Soil contamination with heavy metals has become a major concern due to anthropogenic activities such as mining and smelting, chemical industry and agricultural activities (Zhao et al., 2015; Zhong et al., 2020). Cadmium (Cd) is one of the most hazardous heavy metals, which could be easily accumulated in human body through food chain, resulting in the occurrence of many serious diseases such as blood disorders, organ damage and cancer (Liu et al., 2013a; Briffa et al., 2020). Therefore, the environmental-friendly and sustainable remediation technologies are urgently needed, such as the cost-effective and environmental-friendly techniques of biochar and phytoremediation (Yan et al., 2020; Gao et al., 2022; Shen et al., 2022).

Biochar is a carbon-rich material produced by thermochemical decomposition of biomass under anoxic conditions (Cha et al., 2016; Xie et al., 2022). The biochar with its large specific surface area, great porous structure, active functional groups and high cation exchange capacity has been widely used to adsorb heavy metals in soils and water environment (Narayanan et al., 2021; Chen et al., 2022; Gao et al., 2022; Issaka et al., 2022). For instance, many studies indicated that biochar could effectively immobilize Cd in soils by reducing the mobility and bioavailability, through high polarity and/or abundant chemical functional groups, e.g., hydroxyl, carboxyl, phenolic groups, and π electron-rich domain on the aromatic structures (Cai et al., 2020; Tu et al., 2020). These groups give biochar the property of high polarity, thus increasing the interaction between Cd and biochar, resulting in an increase of Cd immobilization (Sumaraj & Padhye, 2017; Wang R. et al., 2018). Additionally, the application of biochar in soils has often been reported to have positive effects on plant growth (e.g., increasing biomass) (Houben et al., 2013; Li et al., 2019). It can be imagined that when the biomass of plants, especially for plants with high biomass and Cd-tolerance, and/or Cd-hyperaccumulators growing in Cd-contaminated soils increase with application of biochar, more Cd would be stored in plants, leading to higher remediation efficiency.

Phytoremediation approach has also been suggested as a promising environmental-friendly technique to remediate soils contaminated with heavy metals (Sarwar et al., 2017; Shen et al., 2022). The main approach is the use of hyperaccumulator plants which could accumulate more (e.g. more than 100 times) metals in shoots compared to normal plants (He et al., 2017). Alternatively, it’s possible to use plants that can produce high biomass and be tolerant to above-threshold soil metal concentrations (Chen et al., 2022a). Generally, the population biomass is regulated by the planting density, which initially proportionally increases, levels off and then remains constant with increasing density (Li et al., 2016). Thus, it is expected that the large and Cd-tolerant plant with appropriate density could produce great biomass, and subsequently immobilize high Cd in tissues. For instance, the high phytoremediation efficiency of heavy metals in Festuca arundinacea (Qin et al., 2021), Eucalyptus globulus (Luo et al., 2018) and Typha domingensis (Viana et al., 2021) with appropriate plant densities.

Due to human-driven global change, most terrestrial ecosystems are experiencing great fluctuations in resource availability (Li et al., 2022; Tao et al., 2023). The fluctuating resource can influence the growth of plants directly by changing the nutrient uptake (Warren et al., 2003), or indirectly by affecting the activities of soil microbes (Carrero-Colón et al., 2006a; Carrero-Colón et al., 2006b), and then subsequently affects the heavy metal uptake of plants. Alternatively, the changed activities of soil microbes induced by pulsed nutrient can directly affect the heavy metal remediation, as microbial bioremediation has been used as an effective technique in the remediation of contaminated-soils with heavy metal (Wang et al., 2020; Xiao et al., 2020a).

