Effects of Silicon Application on Nitrogen Migration in Soil–Rice Systems under Cadmium Stress

Abstract

Cadmium (Cd) contamination is a serious threat to plants and humans, which subsequently impairs sustainable agricultural production and ecosystem service. Silicon (Si) has been applied to mitigate Cd toxicity, but inevitably has direct and indirect impacts on nitrogen (N) behaviors in soil and plants. However, what role plants play in the N response to Si in soil–rice systems under Cd stress is not known. Therefore, the effects of Si on N migration through different pathways in the soil-rice system were systematically investigated in a rice-cultivation lysimeter experiment. The rice was planted in Cd-contaminated (5 mg kg⁻¹) and uncontaminated soils with three levels of Si application (0, 100, and 200 kg SiO₂ hm⁻²), and the contents of N and Cd in different forms in plants and soils were measured. The group without Cd and Si was set as CK. The study reported that Cd stress caused Cd accumulation in plants, inducing a decrease of 26.0~83.4% in plant dry weights and a decrease of 15.7~46.6% in N concentration compared with CK. Moreover, the leaching of N in soils was increased by Cd, in which the NO₃⁻-N rather than the NH₄⁺-N was leached out. These adverse effects on the plant growth and soil N loss were significantly alleviated by Si application in two ways: (1) the Cd availability in soils was reduced with the acid-extractable Cd (the Cd form with high mobility), decreasing from 1.07 to ~0 mg kg⁻¹; (2) the Cd uptake and translocation in plants were restricted, with the Cd content decreasing by 59.1~96.4% and the translocation index decreasing from 17.7% to 2.2%. The combination of the two mechanisms consequently increased the N absorption of plants from 1.35 to 2.75~3.5 g. The results of the N mass balance calculation showed that, compared with soil N flux, plant-absorbed N contributed predominantly (43.9~55.6%) to the soil total N variation. Moreover, there is a significant trade-off between plant-absorbed N and soil N flux. The magnitude and direction of the soil N flux were greatly and negatively affected by plant-absorbed N during the flooding period. Hence, we conclude that Si application could reduce the leaching of N in soil–rice systems under Cd stress, mainly due to the promotion of the N absorption of plants rather than N immobilization in soils. This study provided new evidence that plants played a dominant role in N response to Si in soil-rice systems under Cd stress.

1. Introduction

Cadmium (Cd) is a heavy metal that can readily accumulate in plants and cause biochemical and physiological disorders due to its high extractability and toxicity [1]. Cd exposure may cause severe chronic health effects including itai-itai disease, lung cancer, and kidney damage [2]. Diet is a major source of Cd exposure, and the consumption of rice (Oryza sativa L.) has been reported as one of the primary sources of Cd intake among the residents in some Asian countries, posing a high risk to human health [3,4]. Furthermore, contamination of soils by heavy metals endangers ecological health, impairs critical agricultural and natural resource systems, leads to socioeconomic instability, and necessitates resource-intensive remediation methods, posing a great threat to sustainability. Therefore, it is imperative to explore effective measures to reduce Cd toxicity in soil–rice systems.

Exogenous silicon (Si) has been widely used to alleviate Cd toxicity in soil–rice systems because of its advantages in terms of nutrient supply, soil conditioning, cost-effectiveness, and disease/insect resistance [5]. Different mechanisms occurring in soils and plants can explain the reduction of Cd toxicity in plants upon Si application. Generally, Si plays an important role in strengthening cell walls to provide a physical barrier to Cd entry [6], in restricting Cd translocation from roots to edible parts [7], and in alleviating oxidative damage induced by Cd stress [8], thereby improving plant growth [9,10,11,12]. In addition, Si application primarily increases soil pH, leading to Cd immobilization and co-precipitation [13,14], minimizes exchangeable speciation of Cd through Cd chelation, and forms an amorphous silica barrier through Si-Cd complexation, which ultimately reduces the bioavailability of Cd in soils [15]. Moreover, these changes due to Si addition could potentially influence the behaviors of N in soil–rice systems.

