MARINA JUSTI AGRONOMIC VALUE OF METAL COMPLEXES WITH ORGANIC ACIDS AND NUTRITIONAL EFFECTS OF IRON SOURCES ON PLANTS LAVRAS – MG 2021 MARINA JUSTI AGRONOMIC VALUE OF METAL COMPLEXES WITH ORGANIC ACIDS AND NUTRITIONAL EFFECTS OF IRON SOURCES ON PLANTS VALOR AGRONÔMICO DE COMPLEXOS DE METAIS E ÁCIDOS ORGÂNICOS E EFEITOS NUTRICIONAIS DE FONTES DE FERRO EM PLANTAS Tese apresentada à Universidade Federal de Lavras, como parte das exigências do Programa de Pós-Graduação em Ciência do Solo, área de concentração em Fertilidade do Solo e Nutrição de Plantas, para obtenção do título de Doutor. Prof. Dr. Carlos Alberto Silva Orientador LAVRAS – MG 2021 Ficha catalográfica elaborada pelo Sistema de Geração de Ficha Catalográfica da Biblioteca Universitária da UFLA, com dados informados pelo(a) próprio(a) autor(a). Justi, Marina. Agronomic value of metal complexes with organic acids and nutritional effects of iron sources in plants / Marina Justi. - 2021. 100 p. Orientador(a): Carlos Alberto Silva. Tese (doutorado) - Universidade Federal de Lavras, 2021. Bibliografia. 1. chelates. 2. metal complexation. 3. iron nutrition. I. Silva, Carlos Alberto. II. Título. O conteúdo desta obra é de responsabilidade do(a) autor(a) e de seu orientador(a). MARINA JUSTI AGRONOMIC VALUE OF METAL COMPLEXES WITH ORGANIC ACIDS AND NUTRITIONAL EFFECTS OF IRON SOURCES ON PLANTS Tese apresentada à Universidade Federal de Lavras, como parte das exigências do Programa de Pós-Graduação em Ciência do Solo, área de concentração em Fertilidade do Solo e Nutrição de Plantas, para obtenção do título de Doutor Aprovada em 21 de dezembro de 2020. Dr. Andrés Calderín García UFRRJ Dr. Marcos Komogawa ESALQ/USP Dr. Marcelo Eduardo Alves ESALQ/USP Dr. Guilherme Lopes UFLA Prof. Dr. Carlos Alberto Silva Orientador LAVRAS-MG 2021 Aos meus pais, Paulo e Maria Emília, pelo amor, apoio e ensinamentos sobre persistência, dedicação e felicidade; aos meus irmãos Joseane e Diego, eternos confidentes e parceiros de jornada Dedico AGRADECIMENTOS Agradeço a Deus pela iluminação diante dos desafios, pelas oportunidades e conquistas alcançadas durante a vida. Agradeço a minha família por todo apoio incondicional e companheirismo. À Universidade Federal de Lavras (UFLA) e ao programa de pós-graduação do Departamento de Ciência do Solo, pela oportunidade concedida para realização do doutorado. À Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES), pela concessão da bolsa de doutorado, e financiamento das ações de pesquisa (CAPES-PROEX-AUXPE 593-2018); ao Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq, processo 307447- 2019-7) e Fundação de Amparo À Pesquisa de Minas Gerais (FAPEMIG) pelo financiamento das ações de pesquisa. Ao professor orientador Carlos Alberto Silva pela orientação, ensinamentos e compreensão mediante desafios. Ao professor Matheus Puggina de Freitas e ao Josué Silla, por toda a colaboração, apoio, ensinamentos e paciência me ajudando com a química teórica e computacional. Ao DQI-UFLA por ceder a estrutura para os cálculos teóricos computacionais. Ao professor Cleiton Antonio Nunes pela ajuda com os estudos de quimiometria. À professora Dra. Maria Ligia de Souza Silva por permitir o uso da estrutura de hidroponia visando a condução dos experimentos com plantas em casa de vegetação do Setor de Nutrição Mineral de Plantas do DCS-UFLA. Àqueles que foram minha segunda família em Lavras, companheiros de casa queridos que tornaram a minha vida alegre Jakeline, Pedro, Jéssica, Sibely, Matheus, Lívia, Monna Lysa e Mariane. Muito obrigada por todos os bons momentos e a amizade que ficarão para sempre na minha memória. Ao Bruno, por ser parceiro, paciente e amoroso me ajudando sempre que possível e me apoiando incondicionalmente nessa jornada. Aos meus amigos e colegas de trabalho do LEMOS, Sara, por ser minha companheira e confidente, por toda ajuda em experimentos e laboratório, Everton, por todo auxílio, experiência, amizade e conversas, Henrique e Pedro por toda ajuda, Rimena, Bruno e Murilo pela companhia e amizade. Todos vocês tornaram o trabalho muito mais leve. Às minhas queridas Cynthia e Geila, por toda amizade e momentos maravilhosos, minha estada em Lavras não seria a mesma sem vocês. Ao Osvaldo e ao Dr. Roberto que me acompanharam durante toda minha vida acadêmica, me orientando, me ajudando em todos os desafios e me ensinando a alcançar a saúde mental. Aos funcionários do Departamento de Ciência do Solo, em especial, Dirce, Mariene, Lívia, Bethânia e Roberto, muito obrigada por tudo. A todos aqueles que passaram pela minha vida neste período e contribuíram de alguma forma para minha vida acadêmica Obrigada! “A persistência é o caminho do êxito” Charles Chaplin ABSTRACT Complexing agents are largely used for agricultural and environmental applications. For years, synthetic aminopolycarboxilic acids, such as EDTA, have been the most common complexing agents used for these applications. However, aminopolycarboxylic acids are expensive and hardly degraded in the environment. Thus, the investigation of natural molecules that can act as complexing agents is essential for sustainable applications. This study aimed at understanding the chemical properties of complexes formed between several metals and low- molecular-weight organic acids, and testing some of these complexes for agricultural applications. Complexes were synthesized at pH 7 using metals Fe, Mn, and Zn, as well as citric, tartaric, malic, and oxalic acids as complexing agents. The 1:1 and 1:2 stoichiometric ratios of reaction (SR; metal: ligand molar ratio) were tested. The chemical species formed were determined through VISUAL Minteq software. Molecular structure and stability were determined through theoretical calculations, and the complexes were submitted to the Fourier Transformed Infrared Spectra (FTIR analysis). The chemical characterization showed that complexed metal fraction, stability, and solubility depended on reaction stoichiometry, metal type, and complexing agent type. Oxalic acid was the smallest molecule used, and its complexes were predominantly of the 1:2 stoichiometry, presenting very low solubility and high stability. Citric and malic acid formed highly soluble complexes. Malic acid only formed 1:1 stoichiometry complexes with all metals and both stoichiometries tested. Citric and tartaric acid tended to form 1:2 complexes with Zn and 1:1 complexes with Mn and Fe. Considering the metals tested, the Fe complexes were more stable than Mn and Zn complexes. The chemometric partial least square technique was used to test the possibility to predict complex properties from spectra. Solubility and complexed-metal fraction were adequately predicted from FTIR spectra. The use of iron complexes as Fe sources in foliar application to maize and nutrient solution addition to soybean and maize was tested in greenhouse experiments. Foliar application of Fe- tartrate and Fe-citrate at 1:2 SR promoted similar plant growth and better Fe nutrition than the application of Fe-EDTA and FeSO4. When applied in the nutrient solution, iron-organic acid complexes kept less iron in the complexed form than EDTA. The effectiveness of complexes in promoting Fe accumulation and plant growth depended on plant species. Despite showing low complexation capacity, the Fe-tartrate at 1:2 SR promoted suitable Fe nutrition in maize plants. As for soybean, Fe-oxalate complexes promoted growth and Fe accumulation in plants similarly to that of the Fe-EDTA treatment. Keywords: chelates. Visual Minteq. computational modeling. foliar fertilization. iron nutrition. metal complexation. molecular structure. stability. RESUMO Agentes complexantes são amplamente usados para aplicações agronômicas e ambientais. Durante anos os ácidos aminopolicarboxílicos sintéticos foram os agentes complexantes mais utilizados para tais aplicações. Entretanto, tais moléculas tem elevado custo e baixa degradabilidade no ambiente. Portanto, a investigação de moléculas naturais que possam agir como agentes complexantes é essencial no âmbito da aplicação sustentável. Este estudo objetivou investigar as propriedades químicas de complexos formados entre vários metais e ácidos orgânicos de baixa massa molecular, bem como testar alguns desses complexos para aplicação agronômica. Os complexos foram sintetizados utilizando os metais Fe, Mn, Zn e os ácidos cítrico, tartárico, málico e oxálico como agentes complexantes. Duas estequiometrias de reação, 1:1 e 1:2 (SR, metal: ligante em proporção molar), foram testadas. As espécies químicas formadas foram determinadas por meio do software VISUAL Minteq. A estrutura molecular e a estabilidade dos complexos foram determinadas