Impacts of carbon nanomaterials on the diversity of microarthropods in turfgrass soil

Nanoscale materials have been produced with unprecedented speed due to their widespread use, and they may eventually be released into the environment. As effective adsorbents for heavy metals, carbon nanomaterials can be used to immobilize metals in contaminated soil, but little information is available regarding their effects on soil microarthropods. This study was designed to investigate the influence of three types of carbon nanomaterials, graphene (G), graphene oxide (GO) and carbon nanotubes (CNTs) on soil microarthropod communities under turfgrass growth conditions. The application of carbon nanomaterials resulted in increased abundance of all soil microarthropods, especially in the GO and CNT treatments. GO also significantly increased the abundances of multiple trophic functional groups, including predators, detritivores, herbivores and fungivores. Further, the dominant genera varied among the treatments. Herbivorous microarthropods predominated in the control, whereas predatory species predominated in the carbon nanomaterial treatments. Carbon nanomaterials also increased the total taxonomic richness, Shannon diversity index, and dominance index of the microarthropod community, but they decreased the evenness index. Higher diversity of soil microarthropods indicates an environment suitable for soil mesofauna and for enhanced decomposition and nutrient cycling in the soil food web.

community in a turfgrass system. We characterized the responses of soil microarthropods to carbon nanomaterials in order to provide a scientific basis for the application of carbon nanomaterials for remediating soil contaminated with heavy metals.

Results
The composition and abundance of soil microarthropods. In total, 24 genera of Acari and 8 genera of Collembola were recorded in this study ( Table 1). The identified mites belonged to three suborders (Oribatida, Prostigmata, and Mesostigmata). The application of carbon nanomaterials significantly increased the species richness of soil microarthropods. Only 6 genera of soil microarthropods were found in the control soil, whereas 17, 18 and 11 genera were observed in G-, CNT-and GO-treated soils, respectively. The dominant genera varied among the treatments. The relative abundance of each genus in the control was greater than 10%. The dominant genera in the G treatments were Palaeacarus (Oribatida), Spinibdella (Prostigmata) and genera belong to Tarsonemidae (Prostigmata), accounting for 47.1% of the individuals. In the GO treatments, Mahunkania and Cheylostigmaeus, belonging to Prostigmata, were the dominant genera and presented 17.6% and 34.3%, respectively, of the total individuals. In the CNT treatments, there were four dominant taxa that together accounted for 76.5% of the individuals: Brennandania 28.1%, Petalomium 15.6%, Trombidiidae 17.2% and Rhodacarellus 15.6%.
The abundance of total microarthropods was strongly affected by the treatments ( Table 1). The average abundance of total soil microarthropods for the GO, CNT and G treatments was 22.7, 7.1 and 5.9 times higher, respectively, compared to the control treatment. However, no significant differences were observed between the G treatment and the control. The highest total abundance of soil microarthropods was observed with GO treatment, in which it was significantly higher than the soil microarthropod abundances in the G and CNT treatments. Prostigmata, with the highest abundance, accounted for 55.6%, 49.1%, 69.1%, and 62.5% of the total soil microarthropods in the control, G, GO, and CNT treatments, respectively. Compared with the control and the other two treatments, GO significantly increased the abundances of Prostigmata and Collembola.
Soil microarthropod trophic groups. All of the collected soil microarthropods except for Neognathus were grouped into fungivorous, detritivorous, herbivorous and predatory guilds according to their feeding habits and the results for their abundances are shown in Fig. 1. The carbon nanomaterials increased the total abundance of each trophic group. Detritivorous microarthropods were not found in the control but were present in all of the carbon nanomaterial treatments. The GO treatment showed the most significant effects, and the abundances of predators, herbivores, and fungivores were increased by 51-fold, 10-fold and 16-fold, respectively, compared to the control. Herbivorous microarthropods dominated in the control, accounting for 44.4% of the total individuals, whereas predators dominated in the G, GO and CNT treatments, representing 37.8%, 52.6% and 68.8%, respectively, of all individuals (Supplementary Dataset 1).
Soil microarthropod community diversity. The diversity indices were significantly influenced by all the carbon nanomaterials (Table 2). Significant increases (P < 0.05) in the Shannon-Wiener index and dominance relative to control values were observed in all of the carbon nanomaterial treatments. Moreover, the GO and G treatments showed significantly greater species richness than did the CNT and control treatments. Although the GO treatment showed lower evenness, the evenness indices were not significantly different among the G, CNT and control treatments.

Discussion
Soil microarthropods are important components of the soil ecosystem. The species composition and abundance of soil microarthropods, the trophic functional groups, and the community indices can be used as sensitive indicators for evaluating ecosystem processes 4 . Organic fertilizers may promote increases in faunal populations by increasing the quality and quantity of food needed by soil microarthropods 5,16,17 . Greater microarthropod densities with increased soil fertility were also observed by Cole et al. 18 . However, Cao et al. 19 found that applications of chemical fertilizer decreased the number of soil mites, perhaps because of the direct toxicity of metal contaminants in the chemical fertilizers 20 . In the present study, the abundance of soil microarthropods significantly increased in the turfgrass plantation under the application of G, GO and CNT. This may be attributable to an increase in the supply of soil nutrients (i.e., soil C) through the addition of carbon nanomaterials. Previous  studies demonstrated that the growth of soil microarthropods abundance depends on increases in the soil C and N concentrations 5 . Both graphene and graphene oxide increased the diversity and richness of soil microarthropods, effects that were mainly attributable to the marked increases in Oribatida and Prostigmata. In a previous study, an increase in the diversity of soil fauna was correlated to an increase in the diversity of available food resources 21 . In general, application of a low level of fertilizer tends to result in increased faunal diversity due to its positive effect on detrital inputs 5,17 . However, graphene oxide showed the most significant effect on the abundance and diversity of soil microarthropods. Previous studies demonstrated that GO can enhance cell adhesion and proliferation and promote biological growth and reproduction, thereby affecting the populations and diversity of total microarthropods 22 .