Silicon has been proven to boost rice growth and parallel an increase in the N content of tissues even under unstressed conditions [16]. In soils, nutrients like N are mobilized by Si due to the competition with other elements for binding sites on organic matter and mineral surfaces [17]. Meanwhile, more NH₄⁺-N loss through ammonia volatilization is expected because of the increased soil and water pH after Si application [18]. In contrast, Zhang et al. [19] reported the conservation effect of the addition of Si fertilizer on NO₃⁻-N in soils. Moreover, Si addition could have indirect influences on N behaviors in soils or plants by affecting Cd bioavailability. A lot of studies have reported that the uptake and translocation of Cd in plants is effectively reduced by the supply of Si, consequently mitigating the negative effects of Cd on N metabolisms in plants [6,7]. In soils, Si application improves the retention of NH₄⁺-N, which is attributed to the decrease in Cd²⁺, which competes with NH₄⁺ for the adsorption sites in soils [20]. Si increases nitrous oxide (N₂O) emissions by chelating with a nitrification inhibitor (heavy metals) that prevents nitrification [21]. These studies primarily separately focused on the direct/indirect impacts of Si on N in soils or plants. However, soil-rice is a continuous system in which the application of Si to alleviate Cd toxicity essentially involves the interactions between Cd, N, and Si in soils and plants. The responses of N to Si through the pathways of plants and soils in soil–rice systems are not synchronously considered. Therefore, the migration and transformation of N in the Cd-contaminated soil–rice system when Si is applied lacks comprehensive and quantitative research.

To address this knowledge gap, we conducted a rice-cultivation lysimeter experiment to investigate the effects of Si application on N migration in a soil–rice system under Cd stress. The specific objectives of this study are to (1) quantify N migration through the soil and plant pathway after Si application; and (2) clarify the mechanism by which Si affects N migration in soil–rice systems under Cd stress.

2. Materials and Methods

2.1. Materials

The soil used in this study was collected from the top layer (0–15 cm) of a paddy field located in Qionglai, Chengdu City, Sichuan Province, China (30°24′ N, 103°32′ E) (Figure 1). The soil was air-dried and ground to pass through a 2 mm mesh. The prepared soil was then used for rice cultivation in an off-site lysimeter experiment. The major physicochemical properties of the soils are listed in

Table 1. Initial physiochemical parameters of soil.

Soil Texture	pH	Organic Matter mg g−1	Available Si mg g−1	NH4+-N mg kg−1	NO3−-N mg kg−1	Total Cd mg kg−1
Silt loam	6.46	47.75	68.82	9.50	65.01	0.26

Note: available Si represents the form of Si which is extracted with weak acid and determined via the Silicon–molybdenum blue colorimetric method.

. Sodium silicate (NaSiO₃·9H₂O) was used as the source of exogenous Si in the experiment, and cadmium bromide (CdBr₂) was added to achieve the preset contamination degree of Cd. Conventional indica rice (“Rong-18B”) was selected as the plant material for the present study.

2.2. Lysimeter Experiment

2.2.1. Experimental Design and Sample Collection

The rice-cultivation lysimeter experiment was conducted at the test site of Sichuan University during the period of 1 June–30 September 2020. Three levels of Si application (0, 100, and 200 kg SiO₂ hm⁻², hereafter referred to as Si₀, Si₁, and Si₂, respectively) with or without Cd contamination (5 mg kg⁻¹) were considered in the lysimeter experiment, resulting in six treatment combinations (CK (Si₀), Si₁, Si₂, CdSi₀, CdSi₁, and CdSi₂) with three replicates for each treatment. The Cd dose was determined based on the background value of the Qionglai paddy field, while the application of Si was based on a previous study by Cuong et al. [16]. The dimensions of the PVC lysimeters used for each treatment were 65 cm × 30 cm × 30 cm. An outlet was set vertically on the sidewall of each lysimeter at a depth of 42.5 cm (Figure 1). A 5 cm filter layer consisting of sand grains (repeatedly washed with water and acid solution) was first loaded at the bottom of the lysimeter. The abovementioned prepared soil was then packed into a lysimeter layer by layer (5 cm) to achieve a bulk density of 1.48 g cm⁻³. The experiment was designed such that only the topsoil layer (0–15 cm) in the lysimeter was contaminated by Cd. To eliminate the influence of soil variation, this was achieved by evenly mixing the corresponding dose of CdBr₂ with soils rather than field experiment and in situ collecting the contaminated soil. These contaminated soils were carefully collected and treated after the experiment. Subsequently, the sidewall of each lysimeter was covered with a thermal insulation film to minimize heat perturbation. Before transplanting the rice, the prepared columns were irrigated with tap water and allowed to rest for 30 days, during which approximately 5 cm of water was maintained at the soil surface. Afterward, four 30-day-old uniform rice seedlings were transplanted into each lysimeter on 1 July. Nitro-compound fertilizer containing N:P:K at 140:90:60 kg ha⁻¹ was applied to each lysimeter at the time of transplantation. The Si was added in the form of Na₂SiO₃ solution three days later. Irrigation was regularly conducted to maintain an approximately 5 cm water layer on the soil’s surface till 30 August, dividing the whole experimental period into the flooding period (1 July–29 August) and non-flooding period (30 August–30 September).