por meio de cálculos teóricos, e os complexos foram submetidos análise de infravermelho com transformada de Fourier (FTIR). A caracterização química mostrou que a fração de metal complexado, a estabilidade e solubilidade dependeram da estequiometria de reação testada, do metal e do agente complexante utilizados. O ácido oxálico, o menor ligante utilizado, apresentou predominantemente complexos com estequiometria 1:2 com baixa solubilidade e alta estabilidade. O ácido cítrico e ácido málico formaram complexos altamente solúveis. O ácido málico apresentou a formação de complexes 1:1 com todos os metais e estequiometrias testadas. O ácido tartárico e o ácido cítrico tendem a formar complexos do tipo 1:2 com Zn e complexos do tipo 1:1 com Fe e Mn. Considerando os metais testados, os complexos de Fe foram mais estáveis que os complexos de Zn e Mn. A técnica quimiométrica de regressão de mínimos quadrados foi usada para verificar a possibilidade de predizer as características químicas dos complexos por meio do espectro de infravermelho. A solubilidade e fração de metal complexada foram adequadamente preditos pelo espectro de infravermelho. O uso de complexos de Fe como fonte do nutriente em aplicação foliar e em solução nutritiva foi testada em condições de casa de vegetação. A aplicação foliar de Fe-tartarato e Fe-citrato na estequiometria 1:2 promoveu crescimento vegetal similar ao Fe-EDTA e ao FeSO4. Quando aplicados em solução nutritiva, os complexos de ácidos orgânicos com Fe mantiveram menos Fe na forma complexada do que o EDTA. A efetividade dos complexos em promover o acumulo de ferro e crescimento vegetal dependeu da espécie testada. Embora tendo menor capacidade complexante, o uso do Fe-tartarato na estequiometria 1:2 promoveu crescimento e acúmulo de Fe adequados nas plantas de milho. No caso da soja, o uso de Fe-oxalato promoveu crescimento e acúmulo de Fe similar ao uso de Fe- EDTA. Keywords: quelatos. Visual Minteq. modelagem molecular. fertilização foliar. nutrição em Fe. complexação de metais. estrutura molecular. estabilidade SUMÁRIO SECTION 1 ............................................................................................................... 11 General Introduction ............................................................................................. 11 Objectives .............................................................................................................. 14 References .............................................................................................................. 15 SECTION 2 - ARTICLES ........................................................................................ 17 ARTICLE 1 - Experimental FTIR and theoretical investigation of chemical properties and molecular structure of metal-carboxylate complexes ................................... 17 Abstract ................................................................................................................. 18 1. Introduction........................................................................................................ 19 2. Material and Methods ....................................................................................... 21 3. Results and Discussion....................................................................................... 24 4. Conclusions ........................................................................................................ 38 5. References .......................................................................................................... 38 ARTICLE 2 - Influence of complex stability on iron accumulation and redistribution for foliar applied iron-organic acid complexes ..................................................... 47 Abstract ................................................................................................................. 47 1. Introduction........................................................................................................ 48 2. Material and Methods ........................................................................................ 49 3. Results ................................................................................................................ 53 4. Discussion ........................................................................................................... 59 5. References .......................................................................................................... 64 Supporting Information ........................................................................................ 68 ARTICLE 3 - Organic acids as complexing agents for iron and their effects on nutrition and growth of maize and soybean ......................................................... 77 Abstract ................................................................................................................. 77 1. Introduction........................................................................................................ 78 2. Material and Methods ........................................................................................ 80 3. Results ................................................................................................................ 82 4. Discussion ........................................................................................................... 91 5. Conclusions ......................................................................................................... 95 6. References .......................................................................................................... 95 FINAL REMARKS ................................................................................................... 99 11 SECTION 1 General Introduction Complexing agents are molecules with one or more chemical groups able to donor electrons. Metals in general have a high affinity for accepting electrons. Complexing agents and metals reaction give rise to a molecule named coordination compound or complex (HOSMANE, 2017). If the complexing agent has two or more groups capable of donor electrons, it is also called chelating agent (HOSMANE, 2017). In the final molecular structure of the complex or chelate, the metal cation occupies the central position, and the complexing agent stays around it, a structural conformation that promotes high stability to the molecule formed (KOLODYNSKA, 2020). Complexing agents are currently used in the health, industrial, environmental, and agricultural domains for several purposes (CLEMENS et al., 1990; KALIA, 2005; RODRÍGUEZ-LUCENA et al., 2010). In agriculture, complexing agents are used to keeping micronutrients in soluble forms and avoid undesirable reactions such as precipitation (e.g., the reaction of iron with phosphate) (ABADÍA et al., 2011). Complexing agents are also used to increase the efficiency of plant uptake and translocation of nutrients (ÁLVAREZ- FERNÁNDEZ et al., 2004; RODRÍGUEZ-LUCENA et al., 2010). Complexing agents are also used to increase desorption of metals from soil and sediments, increasing soluble metal forms, and improving their uptake by plants (GUO et al., 2018). Additionally, complexing agents can be used to reduce metal bioavailability, fixing up the element in complexes of high stability and low solubility (SU et al., 2015). For several years, synthetic aminopolicarboxylic acids, such as EDTA and DTPA, were the most frequently chelating agents used in the agricultural and environmental fields (PINTO et al., 2014). However, these chelating agents have a high cost and are hardly degraded in the environment (NOWACK and VANBRIESEN, 2005). EDTA and its metal complexes can remain for years in soils and sediments, altering the metal equilibria (NOWACK et al., 2006). EDTA was found in significant amounts in subsurface waters and some metal-EDTA chelates like Cu-EDTA are hardly degraded by living microbial community in soils (NOWACK et al., 2006), thus, representing a threat to soil fauna (DUO et al., 2019). Furthermore, the high cost of aminopolycarboxylic acids limits their application for valuable crops (ABADÍA et al., 2011). Therefore, the use of biodegradable, environmentally friendly, and low-cost complexing agents is highly desirable. 