Objectives We studied the formation of iron-plaque on root surface of rice, and the contents of soil Cd fractions in a Cd-contaminated paddy field, in order to clarify the mechanism of Chinese milk vetch (MV) and rice straw synergistic returning to field in controlling Cd uptake of rice.

Methods A field experiment was conducted in the single-cropping rice area along the Yangtsi River in Anhui Province. The five treatments in rice season included mere nitrogen fertilizer application at conventional rate (100N); reducing 20% of nitrogen fertilizer rate combined with MV 22500 kg/hm² (80N+M1); 80N combined with MV 37500 kg/hm² (80N+M2); 100N combined with MV 22500 kg/hm² and rice straw (100N+M1+S); 80N combined with MV 22500 kg/hm² and rice straw (80N+M1+S). Soil samples around the root zone of rice were collected at the tillering, jointing, booting and maturing stages to determine the contents of different Cd fractions. Rice root samples were taken at the same stages and the iron-plaque on the root surface was extracted to analyze the contents of Fe (DCB-Fe) and Cd (DCB-Cd) in the iron plaque. At the maturing stage, the contents of Cd in grains were analyzed.

Results Compared with 100N treatment, the other treatments decreased soil available Cd content at tillering and jointing stages by 12.37%−16.96% and 9.59%−20.30% (P<0.05), respectively; 80N+M1 and 80N+M1+S treatments decreased the available Cd content at booting stage by 18.64% and 20.79%, and the available Cd content at maturing stage by 9.55% and 10.46%, respectively. Compared with 100N treatment, the proportion of acid-Cd in the other four treatment soils were declined by 6.25%−11.48%; the proportion of residual-Cd in 80N+M2 treatment was increased by 93.77%. The DCB-Cd and DCB-Fe contents showed consistent with the extension of rice growth period. Compared with 100N treatment, DCB-Fe content in 80N+M1 and 80N+M1+S treatments were increased by 167.78% and 128.45% at tillering stage, and 67.42% and 97.19% at maturing stage, respectively (P<0.05), and the DCB-Cd contents were increased by 92.79% and 50.47% at tillering stage and 17.75% and 25.44% at maturing stage, respectively (P<0.05). The root Cd content in 80N+M1 and 80N+M1+S treatments were decreased by 11.21% and 12.93%, respectively, and the rice Cd content in 80N+M1+S treatment was the lowest and was decreased by 17.17% compared with 100N treatment (P<0.05). The correlation analysis showed that DCB-Cd was positively correlated among the growing stages of rice. The Cd content was positively correlated with DCB-Cd at booting stage and negative correlated with DCB-Cd and DCB-Fe at tillering and maturing stages. Path analysis showed that the application of MV and rice straw had positive (P<0.05) effects on the contents of DCB-Fe and DCB-Cd, but negative effects on soil available Cd, and rice Cd content, and the effect of rice straw was greater than that of MV.

Conclusions Both MV and rice straw returning to field promote the fixation of available Cd in paddy field and the formation of iron-plaque on the surface of rice roots. The rice straw mainly intercepts Cd absorbed by the roots through the iron-plaque on the root surface, while MV mainly adsorbs and fixes available Cd in the soil environment. The most effective inhibition effect on Cd absorption was observed under 80% nitrogen fertilizer rate combined with MV 22500 kg/hm² and rice straw in Cd polluted paddy soil.