Ten milliliters of percolated water samples were collected through the outlet weekly, filtered, and stored at 4 °C for further analysis. Soil samples were collected at depths of 5, 15, 25, and 35 cm on 29 August and 30 September (after harvest). Part of the fresh soil was extracted using a 2 mol L⁻¹ KCl solution and saturated CaSO₄ solution to determine the NH₄⁺-N and NO₃⁻-N content, respectively. The remaining fresh soil was air-dried and ground for measuring pH, electrical conductivity (EC), and Cd content. The rice plants were manually harvested from each lysimeter at physiological maturity. After harvest, the plant samples were rinsed repeatedly with tap water, followed by deionized water. Subsequently, the individual plants were divided into roots, leaves, stems, and grains and oven-dried at 80 °C for one month to determine the dry weight.

2.2.2. Water, Plant, and Soil Sample Analysis

The NH₄⁺-N and NO₃⁻-N concentrations in the water samples and soil extracts were measured using an automatic chemical analyzer (EasyChem plus, Systea, Italy).

The oven-dried plant samples were milled into flour using a stainless-steel grinder. Some of the prepared samples were digested using HNO₃-HClO₄ (v:v = 4:1). The Cd concentrations in the solutions were determined using an atomic absorption spectrophotometer (AA6880, Shimadzu, Kyoto, Japan). The tissue N content was determined using the Kjeldahl method, as described in the study of Sáez Plaza et al. [22].

The pH and EC of the solution extracted from the soil water (soil:water = 1:10) suspension were measured using a PHS-3C pH meter and DDS-307A conductivity meter (REX, INESA Scientific Instrument Co., Ltd., Shanghai, China), respectively. Soil samples (100 mesh) were digested with HNO₃-HClO₄-HF (v:v:v = 5:1:1) to analyze the total Cd content in the soils. Metal fractions were extracted to evaluate metal bioavailability and transformation using the BCR procedure [23], defining metals into four fractions: acid-extractable (EX-Cd), reducible (RD-Cd), oxidizable (OX-Cd), and residual fractions (RS-Cd). The EX–Cd has the highest mobility in soils and is easily absorbed by the plants while the RS–Cd has the lowest bioavailability. Standard reference material (GSS-1), duplicates and blank samples were used to control accuracy and precision for all the sample analyses.

2.3. Calculation of Nitrogen Mass Balance

Before calculating the N mass balance, we hypothesized that: (1) N loss due to ammonia volatilization and N₂O and NO emissions can be neglected (Q₀ = 0) due to the trace amount; (2) other forms of N are not considered except for NH₄⁺-N and NO₃⁻-N, which account for the majority of N in soils; and (3) N input due to fertilization and N output due to plant uptake only occur in the 0–20 cm soil layer. Therefore, the average soil N flux (i.e., positive indicates downward) during the entire growth period between adjacent layers was calculated based on the mass balance (Figure 1) as follows:

$$\begin{array}{l}{Q}_{20}=F+Q_0-U-∆N_{0~20}\\ Q_{30}=Q_{20}-∆N_{20~30}\\ Q_{40}=Q_{30}-∆N_{30~40}\end{array}$$

(1)

$$Q_0=0$$

(2)

where $Q_z$ (g d⁻¹) denotes the average N flux at a depth of $z$ (cm); $∆N_{z~z+d}$ (g) denotes the N variation in the soil layer of $z~z+d$, particularly, $d$ (cm) is the soil layer thickness; $F$ (g) denotes N input through fertilization; 13 g of fertilizer containing 37% of N was applied for each lysimeter, $F=13\times 0.37=4.81$ g; and $U$ (g) denotes the N uptake of rice during the entire growth period, which is calculated as follows:

$$U={DW}_{grain}{c}_{grain}+{DW}_{leaf}{c}_{leaf}+{DW}_{stem}{c}_{stem}+{DW}_{root}{c}_{root}$$

(3)

where $DW$ (g) and $c$ (g kg⁻¹) represent the dry weight and N content of the different tissues in rice, respectively.

The entire growth period was divided into two periods: the flooding period from the transplantation day (1 July) to 29 August, and the non-flooding period from 30 August to the harvest day (30 September). To analyze N migration in the soil–rice system in detail, the N mass balance was calculated separately for each period, during which the N uptake of the rice was unknown. The N uptake of rice during the second stage was computed as follows:

$$\begin{array}{l}{Q}_{30}={-Q}_{40}-∆N_{30~40}\\ Q_{20}=-Q_{30}-∆N_{20~30}\\ U^{t~t+1}=Q_0-Q_{20}-∆N_{0~20}\end{array}$$