12 Low molecular organic acids and humic substances have carboxylic and phenolic groups that can complex metals (PINTO et al., 2014). Low molecular weight organic acids such as citric acid, malic, and oxalic acid were already successfully used as complexing agents in the remediation process of contaminated soils, increasing metals desorption and promoting high metal bioaccumulation by plants (EVANGELOU et al., 2007). Furthermore, the use of citric acid as a complexing agent for iron in the foliar spray effectively promoted iron nutrition in plants, with a similar agronomic performance of Fe-EDTA (CHAKRABORTY et al., 2014). Several chemical aspects of the organometallic complexes can affect their applicability and effectiveness in supplying nutrients to plants (CARRASCO et al., 2012; Fig. 1). Although some studies showed the efficacy of natural organic molecules as complexing agents, several chemical aspects of the complexes formed still lack to be clarified. The complexes can differ significantly concerning their solubility, complexation capacity, stability, and molecular structure (URRUTIA et al., 2014; GARCIA-MINA, 2006). Molecular structure and spatial conformation of the complex molecule can affect their metabolism and uptake by the soil organisms (JOSHI-TOPE and FRANCIS, 1995), as well as their capacity to interact with mineral surfaces in soil and sediments (TANG et al., 2017). Stability can affect the process of metal displacement from complexes by plants and microorganisms, and if exchange reactions with other cations will occur in the environment (GARCIA-MINA, 2006). The metal complexation capacity is crucial to define the complex agent effectiveness in keeping most of the metal in complexed and soluble forms in soil or nutrient solution (ZANIN et al., 2019; GARCÍA-MINA et al., 2004). Complexation capacity depends on the affinity of the metal for chemical groups of the ligand, and, therefore, can be affected by the metal ion that is supposed to be complexed, by the other ions present in the media, and pH and stoichiometry of reaction as well (GARCIA-MINA, 2006). Solubility of the complex should be high if the intended application is increase the bioavailability of the metal (GARCÍA-MINA et al. 2004), and can be low if the goal is to keep the metal in stable and low soluble forms, reducing its toxicity (SU et al., 2015) Although low molecular weight organic acids can be used for several environmental and agricultural purposes, some chemical properties of their complexes with metals are still not clear. It is important to investigate the stoichiometry of reaction to be used and the stability and molecular conformation of the complexed metal. The same ligand can result in complexes with different stabilities depending on the metal, pH, and stoichiometry used (CHU et al., 2011). Considering the number of organic molecules that can be used in the complexes synthesis and the myriad of conditions that these complexes can be applied, it is necessary to elucidate how 13 factors such as pH, stoichiometry, ligand, and metal affect stability, complexation capacity, and solubility of the complexes formed. Regarding the agricultural use, the effectiveness of a molecule to act as complexing agent for a nutrient also depends on the genetic and physiological characteristics of the plant species (GARCÍA-MINA et al., 2004). Differences in the root architecture, Kasparian strip thickness, and abundance of root transporters can also affect the metal complex absorption (LEŠTAN et al., 2008; Fig. 1). On the leaf, the cuticle morphology and stomata abundance can affect the absorption (ABADÍA et al., 2011; Fig. 1). Plants can also exhibit different strategies to acquire nutrients with different oxidation states, such as iron. Monocotyledonous species release complexing substances (phytosiderophores) that allow them to uptake iron in oxidized forms (KOBAYASHI and NISHIZAWA, 2012). Dicotyledonous species release substances that promote reduction of iron before its acquisition by roots (KOBAYASHI and NISHIZAWA, 2012). Therefore, iron sources probably exhibit different effectiveness in nourishing monocots and dicots species. Additionally, the same micronutrient complex can have different effectiveness when applied via nutrient solution (root uptake) or via foliar spray (leaf penetration). Thus, a molecule used for nutrient solution application cannot be suitable for foliar application (RODRÍGUEZ-LUCENA et al., 2010). Therefore, the growing of plant tests submitted to different application modes of complexed nutrients is essential to define their agronomic value according to the different fertilization management strategies. Furthermore, organic molecules such as humic substances and low molecular weight organic acids play a role as biostimulants, affecting plant physiology and growth (ZANIN et al., 2019). The effects of these molecules on the physiological process, such as respiration and photosynthesis, can vary considerably according to the plant species and used molecule concentration (CONSELVAN et al., 2017). Organic acids are recognized by the ability to reduce abiotic stress effects in the plant physiology, replenishing of Krebs Cycle in nutrient starvation conditions, improvement of antioxidant system in plants, and by increasing metal uptake rate in plant cells (CARL-HUBER et al. 2016). Thus, the use of these molecules as complexing agents can promote effects on plants beyond plant nutrition. In this sense, this study intended to assess several chemical attributes that affect the applicability of complexes formed between Fe, Zn, and Mn, and citric, malic, tartaric, and oxalic low molecular weight organic acids. Furthermore, we conducted experiments with soybean and maize in greenhouse conditions to determine the agronomic value of Fe-organic acid complexes applied to nutrient solution and via foliar spray in promoting adequate plant growth and iron nutrition. 14 Figure 1 - Factors affecting the dynamic of complexed nutrients in plants. Objectives The aims of this study were: i) to determine the thermodynamic stability, molecular structure, metal complexed fraction, and FTIR signature of the complexes formed between organic acids (citric, malic, tartaric, and oxalic acids) and metals (Mn, Fe, and Zn) at two different stoichiometries of reaction; ii) to test the effectiveness of foliar-applied iron-organic acids complex in promoting adequate iron nutrition and plant growth, and to verify if the stability of the complex influences iron accumulation and redistribution in plants; iii) to assess the effectiveness of iron-organic acid complexes applied to nutrient solution in promoting adequate iron nutrition and growth of maize and soybean plants. This study was structured in three chapters, hereafter defined as articles, with the following titles: Article 1 – “Conformational structure and fast identification of chemical properties of metal-carboxylate complexes by FTIR and partial least square regression”; Article 2 – “Influence of Complex Stability on Iron Accumulation and Redistribution for Foliar- Applied Iron-Organic Acid Complexes in Maize”; and Article 3 – “Organic acids as complexing agents for iron and their effects on nutrition and growth of maize and soybean”. 15 References ABADÍA, J.; VÁZQUEZ, S.; RELLÁN-ÁLVAREZ, R.; EL-JENDOUBI, H. et al. Towards a knowledge-based correction of iron chlorosis. Plant Physiology and Biochemistry, 49, n. 5, p. 471- 482, 2011. ÁLVAREZ-FERNÁNDEZ, A.; GARCÍA-LAVIÑA, P.; FIDALGO, C.; ABADÍA, J. et al. Foliar fertilization to control iron chlorosis in pear (Pyrus communis L.) trees. Plant and Soil, 263, n. 1, p. 5- 15, 2004. CARRASCO, J.; KOVÁCS, K.; CZECH, V. r.; FODOR, F. et al. Influence of pH, iron source, and Fe/ligand ratio on iron speciation in lignosulfonate complexes studied using Mössbauer spectroscopy. Implications on their fertilizer properties. Journal of agricultural and food chemistry, 60, n. 13, p. 3331-3340, 2012. CHU, C.; DARLING, K.; NETUSIL, R.; DOYLE, R. P. et al. Synthesis and structure of a lead(II)– citrate: {Na(H2O)3}[Pb5(C6H5O7)3(C6H6O7)(H2O)3]·9.5H2O. Inorganica Chimica Acta, 378, n. 1, p. 186-193, 2011. CLEMENS, D. F.; WHITEHURST, B. M.; WHITEHURST, G. B. Chelates in agriculture. Fertilizer Research, 25, n. 2, p. 127-131, 1990. CONSELVAN, G. B.; PIZZEGHELLO, D.; FRANCIOSO, O.; DI FOGGIA, M. et al. Biostimulant activity of humic substances extracted from leonardites. Plant and Soil, 420, n. 1, p. 119-134, 2017. DUO, L.; YIN, L.; ZHANG, C.; ZHAO, S. Ecotoxicological responses of the earthworm Eisenia fetida to EDTA addition under turfgrass growing conditions. Chemosphere, 220, p. 56-60, 2019. EVANGELOU, M. W. H.; EBEL, M.; SCHAEFFER, A. Chelate assisted phytoextraction of heavy metals from soil. Effect, mechanism, toxicity, and fate of chelating agents. Chemosphere, 68, n. 6, p. 989-1003, 2007. GARCIA-MINA, J. M. Stability, solubility and maximum metal binding capacity in metal–humic complexes involving humic substances extracted from peat and organic compost. Organic Geochemistry, 37, n. 12, p. 1960-1972, 2006. GARCÍA-MINA, J. M.; ANTOLÍN, M. C.; SANCHEZ-DIAZ, M. Metal-humic complexes and plant micronutrient uptake: a study based on different plant species cultivated in diverse soil types. Plant and Soil, 258, n. 1, p. 57-68, 2004. GUO, X.; ZHAO, G.; ZHANG, G.; HE, Q. et al. Effect of mixed chelators of EDTA, GLDA, and citric acid on bioavailability of residual heavy metals in soils and soil properties. Chemosphere, 209, p. 776-782, 2018. HOSMANE, N. S. Chapter 5 - Ligands and d-Block Metal Complexes. In: HOSMANE, N. S. (Ed.). Advanced Inorganic Chemistry. Boston: Academic Press, 2017. p. 75-87. 