The use of plant species for cleaning polluted soils and waters named as phytoremediation has gained increasing attention since last decade, as an emerging cheaper technology.
It depends on the heavy metals you want to remove from the contaminated soil. For example, Oenothera affinis accumulates Cu or Lens culinaris, Cicer arietinum and Lupinus albus accumulate Hg. The application of one plant or another will also depend on the area where the contaminated soil is located. Therefore, a previous study of the area (climatology, environmental impact, etc) should be done.
The aquatic plants like Euryale ferox (prickly water lily), Trapa natans (water caltrop), and Ipomea aquatica (water spinach) are some of the examples in the aquatic habitat especially lentic habitats where these plants incorporate heavy metals and convert them into useful edible part for human beings.
I must differ with respect: contaminated plants should not be put into the food chain Rakesh. In addition, consideration must be given to where the contaminants will be dumped.
(1) It's a question like all mining operations of downsizing and then ligating in this case from organic to mineralized metal(oid) species, and or combinatorial methods and stabilisation sediment strata with antibacterial bentonite downstream. Several technologies are being investigated and to much to go into here, including bacterial methods akin to gold biox recoveries in gold mining. Similar could be said for say water hyacinth, essentially sulphur metabolize, being related to garlic. You may as well call hyacinth floating garlic (also used by miners to chelate lead in their bodies). Not only was this only used to recover dangerous cadmium in remediation, but also able to recover 1kg gold and 1kg silver per hectare of mine waters in high waste systems. Generally speaking ferns, fungi, algae bacteria generally for arsenic due to phospholipids, along with Douglass fir, Hemlock (Tsuga heterophyla) and mountain hemlock (Tsuga martensiana) . There are other methods we will publish shortly. Brack fern (Piterus Vittata), up to several k ppm As in dried fronds and 2.2% in spiked soils. Note fucus is particularly useful since it can grow in brackish waters. And see weeds can accumulate As, as mentioned algae.
Accumulators
Serbertia acuminata for nickel (up to 25.7% Ni in dried Sap). There are more plants that hyper cumulate nickel than any others, many from Cuba (Brooks (1998).Berkheya coddii has been identified in phytomining. Others include Buxaceaae, Euphorbiacea, Flacourtiaceae.
Ononis natrix (goat root from Morocco) for Chromium, (up to 88ppm established in dry stem (Dunn 1996). Commonly over Cr mica (fuchsite), soluble Mn carbonates with Cr, (Ledum groenlandicum) or Labrador tea along with Au, Co, Th, U, W in stems. Haumaniastrum (mint family member) eg H. katangense for Cobalt-Copper, in African copper belt, known as one of several "copper flowers" as it also accumulates Cu, up to 1% Cu, as does Euphorbia in Iran. Also Bacellium homblei copper flower. Rubiacea (bedstraw ) Penats family, Zingiberacea (ginger family-Costa) leaves for Cobalt. In morocco several species accumulate cobalt; including Veronica and Anvellia garcinii. Rubber vine (Cryptostegia grandifolia) and some conifers accumulate copper over mineralisation (Cu-Mo-Au in British Columbia )
Tree fern (Amazonia) for mercury (up to 4,000ppm Hg); Willow for molybdenum (up to 9,600 ppm Mo in leaf ash) , up to17,000ppm Mo in leaves of Epilobium angustifolium; Australian Acacia for strontium and rubidium, also vines such as Clausia (amazons); For sodium Atroplex (salt bush) and Fucus (up to 25% Na) ,Garlic /water hyacinth for Sulphur and selenium (Fabaceae as well), and related Tellurium in Vetche Fabaceae types eg Astragalus s eg Pinus Bnaksian for Tantalum; Birch, willow, Alnus sinuata , some spruce, pine Douglass fir for zinc; Brazil nuts for radium (and selenium); Lodgpole pine for manganese, wide variation occurs along with Fe, some species can accumulate up to 4,190 ppm M as shown in the Eden project.
(2) Secondly the techniques of Parisa Ziarati in preventing the metals entering the food chain and herbaceous/vegetables after remediation is another whole chapter. She has developed some highly efficient systems, and is working on arsenic as well.
(3) Thirdly some accumulation such as Fabaceae species are beneficial to not only prosecting but ethno medicine. We will publish shortly as a medical geology article, in the pipeline.
just in addition to Daniel Tecles comment, it also depends on the tollerance of the plants to metallotoxicity and the amount of contamination soil level. Tamarix species such as Tamarix chnesis and T. ramosissima as well as Eucalyptus species such as E. nitens and E. camaldulensis has often been used in close to tailing dams or mining sites for phytoremediation of heavy metals. Fast grwoing trees with large biomass turn over and large root surface area as well as thier origin (indigenous or exotic) are few of the things to consider in selecting trees for phytoremediaton of soil contaminants. for wetlands the reeds are the best (Arundo donax and Phragmites australis) and from the aquatics water hyacinth is top of the list.iin
Site/ location specific was not mentioned in the question and the potential plant for this could range from submerged to terrestrial plant depending on the objectives of your research. However, i would advise to check the local plant available and leverage on native species of your area. In tropical area of Asian region for example, water hyacinth is a problematic invasive weed. though it is native to South America.