(4)

where $U^{t~t+1}$ (g) denotes the N uptake of rice during the period $t~t+1$, and the superscripts $t$ and $t+1$ represent 30 August and 30 September, respectively. The other terms are the same as those in Equation (1). However, the N flux at the lower boundary of the soil layer ($Q_{40}$) remains unknown. As shown in Figure 1, the filter layer is regarded as a semi-closed system in which external N exchange only occurs at a depth of 40 cm and the outlet. Hence, the N variation in the solution in the saturated filter layer could be computed as follows:

$$∆N_{40~45}=V\eta (c_{40~45}^{t+1}-c_{40~45}^t)=Q_{40}-P$$

(5)

where $V$ (cm³) is the volume of the filter layer, and $V=30\times 30\times 5=4500$ cm³; $\eta$ is the porosity of the filter layer, and $\eta =0.3$; $c_{40~45}^{t+1}$ (g mL⁻¹) is the N concentration of the solution in the filter layer on 30 September; and $P$ (g) is the mass of N percolated through the outlet during the period from 30 August to 30 September, which can be neglected in this study. $U^{t~t+1}$ is acquired by substituting Equation (5) into Equation (4), and then the N uptake of rice at the first stage was calculated as follows:

$$U^{t-1~t}=U-U^{t~t+1}$$

(6)

where $U^{t-1~t}$ (g) denotes the N uptake of rice during the period from $t-1$ to $t$, and superscript $t-1$ represents 1 July; the other terms are the same as those mentioned above. Therefore, the N mass balance in the first stage was calculated as follows:

$$\begin{array}{l}{Q}_{20}=F+Q_0-U^{t-1~t}-∆N_{0~20}\\ Q_{30}=Q_{20}-∆N_{20~30}\\ Q_{40}=Q_{30}-∆N_{30~40}\end{array}$$

(7)

where the terms are the same as those mentioned above.

2.4. Translocation Index of Cadmium in Plants

The translocation index ($TI$) of Cd in rice was calculated as follows [8]:

$$TI=\frac{{Cd}_{grain}}{{Cd}_{grain}+{Cd}_{straw}}$$

(8)

where ${Cd}_{grain}$ and ${Cd}_{straw}$ denote the Cd content (mg kg⁻¹ dry weight) in the grains and straw (stems + leaves), respectively.

2.5. Data Analysis

Analysis of variance (ANOVA) was performed to assess the significant difference between treatments at a significance level of 0.05 using GraphPad Prism 9 software (GraphPad Software, San Diego, CA, USA). The graphics were drawn using GraphPad Prism 9 software too.

3. Results

3.1. Rice Yield

As shown in Figure 2a, the plant dry weight was increased by Si but reduced by Cd. Compared with CK, the dry weight of grains, leaves, stems, and roots increased by 67.7%, 25.6%, 17.3%, and 43.5% in Si₁, and by 146.7%, 52.5%, 30.7%, and 91.6% in Si₂, respectively, whereas those in CdSi₀ were 26.0%, 72.4%, 68.2%, and 83.4%, respectively. However, this reduction in dry weight was significantly alleviated by Si (p < 0.05). In comparison with CdSi₀, the dry weights of grains, leaves, stems, and roots increased by 142.8%, 47.1%, 56.4%, and 44.3% in CdSi₁, and by 199.1%, 87.4%, 82.9%, and 89.0% in CdSi₂, respectively. Regardless of whether the soil was contaminated with Cd, the increasing effect of Si was most pronounced in the grain yield. As shown in Figure 3, the power function of the total dry weight and total Cd content in plants also clearly indicated the dry weight decreased with the Cd content.

3.2. Cadmium in Rice and Soil

As shown in Figure 2b, although the Cd content of grains in CK was lower than the maximum levels of contaminants in food (0.2 mg kg⁻¹) of the National Food Safety Standard (GB2762-2017) [24], the Si application further reduced the Cd content in grains, as demonstrated in Si₁ and Si₂. The Cd content in all tissues of CdSi₀ was significantly higher than in the tissues of CK (p < 0.05), whereas it decreased in different tissues with increasing Si application. For instance, the accumulation of Cd in the root, stem, leaf, and grain of rice was decreased by 60.4%, 71.4%, 72.6%, and 91.4% in CdSi₁, while in CdSi₂, it was reduced by 75.4%, 59.1%, 77.7%, and 96.4%, respectively. Moreover, in all treatments, the Cd content in the roots was higher than that in the aboveground parts (stems, leaves, and grains), consistent with the results of Shi et al. [25]. The TI was calculated to demonstrate the ability of plants to transport Cd. The highest TI was found in CK (39.4%), followed by Si₁ (31.0%), CdSi₀ (17.7%), Si₂ (8.3%), CdSi₁ (6.2%), and CdSi₂ (2.2%), suggesting that Si effectively reduced Cd translocation from straw to grains.