Dominance index (C)
The function of soil and the maintenance of soil quality are closely linked to the micro-fauna community composition; therefore, understanding the responses of soil microarthropod functional groups is useful for revealing the ecological effects of carbon nanomaterials 23 . Based on the presence of Oribatida in soils treated with carbon nanomaterials, these soils tend to contain nearly all functional groups of microarthropods 24 . In the present study, Oribatida was mainly represented by detritivores, which are sensitive to the soil environments 25 . The greater abundances of Oribatida in carbon nanomaterial treatments suggest that the nutrient level and soil quality were improved. Furthermore, detritivorous Oribatida are involved in organic matter decomposition, nutrient cycling and soil formation in the soil ecosystem 1,25 . In the GO treatment, the number of microarthropods in predatory, herbivorous and fungivorous groups increased at a significance level of 51-fold, 10-fold and 16-fold, respectively, in comparison with the control, whereas no significant differences in abundance of these trophic groups were found among the control, G and CNT treatments. This result indicates that different trophic communities have different adaptive mechanisms to the treatments. Predatory microarthropods were found to be dominant in carbon nanomaterial-treated soils, whereas herbivorous species were dominant in the control treatment. The increase in predatory abundance was likely caused by an increase in the availability of prey species, e.g., Enchytraeidae and nematodes 26 . Nematodes are a preferred food source for most predatory species 27,28 . As suggested in previous studies, predators may more directly regulate the abundances of detritivorous and fungivorous microarthropods by limiting their population expansion in the natural soil 29,30 . Further studies should be conducted to identify the mechanisms by which carbon nanomaterials affect soil microarthropods, perhaps through effects on the soil properties.
In conclusion, the application of carbon nanomaterials (G, GO and CNT) in turfgrass soil significantly increased the abundance and diversity of soil microarthropods. The GO treatment produced significant increases in the abundances of four trophic microarthropods. Predatory arthropods predominated in the carbon nanomaterial treatments, whereas herbivorous species predominated in the control. The increased microarthropod populations support enhanced decomposition and nutrient cycling in the soil food web by strengthening both bottom-up and top-down processes.

Materials and Methods
Experimental design. The tested soil was gathered from the top 20 cm of an experimental site at the Tianjin Normal University campus. It is a sandy loam with a pH (in water) of 7.3, water content of 19.4%, conductivity of 2250 μS cm −1 , and organic matter, total phosphorus, and total nitrogen contents of 52.3, 3.75 and 2.15 g kg −1 , respectively. Festuca arundinacea Schreb. was selected as the tested turfgrass. Municipal solid waste (MSW) compost was obtained from Xiaodian composting plants in Tianjin, China. The properties and origins of the carbon nanomaterials, graphene (G), graphene oxide (GO) and carbon nanotubes (CNTs), are reported in Table 3.
For this study 1500 g of tested soil, 50 g of MSW compost and 15 g of carbon nanomaterials were mixed thoroughly and transferred to plastic pots (height 15 cm; inner diameter 20 cm). Treatments without the nanomaterials were used as the controls, and each treatment was replicated three times. Then, 5 g of seeds of Festuca arundinacea was sown in each pot. Cultivation was performed in a greenhouse under natural light conditions (646-27090 LX). During the experiment, the night and day average temperatures were 15 °C and 26 °C, respectively, and the relative humidity was between 35% and 55%. Water was supplied daily to maintain adequate substrate moisture for turfgrass growth. Turfgrasses were mowed once, on day 65.
Characterization of soil microarthropods. Soil samples were taken from the experimental pots on day 130. Soil microarthropods were extracted from 100 g of fresh soil using the Tullgren method 31 and then preserved in 70% alcohol. All extracted faunal samples were sorted and counted under a dissection microscope and  identified to the genus level 32 . The soil microarthropods were grouped into predatory, herbivorous, detritivorous and fungivorous guilds 3, 6, 19, 33-38 . Statistical analysis. The Shannon-Wiener index (H′), Margalef richness index (SR), Pielou evenness index (J) and Simpson dominance index (C) were used to describe the diversity of the soil microarthropod community 39 . H′ = −∑PilnPi; SR = (S − 1)/lnN; J = H′/lnS; C = 1 − ∑(Pi) 2 , where S is the number of species, N is the total number of individuals and Pi = Ni/N is the ratio between the individual number in a genus and the total number of individuals. The responses of the microarthropod community structure to carbon nanomaterials were examined by using analysis of variance (ANOVA). Differences between means were evaluated by using the least significant difference (LSD) test. All statistical tests were conducted at a significance level of p < 0.05 using the SPSS 17.0 software package (SPSS Inc., Chicago, USA). Data presented in the Tables and Figure are means ± standard deviations (SD) of three replicates for each treatment. SigmaPlot 12.5 was used to plot the graph 40 .