Trifolium repens L. is an invasive perennial legume with large biomass and strong adaptability. It could grow as a dominant species in heavy metal contaminated soils, which possesses great potential for phytoremediation (Liu et al., 2018; Nandillon et al., 2019). Many previous studies have found great Cd accumulation in T. repens (Wang L. et al., 2018; Liu et al., 2019; Xiao et al., 2020c). In this study, we carried out a greenhouse experiment to investigate the interactive effects of planting density, biochar and nutrient fluctuation on the growth, Cd and nutrient of Trifolium repens population. We aimed to answer that (1) whether the Cd and nutrient uptake of Trifolium repens increased with higher planting density due to the possible increasing biomass; (2) whether the biochar effect was adjusted by nutrient fluctuation, as previous study pointed that the effect of biochar was not significant without nutrient application (Kuppusamy et al., 2016); (3) whether compared with constant nutrient application, pulsed nutrient application can promote more the growth of Trifolium repens, and subsequently the greater Cd and nutrient accumulation in tissues.

We found that the biomass increased significantly with increasing planting density ( Figures 2 , S1 ), which is consistent with many previous studies (Hagiwara et al., 2010; Li et al., 2013; Springer, 2021; Sun et al., 2022), because that higher planting density can produce higher productivity with sufficient light, water and nutrient (Chu et al., 2008). However, the growth efficiency of T. repens population was higher at the planting density of 3-individuals (e.g., + 107.18% for total biomass from 1-individual to 3-individuals) than of 6-individuals (e.g., + 27.70% for total biomass from 3-individuals to 6-individuals). This could be largely due to the application of biochar (significant interaction effect between planting density and biochar application; Figures 2 , S1 ; Table S1 ). The growth efficiency increased with increasing planting density without biochar application, whereas it was higher at planting density of 3-individuals than of 6-individuals with biochar application ( Figure S1 ). Additionally, when the planting density was 3-individuals, biochar application significantly increased the shoot biomass and total biomass, but this increase disappeared when the planting density was 6-individuals ( Figures 2 , S1 ). Additionally, we found that biochar application significantly decreased the ratio of root to shoot, especially when the planting density was 6-individuals ( Figures 2 , S1 ). These results indicate that seriously competition for resources between individuals at planting density of 6-individuals with biochar application may occur. Previous studies have suggested that biochar application could promote the plant growth, due to its benefits such as increased nutrient supply, water conservation and microbial activity (Sohi et al., 2010; Liu et al., 2013b). However, our results indicate that the effect of biochar application on the growth of plant population could be adjusted by planting density, and the planting density of 3-individuals could be an appropriate density for the plant growth. During the vegetative stage of plants, the levels of N and P in tissues represent the levels of nitrogen compounds and phosphorous compounds, especially protein and nucleic acid, respectively, and their levels could reflect the growth status to some extent (Chapin, 1980). In this study, under the treatment of biochar application, the higher growth efficiency of T. repens population at the planting density of 3-individuals than that at the planting density of 6-individuals might be due to the higher N concentrations and pool sizes in shoots and roots ( Figures 6A, B , 7A, B ). Surprisingly, neither nutrient fluctuation alone nor its interaction with density or biochar did affect the growth of T. repens population ( Figure 2 ; Table S1 ). This indicates that T. repens might not be sensitive to the nutrient fluctuation on the one hand. On the other hand, the nutrient dose (100%-strength Hoagland, 200 ml pot^-1) in the present study might be sufficient for the need of the growth of T. repens population, and then the fluctuating supply was ineffective, which can be proved by no significant differences of N and P concentration in soils between the application of constant nutrient and pulsed nutrient ( Figures 5B, C ; Table S3 ).