16 JOSHI-TOPE, G.; FRANCIS, A. J. Mechanisms of biodegradation of metal-citrate complexes by Pseudomonas fluorescens. Journal of Bacteriology, 177, n. 8, p. 1989-1993, 1995. KALIA, K.; FLORA, S. J. S. Strategies for Safe and Effective Therapeutic Measures for Chronic Arsenic and Lead Poisoning. Journal of Occupational Health, 47, n. 1, p. 1-21, 2005. KOBAYASHI T.; NISHIZAWA N. K. 2012. Iron Uptake, Translocation, and Regulation in Higher Plants. Annual Reviews in Plant Biology, 63, n. 1, p.131–152. LEŠTAN, D.; LUO, C.-l.; LI, X.-d. The use of chelating agents in the remediation of metal- contaminated soils: A review. Environmental Pollution, 153, n. 1, p. 3-13, 2008. NOWACK, B.; SCHULIN, R.; ROBINSON, B. H. Critical Assessment of Chelant-Enhanced Metal Phytoextraction. Environmental Science & Technology, 40, n. 17, p. 5225-5232, 2006. NOWACK, B.; VANBRIESEN, J. M. Chelating Agents in the Environment. In: Biogeochemistry of Chelating Agents: American Chemical Society, 2005. v. 910, cap. 1, p. 1-18. (ACS Symposium Series). PINTO, I. S. S.; NETO, I. F. F.; SOARES, H. M. V. M. Biodegradable chelating agents for industrial, domestic, and agricultural applications—a review. Environmental Science and Pollution Research, 21, n. 20, p. 11893-11906, 2014. RODRÍGUEZ-LUCENA, P.; HERNÁNDEZ-APAOLAZA, L.; LUCENA, J. J. Comparison of iron chelates and complexes supplied as foliar sprays and in nutrient solution to correct iron chlorosis of soybean. Journal of Plant Nutrition and Soil Science, 173, n. 1, p. 120-126, 2010. SU, X.; ZHU, J.; FU, Q.; ZUO, J. et al. Immobilization of lead in anthropogenic contaminated soils using phosphates with/without oxalic acid. Journal of Environmental Sciences, 28, p. 64-73, 2015. TANG, Q.; ZHOU, T.; GU, F.; WANG, Y. et al. Removal of Cd(II) and Pb(II) from soil through desorption using citric acid: Kinetic and equilibrium studies. Journal of Central South University, 24, n. 9, p. 1941-1952, 2017. URRUTIA, O.; ERRO, J.; GUARDADO, I.; SAN FRANCISCO, S. et al. Physico-chemical characterization of humic-metal-phosphate complexes and their potential application to the manufacture of new types of phosphate-based fertilizers. Journal of Plant Nutrition and Soil Science, 177, n. 2, p. 128-136, 2014. ZANIN, L.; TOMASI, N.; CESCO, S.; VARANINI, Z. et al. Humic Substances Contribute to Plant Iron Nutrition Acting as Chelators and Biostimulants. Frontiers in Plant Science, 10, n. 675, 2019. 17 SECTION 2 - ARTICLES ARTICLE 1 - Experimental FTIR and theoretical investigation of chemical properties and molecular structure of metal-carboxylate complexes Marina Justi*, Matheus P. Freitas, Josué M. Silla, Cleiton Antônio Nunes, Carlos Alberto Silva Department of Soil Science, Federal University of Lavras, P.O. Box 3037, 37200-900, Lavras, MG, Brazil Department of Chemistry, Federal University of Lavras, P.O. Box 3037, 37200-900, Lavras, MG, Brazil E-mail: Marina Justi* - marina.justi@gmail.com * Corresponding author This article was prepared in line with the guidelines of the ‘Journal of Molecular Structure’ Highlights Molecular arrangement of metal-carboxylate complexes was unraveled FTIR fingerprints of metal-carboxylate complexes are shown and discussed Coordination modes of metals to carboxylate group can be estimated based on the infrared COO- band shifts Complexed fraction and solubility of metal complexes can be estimated from the combined use of FTIR spectra-partial least squares regression models 18 Abstract Several chemical properties of complexes should be considered to base their suitable applications in agronomic and environmental domains. However, the determination of these characteristics is generally expensive and time consuming. Therefore, the prediction of properties of metal-complexes using a simple and expeditious technique, such as infrared spectroscopy, is highly desirable. This study aimed to investigate the molecular structure and chemical properties of metal complexes, such as stability, metal complexed fraction, and solubility. The complexes were submitted to experimental FTIR analysis and the spectra dataset were regressed against the metal-complexes properties aforementioned using partial least squares (PLS) to obtain a predictive model. Low molecular weight carboxylic acids (citric acid, malic acid, tartaric acid, oxalic acid) were used as complexing agents for three metals (Mn, Zn and Fe) in two stoichiometric ratios of reaction (1:1 and 1:2 metal: ligand molar ratio). Solubility and complexation ratio appeared to be dependent on the ligand and stoichiometry. Citrate and oxalate have formed the most stable complexes with the tested metals. Metal- oxalates tend to form 1:2 complexes at the pH and stoichiometries tested, whereas citrate, malate and tartrate tend to form 1:1 complexes. All the complexes exhibited monodentate coordination of carboxylic groups with metals. Overall, predictive models were built to estimate the solubility and the ratio of complexed metals. Furthermore, the coordination mode of ligands to metals could also be assigned from infrared band shifts of the carboxylate groups. This work revealed crucial chemical properties to base the use of metal-complexes, and show that the fast and non-destructive FTIR technique can be used for predict these properties. Keywords: Partial least square regression, stability, monodentate coordination, stereochemistry 19 1. Introduction Complexing agents are extensively applied for agricultural and environmental purposes [1–3]. In agriculture, they are commonly used to increase the bioavailability of nutrients, as well as to reduce the reactivity of these nutrients with soil colloids and other chemical species present in soil or nutrient solution [2,4]. In the environmental domain, complexing agents are commonly used for metal remediation processes, to desorb metals of environmental matrices, such as soil and sediments, and increase the metal bioavailability and efficiency of phytoextraction [1]. They can also be used for remediation processes that involve stabilization, and fixation of metals in stable and low soluble forms to prevent leaching of metals into the ground water[5]. In this case, the main goal is to make the metal less available for plants and microorganisms, reducing metal activity in soil solution [5,6]. Synthetic aminopolycarboxylic acids (such as EDTA and DTPA) have been used as the main chelating agents for agricultural and environmental applications for several years [1,7,8]. However, the low biodegradability of these molecules limits their use for large-scale field applications [8,9]. Therefore, the use of natural and biodegradable complexing agents, such as low-molecular-weight organic acids, is a more environmentally friendly option [1,10,11]. Organic acids (e.g., citric acid, malic acid, tartaric acid, and oxalic acid) are naturally found in soil and plant system and play several roles as complexing agents [12]. The release of these molecules by plants in the rhizosphere is a recognized mechanism to improve metal uptake and accumulation [13]. Although the organic acids exhibit a lower metal complexing ability than aminopolycarboxylic acids, they have been used successfully for environmental [14,15] and agricultural applications [16,17]. Several chemical conditions can control the formation of complexes in the environment, such as pH, metal type, and stoichiometry of reaction [18,19]. Furthermore, there are chemical properties of complexes that need to be understood to improve their performance in different applications and environments, such as their molecular conformation, stability, and solubility. It was found that the ability of soil bacteria to capture and metabolize metal complexes is associated with their stability and stereochemistry [20]. Several coordination modes in complexes involving the carboxylate group and metals are recognized, such as monodentate, bidentate and chelating, and these chemical bonds can affect the bioactivity of complexes [21]. Stability and solubility are also key