Figure 4 shows the distribution of various forms of Cd along the soil profile on 29 August 2020. The Cd-free treatments (CK, Si₁, and Si₂) had a homogeneous vertical distribution of total Cd, with a concentration of approximately 0.5 mg kg⁻¹. In contrast, the Cd treatments (CdSi₀, CdSi₁, and CdSi₂) had an inhomogeneous vertical distribution, with Cd concentrated in the topsoil layer (0–5 cm) with a concentration of approximately 6 mg kg⁻¹. Regarding the speciation of Cd, RS-Cd accounted for most of the total Cd content for the entire soil profile in the Cd-free and Cd treatments, followed by EX-Cd, RD-Cd, and OX-Cd. In the topsoil layer (Figure 4a), a remarkable increase of EX-Cd (1.07 mg kg⁻¹) was observed in CdSi₀ while the content of EX-Cd was marginally detected, and RS-Cd (5.35 mg kg⁻¹) accounted for the majority of the total Cd in CdSi₁ and CdSi₂. All forms of Cd in the lower soil layers exhibited similar distributions in the Cd- and Cd-free treatments.

Figure 5 shows the distribution of various forms of Cd in the soil profile on 30 September 2020. Compared to the profile on 29 August, the Cd-free treatments and CdSi₀ were relatively stable in all forms of Cd. However, CdSi₁ and CdSi₂ had a sharp decrease in all forms of Cd in the topsoil layer (Figure 5a). In the lower soil layers (Figure 5b–d), various forms of Cd in all treatments decreased to some extent, especially the RS-Cd. Interestingly, a notable increase in total Cd was observed at a depth of 15 cm only in CdSi₁, with 83.5% of the total Cd being RS-Cd.

3.3. Nitrogen in Soil–Rice System

3.3.1. Total Nitrogen in Rice

As shown in Figure 6, grains had the highest N content, followed by leaves, stems, and roots in all treatments. The N content of grains and leaves was lower in CdSi₀ than in CK, whereas no significant difference was observed in the stems and roots between the two treatments. Si substantially increased the N content of grains, leaves, and stems but had a negligible effect on the N content of the roots. For example, the N content of grains increased by 4.7% and 9.6% in Si₁ and Si₂, respectively, compared with CK. Moreover, the decrease in N content induced by Cd could be reduced by Si; for example, the N content of grains was increased by 10.1% and 14.1% in CdSi₁ and CdSi₂, respectively, compared with that in CdSi₀.

3.3.2. NH₄⁺-N and NO₃⁻-N in Soils

Figure 6 shows the NH₄⁺-N content of the soil on 29 August. The NH₄⁺-N content at each depth (5, 15, 25, and 35 cm) in CdSi₀ was higher than that in the CK. The NH₄⁺-N content decreased to 5 mg kg⁻¹ at a depth of 35 cm in CK, but it increased steeply at the same depth in CdSi₀. The NH₄⁺-N content of soils at depths of 5, 15, and 25 cm in Si₁ and Si₂ was lower than in CK, whereas in the soil at 35 cm depth, the content was significantly higher in Si₁ and Si₂ than in CK. In CdSi₁ and CdSi₂, the soil NH₄⁺-N content was intermediate between that of CdSi₀, Si₁, and Si₂.

Figure 7b shows that the soil NH₄⁺-N content in all treatments on 30 September decreased to different degrees compared with that in the soils on 29 August. In CdSi₀, the NH₄⁺-N content of soils at depths of 15, 25, and 35 cm decreased significantly compared with that in soils on 29 August. It is worth noting that there is a dramatic decrease in the NH₄⁺-N content of soils at depths of 15 cm in Si₁ and Si₂.

Figure 8a shows the NO₃⁻-N content of the soils on 29 August. In CK, the NO₃⁻-N content decreased gradually from the surface down, until it rose sharply to the maximum value at 35 cm. The application of Si led to a more uniform distribution and lower NO₃⁻-N along the soil profile compared with CK, as indicated by the NO₃⁻-N content in Si₁ and Si₂. A lower concentration of NO₃⁻-N along the soil profile was observed in CdSi₀ compared with CK, especially at a depth of 35 cm. Compared to CdSi₀, the NO₃⁻-N content of the top layer decreased with the Si dose, whereas the NO₃⁻-N content of the lower layer increased with the Si dose in CdSi₁ and CdSi₂. Similar distributions of NO₃⁻-N were observed in all treatments on 30 September, as shown in Figure 8.

3.3.3. Nitrogen Migration on Different Pathways

Table 2. Mass balance of nitrogen in the soil-rice system through the experimental period.

