In recent years, increasing implementation of biological, cultural, and mechanical weed-control methods is desired; however, many of these techniques are not viable in established turfgrass systems. The use of freezing or frost for weed control has previously been researched; however, is not well elucidated. Field and greenhouse experiments were conducted to evaluate liquid carbon dioxide (LCD) for weed control in established turfgrass systems. LCD was applied with handheld prototypes that were modified to reduce the amount of LCD required for weed control. Common annual and perennial turfgrass weeds included common chickweed, corn speedwell, goosegrass, large crabgrass, smooth crabgrass, Virginia buttonweed, and white clover. Turfgrass tolerance was evaluated on the following species: hybrid bermudagrass, Kentucky bluegrass, tall fescue, and zoysiagrass. The final modification allowed for lower output (0.5 kg LCD min−1) when compared with the initial prototype (3 kg LCD min−1). In general, weed control increased as LCD increased. When comparing weed species life cycles, annuals were controlled more than perennials (P < 0.0001) at 14 and 28 d after treatment (DAT). Further, exposure time affected control as white clover, Virginia buttonweed, and large crabgrass control was greater (18, 14, 15%, respectively) from the longer exposure time (30 vs. 15 s), although equivalent amounts of LCD (30 kg m−2) were applied. These data also suggest that plant maturity affects control, as large crabgrass control in one- to two- and three- to four-leaf stages (> 90%) was greater than in the one- to two-tiller stage (< 70%). Turfgrass injury at 7 DAT was unacceptable (> 30%) on all species, but declined to 0% by 28 DAT. These data suggest that LCD has the potential to provide an alternative for weed control of select species where synthetic herbicides are not allowed or desired.