factors influencing the availability of metal complexed for plant acquisition [4,22]. Stability is an index to infer the complexing agent and metal affinity to form the complex. It depends on the energy associated with bonds and molecular 20 conformation [23]. Metal complexed fraction is related with stability; generally, the greater is the complexed fraction, the higher is the chemical stability of the complex [24]. The metal complexed ratio also indicates the ability of the complexing agent to keep the metal complexed in specified conditions of reaction (pH, stoichiometry, metal concentration) [25]. Several techniques are used for the determination of complex chemical properties. Thermodynamic stability determination is not a simple technique and requires a number of calculations with variables related to energetic parameters [23,26]. Thus, the use of theoretical calculations and computational modeling to assess the molecular structure and their thermodynamic parameters have been shown to be a very useful tool [27,28]. Determination of the complexation ratio usually involves chromatographic and spectrometric analyses or separation methods based on ultrafiltration, which are expensive and time consuming, limiting their application [18,29]. Therefore, the chemical speciation obtained by specific software has been largely used to determine the products of complexation reactions in solution [30–32]. Fourier Transform Infrared (FTIR) spectroscopy is a valuable tool to study the changes in functional groups signatures and in the molecular structure of metal complexes. This technique has been used to identify functional groups involved in the complexation reaction and conformational features of the formed complexes [33,34]. The band shift relative to specific vibrational modes can be used as an index of coordination modes and aspects of the molecular structure [35]. Multivariate calibration methods have been successfully applied to model spectroscopic data, with the aim of constructing predictive models for the determination of chemical properties of compounds [36]. Partial Least Squares (PLS) regression is one of the most popular multivariate regression techniques used to predict sample properties from FTIR spectra [37,38]. This technique is often used to predict the content of specific components from samples, such as fatty acids [39], flavonoids, and carotenoids [40] in food samples, as well as to estimate the content of soil organic matter and its properties in soils and sediments [41], and to predict aromaticity, C/N ratio, cation exchange capacity, and dissolved organic carbon content in plant substrates [42]. Considering that FTIR is a fast, non-destructive method and does not require sample pre-treatment, this analysis and the calibration techniques used to predict sample properties from spectra have attracted much interest within the green-chemistry scope and its multiple applications [36]. The possibility of determining properties that are conventionally highly cost and time-consuming from a simple, non-destructive, and inexpensive method, such as FTIR, is highly desirable [36,42]. Therefore, the main aim of this study was to find out key properties of metal-carboxylate complexes (molecular structure, stability, solubility, and metal complexed 21 ratio), that affect their applicability and performance in environmental and agricultural uses, and to study the use of FTIR spectra dataset to predict these chemical properties. Accordingly, in the synthesis of complexes, three different metals (Fe, Mn, and Zn) were mixed with different low molecular weight organic acids as organic ligands (citric, malic, tartaric, and oxalic acids). The key chemical properties of complexes were determined with known methods and briefly discussed. Sequentially, the possibility to use FTIR spectra to verify aspects of molecular structure and to predict metal complexed fraction, solubility, and stability of complexes was investigated. Hypothetically, FTIR combined with PLS regression can be useful and a suitable approach to predict the chemical properties of metal organic acid complexes. 2. Material and Methods 2.1 Synthesis The low molecular weight organic acids studied as complexing agents were: tricarboxylic citric acid (pKa1 = 3.13, pKa2 = 4.76, pKa3 = 6.40), dicarboxylic DL-malic acid (pKa1 = 3.46; pKa2 = 5.10), dicarboxylic L-tartaric acid (pKa1 = 3.04; pKa2 = 4.37), and dicarboxylic oxalic acid (pKa1 = 1.23; pKa2 = 4.19). The 1:1 and 1:2 (metal: ligand molar ratio) stoichiometric ratios (SR) were chosen for organic acid reactions with metals because previous studies have shown that these are the most energetically favorable for transition metals and organic ligands, such as citric and oxalic acids [39,43]. Stock solutions of metal-sulfates (MnSO4.4H2O, FeSO4.7H2O, and ZnSO4.7H2O) and organic acids were prepared at 0.1 mol L- 1. The complexes were prepared by mixing 50 mL of the metal-sulfate stock solution with 50 mL (1:1 SR) or 100 mL (1:2 SR) of organic acid stock solutions. The pH was adjusted to 7.0 using a KOH 0.5 mol L-1 solution. The salts used in the formulation of complexes were analytical grade reagents. Regarding the complex solubility characterization, triplicate solutions at 0.01 mol L-1 concentration were prepared and shaken for 30 min. In sequence, the solutions were centrifuged to separate the precipitate, and the metal concentration was determined in the supernatant in an atomic absorption spectrometer (AAS). 2.2 Metal chemical speciation To estimate the amount of metal complexed by the organic acids, as well as the main species of complexes formed in solution, the software of chemical speciation VISUAL Minteq version 3.1 was used. For the estimations, the metal (M) salts and complexing agents were added as initial compounds with the same concentration used in the laboratory synthesis (0.01 22 mol L-1). The titration mode was used to reach the pH 7, and the titrant used was KOH 0.5 mol L-1, mirroring the laboratory conditions. In the titration mode, several steps are previously defined, and the result of chemical speciation in each step can be selected to visualization. The temperature used was 25ºC and the CO2 partial pressure and its dissolution in the solutions were not considered. The chemical species of the VISUAL Minteq outputs were classified as free- metal or complexed-metal. Free-metal represented the sum of ionic and non-complexed forms, such as M2+, MOH+ and MSO4 (aq). The complexed metal represented the sum of all other forms of metals associated with the organic ligands (complexes), with negative, positive and neutral charges. Complexed-metals formed with one metal atom and one molecule of ligand were classified as 1:1 type, such as the M(Citrate)- and M(Malate) chemical species. Complexed-metals formed with one metal atom and 2 ligand molecules were classified as 1:2 type, such as the [M(Oxalate)2] -2 and [M(Tartrate)2] -2 chemical species. 2.3 Computational modeling The geometry optimization of isolated organic acids and the complexes formed between metals and organic acids was quantum-chemically performed using the Gaussian 09W software. Both 1:1 and 1:2 (metal: ligand) stoichiometries were designed as input molecules for all metal complexes [28,43]. For complexes, the most likely geometry was considered with three carboxyl groups connected to metals for CA and with two carboxyl groups for MA, TA and OA at a 1:1 stoichiometry. Regarding the 1:2 stoichiometry, the most likely geometry was considered with three or two carboxyl groups of each organic acid molecules binding to metals as input geometries. All the input geometries were subsequently optimized at the density functional theory (DFT) ωWB97XD method, using the LanL2DZ basis set [28], which is a suitable and effective core potential for post-third row atoms, such as Fe, Mn and Zn. The calculations were performed considering both the gas phase and water as implicit solvents [28,39] (through the polarizable continuum model using the integral equation formalism)[44]. After the geometry optimization, the thermodynamic parameters were studied in order to compare the stability of the complexes tested. Accordingly, the standard Gibbs free energies (Gº) for the ligand, iron and complex were obtained and, then, used to calculate the standard Gibbs free energy of complexation (ΔGº), according to the following Equation 1: ΔGº = Gº complex – (Gº organic acid + Gº iron) (Eq. 1) 2.4 FTIR Analysis 23 For the Fourier Transform Infrared analysis, the solutions of metal-complexes were frozen and lyophilized to obtain solid samples. The samples were homogenized and analyzed in a Cary 630 Agilent® equipment with a ZnSe crystal with an ATR configuration . The scans were done at 4 by 4 cm-1 resolution, with 64 scans, and from 650 to 4000 cm-1. The software Fityk [44] was used for peak processing (peaks deconvolution). 