Depth (cm)	Nitrogen Variation in Soil (g)						Plant Total Nitrogen (g)						Soil Nitrogen Flux (mg d−1)
	CK	CdS0	S1	S2	CdS1	CdS2	CK	CdS0	S1	S2	CdS1	CdS2	CK	CdS0	S1	S2	CdS1	CdS2
1 July–29 August
0–20	−5.37	−5.62	−6.02	−6.40	−6.02	−5.92	3.02	0.65	5.78	8.13	2.22	2.79	39.13	82.80	4.08	−28.97	63.25	52.21
20–30	−0.54	−0.67	−0.62	−0.72	−0.75	−0.44	-	-	-	-	-	-	48.19	93.98	14.39	−17.04	75.80	59.58
30–40	0.07	−0.52	−0.48	−0.67	−0.39	−0.48	-	-	-	-	-	-	47.06	102.60	22.31	−5.80	82.30	67.64
30 August–30 September
0–20	−0.44	−0.39	−0.30	−0.19	−0.24	−0.38	0.88	0.71	0.69	0.46	0.53	0.71	−14.68	−10.65	−13.22	−9.14	−9.69	−11.08
20–30	−0.17	−0.14	−0.12	−0.09	−0.10	−0.19	-	-	-	-	-	-	−8.95	−5.94	−9.35	−6.08	−6.51	−4.78
30–40	−0.27	−0.18	−0.28	−0.18	−0.19	−0.14	-	-	-	-	-	-	−0.07	−0.10	−0.09	−0.08	−0.14	−0.03
1 July–30 September
0–20	−5.81	−6.00	−6.32	−6.58	−6.26	−6.30	3.90	1.35	6.47	8.60	2.75	3.50	21.20	51.65	−1.69	−22.36	38.94	31.11
20–30	−0.72	−0.81	−0.73	−0.81	−0.85	−0.63	-	-	-	-	-	-	29.14	60.67	6.48	−13.39	48.36	38.13
30–40	−0.20	−0.69	−0.75	−0.85	−0.58	−0.63	-	-	-	-	-	-	31.35	68.37	14.84	−3.89	54.82	45.08

Note: Negative nitrogen variation indicates a decrease, positive in plant total nitrogen, and downward in nitrogen flux.

shows the results of the mass balance calculation of N in the soil–rice system throughout the experimental period. The total N absorbed by plants in CdSi₀ was lower than that in CK, whereas the plants in Si₁ and Si₂ absorbed more N than those in CK. The application of Si could significantly alleviate the inhibition of N uptake due to Cd. For example, the total absorbed N was increased from 1.35 g in CdSi₀ to 2.75 g in CdSi₁ and 3.50 g in CdSi₂. Moreover, considerably more N was absorbed by plants during the flooding period (1 July–29 August) than during the non-flooding period (30 August–30 September). For instance, 3.02 g N was absorbed during the flooding period, while 0.88 g N was absorbed during the non-flooding period in CK.

As indicated in Table 2, the soil N flux for CdSi₀ was higher than that for CK from 1 July to 30 September at a certain depth; for instance, at 0–20 cm depth, the soil N flux for CdSi₀ was 51.65 mg d⁻¹, which was higher than the 21.2 mg d⁻¹ for CK at the same depth. This explains why more N moved downward in CdSi₀ than in CK throughout the experimental period (1 July–30 September). By comparing soil N flux between CdSi₀, CdSi₁, and CdSi₂ at certain depths, it can be seen that the soil N flux decreased with an increase in the Si concentration. For example, at 0–20 cm depth, the soil N flux for CdSi₀, CdSi₁, and CdSi₂ was 51.65, 38.94, and 31.11 mg d⁻¹, respectively. Regarding Si_1, with the addition of 100 kg SiO₂ hm⁻², the soil N flux at 0–20 cm depth was −1.69 mg d⁻¹ (negative flux represents upward movement), while the soil N flux remained downward (6.48 and 14.83 mg d⁻¹) at depths of 20–30 cm and 30–40 cm. Regarding Si_2, with the addition of 200 kg SiO₂ hm⁻², upward movement was identified for the soil N flux at each depth. The soil N flux decreased with an increase in depth; for example, it was −22.36, −13.39, and −3.89 mg d⁻¹ at depths of 0–20, 20–30, and 30–40 cm, respectively.

A nitrogen mass balance calculation was carried out for different periods to describe the N migration over time. During the flooding period (1 July–29 August), the N fluxes at each depth in all treatments decreased, except for Si₂. The magnitudes of these downward soil N fluxes were in the following order: CdSi₀ > CdSi₁ > CdSi₂ > CK > Si₁. During the non-flooding period (30 August–30 September), the soil N fluxes at each depth in all treatments were upward, and no significant difference was observed between treatments. Furthermore, the soil N flux magnitude during the non-flooding period was much smaller than that during the flooding period and showed a vertical distribution, with a gradual decline from top to bottom.

