective densities. Increasing ethanol content under Stoichiometric and Lean Stratified conditions decreases number and mass concentrations with only small changes in density; but under Lean Homogeneous conditions increases mass and number concentrations and significantly increases density. Using the measured effective densities and particle size distributions measured with a TSI EEPS, solid particle mass is calculated using the integrated particle size distribution (IPSD) method and compared to black carbon mass (BC) measured with a Micro Soot Sensor (MSS). In most cases, the IPSD method gives higher mass concentrations than the MSS BC concentrations. This is consistent with the presence of non-light absorbing material – likely ash and tightly bound hydrocarbons – not detected by the MSS. However, errors in measurement of the upper end of the EEPS size distribution, where most of the mass is found, may play a role.