2.5 Partial Least Squares Regression Spectral data for 24 complex samples (Fe, Mn, and Zn complexes of citric, malic, oxalic, and tartaric acids, at both 1:1 and 1:2 stoichiometric ratios) were used for PLS regression analysis. The FTIR spectra (absorbance from 4000 to 650 cm-1) of complexes were regressed against the respective chemical attributes (solubility, total metal complexed fraction, and stability) using partial least squares (PLS) regression, thus yielding PLS calibration models. The models obtained can be used to estimate chemical parameters of the complexes without to need using classical methods. The calibration performance was evaluated using the root mean square error of calibration (RMSEc) and the squared correlation coefficient of calibration (R2 cal)[42,45]. RMSEc is a measure of how well the model fits the experimental results, and it was calculated through the equation 2 [45]: RMSEc = √ ∑ (�̂�𝑖−𝑦𝑖)2𝑛 𝑖=1 𝑛 (Eq. 2) where yi is the reference value of the dependent variable, �̂�i is the predicted value, y is the mean value, and n is the number of samples. The leave-one-out cross validation method and a y-randomization test were used to validate the PLS models [46]. The root mean-square error of cross validation and the determination coefficient of cross validation were calculated according to the recommended literature [45]. The performance parameters root mean square of error of y-randomization and determination coefficient in the y-randomization were also calculated according to [45]. The models were tested through an external group of test samples, which were not included in the calibration process and used to verify the predictive capacity of the PLS models. From this step, the root mean square error of prediction (RMSEp), and the coefficient of prediction (R2pred) were determined and used as statistical parameters to evaluate the model performance on prediction [45]. The multivariate modeling and calculations were carried out using the Chemoface version 1.4 software [47]. 24 3. Results and Discussion 3.1 Complex stability and molecular structure modeling The computationally optimized molecular structures of the organometallic complexes are shown in Figure 1 (iron complexes), Figure 2 (manganese complexes), and Figure 3 (zinc complexes). Complexes of 1:2 type for all organic acids exhibited four bonds with metals, yielding a tetrahedral geometry (Figure 1, 2, 3). Although the citric acid has three carboxylic groups, in the 1:2 configuration only two groups of each molecule interacted with metals. The isomer with three carboxylic groups interacting with metal was tested, but it was less stable than the isomer with two carboxylic groups, for all metals. This probably occurred due to a high repulsion between the negatively charged oxygens, increasing the steric hindrance in the isomer with three carboxylates positioned in the center around the metal. Three bonds with the metal were formed in 1:1 citrate complexes, configuring a pyramidal arrangement. In 1:1 malate, tartrate and oxalate complexes, two bonds are formed with metals. Interactions of carboxylate groups with metal exhibited a monodentate configuration for all complexes studied (Figures 1, 2 and 3). The Gibbs free energies also obtained through the theoretical calculations for the complexes formed between metal and organic acid as ligand are shown in Table 1. Complexes with Fe, in general, exhibited the lowest values of energy, suggesting greater stability of Fe complexes compared to Mn and Zn complexes. Zn complexes exhibited a lower stability compared with Mn and Fe. Stability of metal complexes generally increases with a decrease in the size of metal cations [23,50]. Zinc has an ionic radius (0.74 Å) larger than Mn and Fe (both of them around to 0.7Å), which partially explains the lower stability of zinc complexes. In general, the 1:2 complex type (with four bonds with metal) was more stable than 1:1 (with three or two bonds to metal) for all the ligands and metals considered. The number of bonds formed around the metal ion could be ascribed to its coordination number [50]. The coordination number indicates the number of electron pairs that metal should receive to complete their orbitals in the ion valence layer [51]. This number is not unique because other factors, such as ligand size and charge, could affect their binding capacity [52]. Zn coordination number is usually 4, but it can also present a coordination number 6. Fe and Mn present octahedral coordination (coordination number 6) in most cases [52]. Therefore, differences in coordination number could explain the differences in stability. In the 1:2 stoichiometry ratio of metal-organic acids, more electron pairs are donated to metal ions, either achieving their coordination number (in the case of Zn) or approaching to it (in the case of Mn 25 and Fe). The coordination number could be reduced in cases of high steric hindrance of ligands with large molecular size [43,52], such as in the 1:2 citrate complexes that exhibit only two carboxylic groups in each citrate ligand interacting with metals, instead of three groups. Furthermore, carboxylate groups rotate from its free to complexed forms [43], affecting the neat stability values. Other variables, such as molar concentration, pH and reaction stoichiometry, can influence the formation and stability of the complexes [18], but these factors were not considered in the theoretical calculations done in this study. Thus, even being considered a more stable compound according to computational modeling, the formation of a complex could be conditioned by the solution parameters as well. For this reason, the chemical speciation, considering pH, molarity and stoichiometry, was also estimated using the Visual Minteq software. Table 1. Sum of electronic and thermal free energies (ΔG°, kcal mol-1) of organometallic complexes in implicit water, according to the metal and stoichiometry. Ligand Stoichiometry Fe Mn Zn Citrate 1:1 -72.4 -58.0 -32.9 1:2 -90.3 -74.0 -48.9 Tartrate 1:1 -48.7 -39.8 -12.0 1:2 -88.6 -47.9 -35.8 Malate 1:1 -58.3 -23.8 -19.1 1:2 -93.8 -34.9 -26.9 Oxalate 1:1 -56.8 -46.1 -18.6 1:2 -91.9 -82.8 -56.5 26 Figure 1. Molecular structure of iron complexes with citrate (A), malate (B), tartrate (C) and oxalate (D). The complexes at left are of 1:1 type, and 1:2 type at the right side. 27 Figure 2. Molecular structure of manganese complexes with citrate (A), malate (B), tartrate (C) and oxalate (D). The complexes at left are of 1:1 type, and 1:2 type at the right side. 28 Figure 3. Molecular structure of zinc complexes with citrate (A), malate (B), tartrate (C) and oxalate (D). The complexes at left are of 1:1 type, and 1:2 type at the right side. 3.2 Chemical Speciation: Effect of Ligands, Stoichiometry and Metals Organic acids, metals, and stoichiometry of reaction influenced the free and complexed metal ratios in solution, as well as the amount of each type of the complex formed (1:1 or 1:2) (Figure 4). For all three metals, citrate at 1:2 SR kept the highest ratio (>99%) of the metal in the complexed form, while TA at 1:1 SR has the lowest metal complexed ratio for Fe and Zn (58% and 56% for Fe and Zn, respectively). Regarding Mn complexes, the lowest complexed ratio was verified for MA at 1:1 SR (61%). Among the complexing agents, oxalate is more prone to form complexes of 1:2 type ([Fe(Oxalate)2 -2], [Mn(Oxalate)2 -2], and [Zn(Oxalate)2 -2]). 