4. Discussion

4.1. Effects of Si on Plant Growth and N Content

Silicon (Si), a beneficial element, has been proven to improve rice growth and grain yield, particularly under stressful conditions [26,27]. Detmann et al. [26] reported that the harvest index of rice was improved, whereas N levels were unaltered, by Si addition. However, in our present study, increases in both the dry weight and N content of all rice parts was observed after the Si addition (Figure 2 and Figure 6), which is consistent with the findings of Cuong et al. [16]. The increases might be attributed to the positive effect of Si in increasing growth and yield characteristics, enhancing the pollen viability and photosynthetic activity [26], and improving nutrient uptake [28,29].

Cadmium seriously inhibits rice growth by suppressing photosynthesis [8,10], generating excessive reactive oxygen species (ROS) [3,30], and restricting the uptake of nutrients [31]. Moreover, Cd stress can greatly reduce the activities of nitrate reductase (NR), glutamine-synthetase (GS) [32], and glutamate synthetase (GOGAT), which are three key enzymes involved in N assimilation in rice [33]. However, these Cd stresses in rice plants were significantly alleviated by Si addition [34]. Silicon can significantly reduce Cd uptake by maintaining cell wall integrity [6], as it increases the amount of Cd that binds to the cell wall. Furthermore, Si application can promote Fe plaque formation on root surfaces, which hinders the entry of Cd into plants [7]. The precipitation of Si and Cd in the apoplast of roots and leaves is another important reason for the alleviation of Cd toxicity in plants [25]. Moreover, the exogenous application of Si can separate more Cd into vacuoles in roots [25] and restrict its upward movement [9,12], resulting in low Cd content in edible parts and high Cd content in roots (Figure 2b).

Rizwan et al. [35] and Yu et al. [36] revealed that Si addition reducing Cd concentration in plants could be due to changes in Cd speciation in the soil. Silicon application is beneficial for converting Cd into an inactive form by forming silicate complexes. This conclusion was supported by the negative correlation between Si and EX-Cd (high bioavailability) in CdSi₁ and CdSi₂ and the positive correlation between Si and RS-Cd (low bioavailability) in our study (Figure 4b). In addition, reduced Cd uptake by plants may be associated with an increase in soil pH, which ultimately reduces the Cd bioavailability [3].

4.2. Effects of Si on N Migration and Transformation in Soils

In the absence of Cd stress, although the addition of Si benefits the N uptake of plants [37], it seems to lead to more NH₄⁺-N leaching to deep soil and even underground water. Nutrients like N are mobilized from the soil by Si addition, because Si can compete with other elements for binding sites on organic matter and mineral surfaces [17]. Then, the mobilized N migrates downward with soil water during the flooding period, e.g., the distinctly higher NH₄⁺-N content at a depth of 35 cm in Si₁ and Si₂ compared with CK (Figure 7a). However, it is not the case that the more Si is applied, the more NH₄⁺-N is leached out, e.g., the NH₄⁺-N content at a depth of 35 cm was lower in Si₂ than in Si₁. An important property of silicic acid (H₄SiO₄) is that it can polymerize at high concentrations, forming Si-O-Si oligomer chains that have a higher binding affinity for N [38]. The lower NO₃⁻-N content at all depths in Si₁ and Si₂ compared with CK could be attributed to the promoting effect of Si addition on the plant absorption and leaching of NO₃⁻-N. However, the two processes are difficult to distinguish based only on the distribution of NO₃⁻-N in soils, as they occur simultaneously.

Under Cd-stress conditions, the competition between Cd²⁺ and NH₄⁺ for adsorption sites in the soil causes more NH₄⁺-N to release in the soil solution. These un-adsorbed NH₄⁺-N are available for plants, but also easily migrate downward with soil water, which means less NH₄⁺-N remains in upper soils. The downward-migrated NH₄⁺-N indeed increased the NH₄⁺-N content in deep soil, e.g., the slight increase in NH₄⁺-N content at a depth of 35 cm in CdSi₀. However, higher NH₄⁺-N content at all depths was observed in CdSi₀ compared with CK. In fact, with Cd inhibition on plant growth, less NH₄⁺-N is absorbed by plant roots. Moreover, the conversion of NH₄⁺-N to NO₃⁻-N is reduced by Cd because ammonium oxidizers are sensitive to heavy metals [39,40]. Both of the two processes could promote the retention of NH₄⁺-N in Cd-contaminated soils. As for NO₃⁻-N, the decrease in plant absorption increasing the leach out of NO₃⁻-N, as well as the suppressed conversion of NH₄⁺-N to NO₃⁻-N, jointly result in lower NO₃⁻-N content in CdSi₀ than in CK. All these influences of Cd on NH₄⁺-N and NO₃⁻-N are diminished by Si addition, thus promoting the adsorption and absorption of NH₄⁺-N and NO₃⁻-N in the upper soil layer and reducing its migration to the lower soil layers.

