We use cosmic microwave background (CMB) temperature maps from the 500 deg2 SPTpol survey to measure the stacked lensing convergence of galaxy clusters from the Dark Energy Survey (DES) Year-3 redMaPPer (RM) cluster catalog. The lensing signal is extracted through a modified quadratic estimator designed to be unbiased by the thermal Sunyaev–Zel’dovich (tSZ) effect. The modified estimator uses a tSZ-free map, constructed from the SPTpol 95 and 150 GHz data sets, to estimate the background CMB gradient. For lensing reconstruction, we employ two versions of the RM catalog: a flux-limited sample containing 4003 clusters and a volume-limited sample with 1741 clusters. We detect lensing at a significance of 8.7σ(6.7σ) with the flux (volume)–limited sample. By modeling the reconstructed convergence using the Navarro–Frenk–White profile, we find the average lensing masses to be {M}200{{m}}=({1.62}-0.25 0.32 [{stat}.] /- 0.04 [{sys}.]) and ({1.28}-0.18 0.14 [{stat}.] /- 0.03 [{sys}.]) × {10}14 {M}ȯ for the volume- and flux-limited samples, respectively. The systematic error budget is much smaller than the statistical uncertainty and is dominated by the uncertainties in the RM cluster centroids. We use the volume-limited sample to calibrate the normalization of the mass-richness scaling relation, and find a result consistent with the galaxy weak-lensing measurements from DES.