29 Furthermore, regarding the metals, Zn is the most prone to form 1:2 type complexes. The 1:2 Zn complexes were formed with both tested SR for citrate, tartrate and oxalate, whereas Fe and Mn 1:2 complexes were formed only with oxalate. Metal malate complexes at 1:2 SR were not observed for any metal (Figure 4). The chemical speciation is not in full agreement with the results found by the theoretical calculations. Although the 1:1 stoichiometry complex simulation, in general, presented lower stability than 1:2 stoichiometry complex according to the theoretical calculations, the former was predominant in solutions according to Minteq outputs. In line with above-mentioned, metal coordination number is not unique, and, in general, the metals can adopt more than one coordination number, depending on the ligand and reaction conditions [52]. Coordination number is highly dependent on metal size, ligand size and electron-donating ability [52]. According to chemical speciation (Figure 4), ligands with a larger size, such as citrate, formed 1:2 complexes with Zn, the element with a higher ionic radius (0.74 Å) than Mn and Fe (Figure 4). Metals with lower ionic radius than Zn, such as Mn2+ (0.70 Å) and Fe2+ (0.70 Å), did not present 1:2 complexes with citrate (Figure 1) in the tested conditions. Oxalate is more prone to form 1:2 complexes due to its low molecular weight/size, which minimizes the steric hindrance [39]. However, even with oxalate, the proportion of 1:2 for Mn and Fe is lower than with Zn, the element among the nutrients studied with the highest ionic radius. Although the 1:2 stoichiometry complexes were more stable according to the theoretical calculations, the occurrence of these complexes in solution should be associated to a lower pH (pH<4) and a higher proportion of ligand relative to metal [27]. In the phloem sap of Fe deficient plants, e.g. the Fe: citrate proportions found are around 1:75 – 1:250, and, in this case, the 1:2 and 1:3 Fe- citrate species are highly predominant over 1:1 species [53]. The stoichiometry also affects the amount of total complexed metals and the predominant type of complexes. Oxalate and Zn, the ligand and metal more prone to 1:2 complex type formation, exhibit almost the total of complexed ratio composed of 1:2 complexes at the 1:2 stoichiometry preparation (Figure 4). Conversely, in the 1:1 stoichiometric ratio solution, proportionally, less Zn-oxalate of 1:2 type is formed over 1:1 type, due to the insufficient amount of the organic ligand. 30 Figure 4. Metal complexed fraction and free metal fraction at pH 7 and 0.01 mol L-1 as related to metal (Fe, Mn, and Zn, Fig. 1A, 1B, and 1C, respectively), complexing agent (citrate, malate, tartrate, and oxalate), and the stoichiometric ratio of the reaction (1:1 and 1:2 metal: ligand molar ratio). 3.4 Solubility Soluble and insoluble portions of the metals tested, as related to organic acids used as complexing agents and stoichiometry of reaction, are shown in Figure 5. Organic acids and stoichiometry influenced the soluble and insoluble metal ratios. In general, oxalate complexes exhibited the lowest soluble ratios for Fe (< 30%), Mn (<25%), and Zn (<30%) (Figure 5). Oxalate complexes generally have low solubility [54]. Metal-oxalate complexes form polymeric chains, resulting in crystalline precipitate materials [55]. Conversely, citrate and malate complexes with the three metals used in this study exhibited high solubility (Figure 5). Significant low ratios of insoluble material of citrate and malate (Figure 5) could be ascribed to the formation of metal oxides with the minimum metal ratio that remained not complexed (Figure 5). Tartrate generates salts with high or moderate solubility [56]. In this study, tartrate exhibited soluble metal ratio higher than oxalate, but lower than citrate and malate, for all the studied metals and stoichiometries. Therefore, the insoluble fraction of metals in tartrate solutions could be ascribed to the i) possible precipitation of tartrate salts [56] and/or ii) lower complexation capability of this ligand (Figure 5), resulting in more free metal ratios, which precipitates as hydroxide/oxide forms at pH 7. 31 Figure 5. Soluble and insoluble ratios at pH 7 and 0.01 mol L-1 as related to metal (Fe, Mn, and Zn, Fig. 1A, 1B, and 1C, respectively), complexing agent (citrate, malate, tartrate, and oxalate), and stoichiometric ratio of the reaction (1:1 and 1:2 metal: ligand molar ratio). 3.3 Fourier Transform Infrared The infrared spectra of citric, tartaric, malic, and oxalic acid and their complexes with Fe, Mn, and Zn are shown in Figure 6. A general assignment of characteristic vibrational modes found in the IR spectra according to theory and literature [33,34,43,57–60] are shown in Table 2. 32 Figure 6. Infrared spectra of organic acids (citrate, malate, tartrate, and oxalate) and their respective metal-complexes at 1:1 and 1:2 stoichiometric ratios. 33 Table 2. General assignments of FTIR bands of metal-organic acid complexes according to theory and literature [33,34,43,57–60]. Vibration Wavenumber (cm-1) v (O−H) 3350-3250 v (C−H) 3100-2800 v (COO-)asym 1620-1540 v (COO-)sym 1440-1360 ẟ (O−H) 1440-1395 ẟ OH···O out of plane 1320-1280 v (C−O) 1080-1020 SO2 1120-1050 v (C−C) + ẟ (O−C−O) 820-760 v: stretching; ẟ: rocking 3.3.1 Metal-carboxylate coordination according to COO- vibration stretching shifts The interactions of metals with carboxylates occur through the carboxylic groups; thus, the most useful typical bands of metal-carboxylates are the COO- stretching vibrations [33]. The frequencies and intensities of the carboxylate bands are sensitive to the structure and molecular environment of the carboxylate group, the nature of the complexed metal, and the structure of the complex formed [61]. The shift in the COO- asymmetric stretching is associated to different coordination modes of the carboxylic groups, a stereochemical aspect that affects the metal-complex bioavailability. In the monodentate mode of coordination, a strong reduction of equivalence of the two oxygens of COO- group occurs, and, in general, an increase in the COO- asymmetric stretching frequency is observed [63,64]. A decrease in the COO- asymmetric stretching wavenumber is particularly evident for chelating configuration of complexation [35]. Therefore, the formation of complexes with metals affects the distance between the asymmetric and symmetric COO- bands. The separation of the bands (ΔCOO-) is also used as an indicative of the type of metal-carboxylate coordination binding [61]. Thus, the detailed assignment of asymmetric and symmetric vibrations for each organic acid and their respective metal complexes are shown in Table 3. In comparison with its corresponding free deprotonated ligand, a large splitting of COO- stretching frequencies (generally higher than 200 cm-1) is often an indication of monodentate coordination in a metal carboxylate [33,65]. Reduced splitting of COO- stretching frequencies 34 in comparison with ionic free deprotonate ligand are indicative of chelating or bridging configuration, but bridging splitting is larger than chelating [66,67]. According to chemical speciation outputs, at pH 7, the free carboxylic acids are completely deprotonated and, therefore, it can be assumed that the COO- is present instead of COOH. In the citrate complexes, a larger splitting of carboxylic bands than free citrate ions were observed for Mn complexes at both stoichiometries, suggesting a monodentate coordination. This configuration is confirmed by the molecular structures obtained from computational modeling (Figure 2). Regarding Fe and Zn, at both stoichiometries, ΔCOO- presents values approximately equal to citrate (Table 3). Although these ΔCOO- values suggest a bidentate bridging interaction or ionic interaction with carboxylic groups, all these complexes exhibited monodentate interaction (Figure 1 and Figure 3). A low splitting of COO- stretching vibrations for monodentate complexes are also observed for Cd and Pb citrate complexes [27]. Furthermore, Deacon and Philips [66] also found acetate monodentate complexes of Zn with small ΔCOO- values, similar to free ionic COO-. The shift of COO- asymmetric band in a ligand, and, in consequence, the ΔCOO- values, is also dependent on the properties of metal ion interacting with carboxylate [68]. Furthermore, water molecules can interact with the free O from the carboxylate group bound to the metal, forming a pseudo-bridging mode, reducing the ΔCOO- value [67]. For tartaric and malic acids, in general, all the metal complexes presented ΔCOO- values higher than the corresponding free acid (Table 3), also indicating a monodentate interaction. The theoretical modelling mirrored this result, demonstrating that two carboxylate groups of tartrate and malate are bound to metals in a monodentate configuration (Figures 1, 2, and 3). In case of oxalate, all complexes presented ΔCOO- values remarkably higher than free oxalate ions, indicating monodentate configuration, in line with the theoretical calculation outputs. Therefore, a good association was found between the splitting of COO- vibrations and coordination modes of carboxylates in the structure of metal complexes with malate, tartrate, and oxalate. Only some exceptions to the rule were found for Zn and Fe complexes of citrate. Thus, an additional examination of the molecular structure performed in addition to FTIR spectroscopic analysis is recommendable to confirm the coordination modes. 