4.3. Effects of Si on the Interactions of Nitrogen between Different Pathways

After the qualitative analysis (Section 4.2), an N mass balance calculation was performed to quantitatively describe the N migration via different pathways (plant absorption and soil N flux). As shown in Table 2, variations in total N (NH₄⁺-N and NO₃⁻-N) in the soils were mainly attributed to plant-absorbed N and soil N flux. In most treatments, plant-absorbed N contributed predominantly to the soil total N variation; for example, the fraction of plant-absorbed N was 67.2% in CK during the entire experimental period. However, the fraction of plant-absorbed N varied with treatment and period; the fractions were in the following order: Si₂ > Si₁ > CK > CdSi₂ > CdSi₁ > CdSi₀. With the same soil background value of N and the same external N input (fertilization) in the soil–rice system, Si promotion on plant dry weight and N content resulted in the highest plant-absorbed N, and thus the highest fraction, and vice versa for Cd. The plant-absorbed N also exhibited strong heterogeneity over time. Most plant N absorption occurred during the flooding period for all treatments. For example, in the CK, N absorption during the flooding period accounted for 77.4% of the total N absorption. Studies have shown that plant N accumulation during the flooding period is higher than that during the non-flooding period [41]. The results of [42] also showed that more than 50% of the N fertilizer was absorbed by rice during the flooding period. In addition, some studies have concluded that water stress also decreases N accumulation in plants, especially during the non-flooding period [43].

During the flooding period, the soil N fluxes decreased in all treatments, except for the upward soil N flux in Si₂. The magnitude of the downward flux was in the following order: CdSi₀ > CdSi₁ > CdSi₂ > CK > Si₁, which was the opposite of the order of the amount of N absorbed by the plants. The soil N flux was determined to a great extent by the migration of NO₃⁻-N, because of the large magnitude and high mobility of NO₃⁻-N relative to NH₄⁺-N. Studies have reported that soil flooding can increase NO₃⁻-N migration, and more than 90% of NO₃⁻-N is carried to deeper soil [44]. Thus, the largest downward soil N flux in CdSi₀ indicates an increasing risk of NO₃⁻-N loss through deep percolation induced by Cd. In contrast, Si application caused a significant reduction in the downward N flux in soils with and without Cd. Notably, the opposite direction of soil N flux occurred in Si₁ and Si₂. Therefore, the direction of soil N flux should be downward in Si₁ and Si₂ when only the promotion of Si on leaching NO₃⁻-N in soils (described in Section 4.2) is considered. However, with an increase in Si application, the increasing amount of plant-absorbed N exceeded the amount of leached NO₃⁻-N and the upward soil N flux appeared in Si₂. Therefore, we conclude that there is a significant trade-off between plant-absorbed N and soil N flux. The more N that the plant absorbed, the less N migrated to the soil. Therefore, the magnitude and direction of the soil N flux were greatly affected by plant-absorbed N. Plants play a dominant role in N response to Si in the soil–rice system. During the non-flooding period, the direction of soil N fluxes changed upward; the magnitude decreased in all treatments with no significant difference between each other. The reason is that a small quantity of N absorption by plants still existed, and N migrated to the surface soil without water leaching under the influence of evapotranspiration.

5. Conclusions

The present study demonstrated that Cd negatively affected plant growth and N accumulation, leading to a decrease in the N absorption of plants. Moreover, the leaching of NO₃⁻-N, rather than NH₄⁺-N, in soils was increased by Cd. The application of Si significantly alleviated these negative effects by reducing Cd availability in soils and restricting the Cd uptake and translocation in plants, sequentially reducing the Cd concentration in different plant tissues, and ultimately leading to an increase in the N absorption of plants. Variations in total N (NH₄⁺-N and NO₃⁻-N) in the soils were mainly attributed to plant-absorbed N and soil N flux, of which plant-absorbed N accounted for the majority. The result of the nitrogen mass balance calculation showed that there was a significant trade-off between plant-absorbed N and soil N flux. The magnitude and direction of the soil N flux were greatly and negatively affected by plant-absorbed N. Therefore, we conclude that Si application could reduce the leaching of N in soil–rice systems under Cd stress, mainly due to the promotion of the N absorption of plants rather than N immobilization in soils. This study improves our understanding of the interaction between Cd, Si, and N in plants and soils when we use Si amendment to alleviate Cd toxicity in a soil-rice system with Cd contamination. We advocate supplementing Cd-contaminated soil with Si fertilizers during the flooding period as a practical approach to alleviate Cd toxicity and reduce percolation loss of N, which is significant for the sustainable development of agriculture. Further studies are needed to investigate the role of microbial communities in the N response to Si, while also considering the N loss occurring at the upper boundary, such as via N₂O emissions and ammonia volatilization.