Experimental Determination of the Fate of Rising CO 2 Droplets in Seawater PETER G. BREWER,* EDWARD T. PELTZER, GERNOT FRIEDERICH, AND GREGOR REHDER † Monterey Bay Aquarium Research Institute, 7700 Sandholdt Road, Moss Landing, California 95039 Direct oceanic disposal of fossil fuel CO 2 is being considered as a possible means to moderate the growth rate of CO 2 in the atmosphere. We have measured the rise rate and dissolution rate of freely released CO 2 droplets in the open ocean to provide fundamental data for carbon sequestration options. A small amount of liquid CO 2 was released at 800 m, at 4.4 °C, and the rising droplet stream was imaged with a HDTV camera carried on a remotely operated vehicle. The initial rise rate for 0.9-cm diameter droplets was 10 cm/s at 800 m, and the dissolution rate was 3.0 μmol cm -2 s -1 . While visual contact was maintained for 1 h and over a 400 m ascent, 90% of the mass loss occurred within 30 min over a 200 m ascent above the release point. Images of droplets crossing the liquid-gas- phase boundary showed formation of a gas head, pinching off of a liquid tail, and rapid gas bubble separation and dissolution. Introduction The disposal of fossil fuel CO2 in the deep ocean has been discussed (1, 2) as a means of ameliorating greenhouse gas induced climate change (3, 4).Although modelsofthisprocess have been formulated (5, 6),and laboratorysimulationshave been carried out (7-9), there have been few direct oceanic experiments reported (4, 10, 13). With the availability of advanced Remotely Operated Vehicle (ROV) technology it has now become possible to carry out controlled releases of many chemical species in the deep sea and to observe and measure the processes taking place. In earlier work (10) we have reported on the controlled mid-water release of both CH4 and CO2 into contained spaces to form hydrates trapped at depth for imaging. Here we report on the more difficult problem ofobservingand accuratelymeasuringthe behavior of a rising stream of freely released CO2 in order to evaluate the behavior and dissolution rate of droplets in dynamic motion in the open ocean. An accurate description of the fate of CO2 injected into ocean water is necessaryfor predictingthe behavior oflarge- scale ocean disposalschemes.Atshallowdepths(above about 350m,dependingupon the localtemperature gradient)CO2 is in the gas phase and will readily dissolve. Below this depth the gas hydrate phase boundaryoccurs,again dependent on local P,T conditions, and a shell of CO2-hydrate may be generated (11, 12) which can have profound consequences; and below about 400 m depth the gas -liquid transition occurs.Each ofthese regimes can have distinct reaction rates and characteristics.The lowcompressibilityofseawater,and the high compressibility of liquid CO2, results in a density ratio reversal at high pressure such that below about 3000 m depth a gravitationally stable release can be achieved (13) under typical ocean conditions. Release at depth leads to longoceanicresidence timesand effective sequestration from theatmosphere(14).However,although release at great depth leads to long residence times, the cost and difficulty of this suggeststhe need for carefulevaluation ofmid-depth releases (6), which may be very effective near deep water mass formation areas. Laboratory studies have been successful in identifying both the CO2 hydrate phase boundaryand the characteristics of the solid hydrate itself (15). It is difficult for a laboratory study to investigate the fate of CO2 droplets and bubbles at high pressure and in complexfree motion involvingchanging bubble dynamicsand surface film properties.Thusnumerical models have been created, and these have lead to differing conclusions. Holder et al. (16) modeled the behavior of a rising CO2 plume in the case ofa constantlygrowingfilm ofsolid hydrate. Since CO2-hydrate is more dense than either seawater or liquid CO2 itself, then this would slow the rise rate and eventually cause sinking. Herzog et al. (5) modeled the dissolution rate ofliquid CO2 without a hydrate film released at depths between 500 and 2000 m. They concluded that if the initialdrop radius is less than 1cm,complete dissolution would occur within less than 200 m ascent from the release point. A detailed set of model calculations incorporating plume dynamics has been carried out (6, 17), and yet no field data exist to compare models against actual observations. Experimental Section Our experiments tookplace at 800m depth in MontereyBay, California. The depth was chosen to match plans for an internationalexperiment designed primarilyto studyplume dynamics of CO2 released in the deep ocean (18). Remotely Operated Vehicle. We used the ROV Ventana (10),deployed bythe RVPoint Lobos for our work.The viewing system on Ventana used for real-time observation and permanentlyrecordingvisualimagesutilizesthedigital“high- definition” television format. The system captures images with a specially modified Sony HDC-750 high-definition television camera which digitizes the picture data and formats it to the SMPTE 292M HDTV interface standard with 2:1 interlace.The resolution ofthe images is 1080pixels vertically and 1920 pixels horizontally which is about five times the resolution of conventional video. The image data is continu- ally uplinked from the ROV to the ship for display and simultaneously recorded with a Panasonic HD2000 high- definition videotape recorderwithout anyfurthertranscoding steps. Still frame grabs obtained from the 30 frames/s recordingwere captured (ViewgraphicsCorp.)and processed using standard image processing software (Photoshop). CO 2 Release and Imaging. Early work had shown that it was impossible to obtain accurate size data and maintain within the field of view a CO2 droplet cloud that was free to move in all dimensions. Lack of a dimensional reference, effects of the vehicle motions, strong lateral forcing due to local currents, and the similarity of appearance of a droplet ofliquid CO2 to the ubiquitousgelatinousmarine organisms, *Correspondingauthor phone: (831)775-1706;fax: (831)775-1620; e-mail: brpe@mbari.org. † Present address: GEOMAR, Center for Marine Research, Wisch- hofstrasse 1-3, D-24148, Kiel, Germany. Environ. Sci. Technol. 2002, 36, 5441-5446 10.1021/es025909r CCC: $22.00  2002 American Chemical Society VOL. 36, NO. 24, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 5441 Published on Web 11/01/2002