We found that the concentration of Cd in shoots was higher with biochar application than without biochar application, however, the difference decreased with increasing planting density ( Figure S2A ; Table S2 ). Under biochar application, the decrease of Cd concentration in shoots with increasing planting density might be due to the dilution effect with large biomass ( Figure 2A ). However, this explanation was not applicable to the unchanged Cd concentration without biochar application. Biochar application increased the Cd uptake in shoots but decreased it in roots, however, the effect of biochar application on Cd uptake was also adjusted by planting density. Biochar application benefited more the Cd uptake of T. repens population with lower planting density (i.e., 1-individual) in comparison to higher planting density (i.e., 3- and 6-individuals) ( Figures S2 , S3 ). This result indicates that increasing planting density might decrease the Cd uptake efficiency, although the biomass significantly increased. Overall, our results indicate that biochar application can largely promote the Cd uptake of T. repens population, which have been found in many previous studies (Wang et al., 2019; Cai et al., 2020; Tu et al., 2020; Xiao et al., 2020b), however, this promotion in the present study depends on the planting density. Generally, rhizobium could fix and provide large amount of the nitrogen nutrients to T. repens, and poor nitrogen nutrients in soil could promote the nitrogen fixation (Hoglund and Brock, 1987). In this study, at the early stage, rhizobium might not provide enough N at the planting density of 1-individual due to the unstable soil microbial composition in comparison to the planting density of 3- and 6-individuals. Meanwhile, biochar could not only immobilize Cd, but also absorb N, which could lead to the shortage of N in soils (Kuppusamy et al., 2016). When the soils were supplied with pulsed nutrients (especially low volumes of Hoagland solution for 3 weeks at the early stage), T. repens has to response to the N shortage and improve the uptake efficiency of N from the soils, which inevitably absorb more Cd into the plants. Alternatively, higher planting density is likely to cause strong intraspecific competition for light, nutrients and water in a certain spatial range, and in this circumstance, biochar application might have neutral and even negative impact. Biochar application significantly decreased Cd uptake in roots, and the reduction was greater at the planting density of 6-individuals than 3- and 1-individual(s), which possibly resulted from the lower ratio of root to shoot. Obviously, fewer root systems absorbed less Cd. Surprisingly, the soil bioavailable Cd increased under the circumstance that more Cd was accumulated in the T. repens population at higher planting density (i.e., 3- and 6-individuals). We guess that compared with 1-individual, more individuals in pots could change the soil properties more, which may lead to some chemical reactions or microbial activities that resulting in more bioavailable Cd.

Although nutrient fluctuation had no effect on the growth of T. repens population, but its interaction with planting density had significant effects on Cd uptake in tissues ( Figure S5 ; Table S2 ). Constant nutrient promoted Cd uptake in roots and the whole population at planting density of 3-individuals, however, this promotion decreased from planting density of 3-individuals to 6-individuals, which might mainly due to higher root Cd pool size ( Figures 4B , S5B ). We calculated average individual plant weight for each planting density treatment (the average biomass divided by the number of individual(s)) in constant and pulsed nutrients, respectively. The average root weight in the treatment of constant nutrients (0.66 g) at the planting density was higher than that of pulsed nutrients (0.47 g), thus more Cd was absorbed. Pulsed nutrient increased the Cd uptake in roots and the whole population with increasing planting density from 1-individual to 6-individuals ( Figures S5B, C ). But there was no significant difference for Cd uptake at the planting density of 6-indivduals between constant nutrient and pulsed nutrient ( Figure S5 ). Thus, our results suggest that constant nutrient could promote the Cd uptake of T. repens population at an appropriate planting density, such as planting density of 3-individuals in the present study. The nutrient (N and P) uptake of T. repens population was higher at higher planting density (i.e., 3- and 6-individuals) than at lower planting density (i.e., 1-individual), which could be due to the higher biomass ( Figure 2 ). At planting density of 3- and 6-individuals, biochar application significantly increased root N uptake ( Figures 5B , 6B ) but decreased root P uptake ( Figures 6D , 7E ), which indicates that biochar promotes more N uptake in comparison to P uptake.

We conclude that the growth and Cd and nutrient uptake of T. repens population increased with increasing planting density. The effects of biochar and nutrient fluctuation on the growth and Cd uptake of T. repens population were influenced by planting density. With biochar application, the shoot biomass and Cd uptake were increased more at the planting density of 3-individuals. Constant nutrient had no effect on biomass, but significantly increased root Cd uptake of T. repens at the planting density of 3-individuals. Thus, our results suggest that an appropriate planting density, such as the planting density of 3-individuals per pot in the present study, together with biochar and constant nutrients could promote the growth and Cd uptake of T. repens population.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

W-LZ: Conceptualization, Funding acquisition, Writing – original draft. Y-FW: Methodology, Resources, Writing – review & editing. JM: Data curation, Methodology, Writing – review & editing. PZ: Data curation, Methodology, Writing – review & editing. JC: Data curation, Methodology, Writing – review & editing. CS: Data curation, Methodology, Writing – review & editing.