35 Table 3. Carboxylate symmetric and asymmetric stretching frequencies for different metal carboxylic acid complexes. Metal-Organic Acid Stoichiometry v (COO - )asym v (COO - )sym Delta (COO - ) Citrate - 1559 1357 202 Fe 1:1 1580 1380 200 1:2 1578 1377 201 Mn 1:1 1587 1364 223 1:2 1586 1365 221 Zn 1:1 1573 1380 193 1:2 1575 1375 200 Tartrate - 1578 1394 184 Fe 1:1 1587 1358 236 1:2 1592 1358 234 Mn 1:1 1584 1380 204 1:2 1585 1374 211 Zn 1:1 1589 1379 210 1:2 1586 1375 211 Malate - 1582 1375 207 Fe 1:1 1611 1371 240 1:2 1588 1365 223 Mn 1:1 1592 1383 209 1:2 1604 1375 228 Zn 1:1 1590 1358 231 1:2 1602 1354 248 Oxalate - 1573 1306 267 Fe 1:1 1634 1311 323 1:2 1629 1312 317 Mn 1:1 1628 1310 318 1:2 1636 1307 329 Zn 1:1 1650 1315 335 1:2 1637 1338 299 36 3.3.2 Partial least squares analysis and complex properties prediction from FTIR PLS multivariate regression analysis was used to predict the relationship between the infrared spectral data and the solubility, stability, and total complexed metal fraction of the complexes. Although all these chemical attributes were tested, only solubility and stability were properly predicted using PLS. The statistical parameters obtained from the PLS regression for solubility and complexed fraction are shown in Table 4. Multivariate calibration with PLS was applied to a set of 24 complexes, which considered three metals and four organic acids at two stoichiometric ratios. The number of latent variables used for each model was stablished according to the minimal value of root mean square error of cross-validation (RMSE cv) [45]. Table 4. Parameters of PLS regression for models of metal solubility and metal complexed ratio. PLS Parameter Solubility (% metal in soluble form) Complexed ratio (% metal in complexed form) LV 3 6 RMSE c 13.39 3.33 R2 cal 0.83 0.94 RMESE y-rand 25.79 7.66 R2 y-rand 0.38 0.64 cR2 p 0.61 0.51 RMSE cv 22.07 8.98 R2 pred 0.57 0.63 LV: latent variable; RMSE: root mean square error; R2: determination coefficient; RMSEc: RMSE of calibration; R2 cal: R2 of calibration; RMSEy-rand: RMSE of y randomization; R2 y-rand: R2 of y-randomization; r2p: r2 of y- randomization prediction; RMSE cv: RMSE of cross validation; RMSE p: RMSE of prediction; R2 pred: R2 of prediction For PLS regression-based models, values for the determination coefficients of calibration (R2 cal) > 0.6 and prediction (R2 pred) > 0.5 are acceptable [69]. Thus, the PLS models were highly predictive, both for solubility and ratio of complexation (Table 4). These models were also reliable and not prone to overfitting, since the cR2 p ≥ 0.5 for both solubility and complexation ratio models indicated R2 cal are significantly reduced after randomization (R2 y- rand) of the dependent variables block (y, the solubility and complexation ratio values) [45]. The predictive error of the models expressed as RMSEP is analytically acceptable [45] for the prediction of soluble ratio and for total complexed ratio, corresponding to a relative analytical of 13.39 and 3.33% (Table 4). In general, the PLS models had a greater performance to predict the metal complexed ratio than for complex solubility. The bands changes due to complexation 37 is unequivocal [61,65]. In line with this study, several authors attested the relationship between complexation and the complex spectroscopic properties [33,43]. To illustrate the predictive ability of the developed models, examples of the regression lines of the cross-validated PLS-1 models for each parameter are shown in Figure 4. Figure 4. Predicted x Measured values for metal solubility (left) and metal complexed ratio (right). The coordination mode of metal complexes can affect their recognition by cell membranes of soil bacteria and the uptake and metabolism of complexes by plants [25], and therefore the efficiency of bioremediation or plant nutrition. Complexation capacity, represented in this work as metal complexed ratio, and solubility are essential properties to base the choice for the most effective complexing agent for Fe, Mn, and Zn, depending on the end purpose. Agronomic applications require complexes with high solubility and high complexation capacity because the main objective is to keep the nutrient in soluble forms and avoid reactions with other chemical species to guarantee plentiful plant acquisition [4,24]. Regarding environmental applications, if the intention is to increase metal extraction by plants (phytoremediation), it is also suitable to use complexes with high solubility, stability and complexation capacity [22,70], such as citrate and malate. Otherwise, if the objective is to inactivate toxic metals, to avoid their negative effects on plants and microorganisms, metal complexes with high stability and low solubility [5,70], such as metal-oxalates, should be accounted for. Therefore, the prediction of these properties through PLS regression using FTIR data, a fast and non-destructive method, is a suitable technique to study the chemical properties of Mn, Zn, and Mn complexes and facilitate the decisions regarding their applicability. 38 4. Conclusions Upon to the synthetic conditions of this study, Mn and Fe reacted with all organic acids and formed complexes with 1:1 stoichiometry, except for oxalate, which also presented complexes of 1:2 type. Regarding to Zn, 1:2 type complexes were also formed with tartrate and citrate at 1:2 stoichiometry of reaction. In the structure of all complexes, the carboxylate groups involved in the complexation interacts in a monodentate configuration with metals. The prediction of the structure and type of coordination interaction was provided by FTIR spectra combined with PLS regressions models. However, there are some few exceptions to general rules of splitting value association with coordination modes, which can affect the precision of conformational identification. The combination of FTIR spectroscopy with statistical multivariate techniques demonstrated a great potential of this approach in predicting complex properties such as solubility and stability. Acknowledgements Many thanks to Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES), by the scholarship provided to the first author and to the research funding (CAPES-PROEX- AUXPE 593-2018); to the Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq, processo 307447-2019-7) for funding this research. 5. References [1] M.W.H. Evangelou, M. Ebel, A. Schaeffer, Chelate assisted phytoextraction of heavy metals from soil. Effect, mechanism, toxicity, and fate of chelating agents, Chemosphere. 68 (2007) 989–1003. https://doi.org/10.1016/j.chemosphere.2007.01.062. [2] S. López-Rayo, P. Nadal, J.J. Lucena, Novel chelating agents for iron, manganese, zinc, and copper mixed fertilisation in high pH soil-less cultures, J. Sci. Food Agric. 96 (2016) 1111–1120. https://doi.org/10.1002/jsfa.7183. [3] C. Martín-Fernández, Á. Solti, V. Czech, K. Kovács, F. Fodor, A. Gárate, L. Hernández-Apaolaza, J.J. Lucena, Response of soybean plants to the application of synthetic and biodegradable Fe chelates and Fe complexes, Plant Physiol. Biochem. 118 (2017) 579–588. https://doi.org/10.1016/j.plaphy.2017.07.028. 39 [4] P. Rodríguez-Lucena, L. Hernández-Apaolaza, J.J. Lucena, Comparison of iron chelates and complexes supplied as foliar sprays and in nutrient solution to correct iron chlorosis of soybean, J. Plant Nutr. Soil Sci. 173 (2010) 120–126. https://doi.org/10.1002/jpln.200800256. [5] Z. Derakhshan Nejad, M.C. Jung, K.H. Kim, Remediation of soils contaminated with heavy metals with an emphasis on immobilization technology, Environ. Geochem. Health. 40 (2018) 927–953. https://doi.org/10.1007/s10653-017-9964-z. [6] A. Singh, S.M. Prasad, Remediation of heavy metal contaminated ecosystem: an overview on technology advancement, Int. J. Environ. Sci. Technol. 12 (2015) 353– 366. https://doi.org/10.1007/s13762-014-0542-y. [7] S. López-Rayo, I. Sanchis-Pérez, C.M.H. Ferreira, J.J. Lucena, [S,S]-EDDS/Fe: A new chelate for the environmentally sustainable correction of iron chlorosis in calcareous soil, Sci. Total Environ. 647 (2019) 1508–1517. https://doi.org/10.1016/j.scitotenv.2018.08.021. [8] B. Nowack, Environmental chemistry of aminopolycarboxylate chelating agents, Environ. Sci. Technol. 36 (2002) 4009–4016. https://doi.org/10.1021/es025683s. [9] D. Leštan, C. ling Luo, X. dong Li, The use of chelating agents in the remediation of metal-contaminated soils: A review, Environ. Pollut. 153 (2008) 3–13. https://doi.org/10.1016/j.envpol.2007.11.015. [10] A. Ullmann, N. Brauner, S. Vazana, Z. Katz, R. Goikhman, B. Seemann, H. Marom, M. Gozin, New biodegradable organic-soluble chelating agents for simultaneous removal of heavy metals and organic pollutants from contaminated media, J. Hazard. Mater. 260 (2013) 676–688. https://doi.org/10.1016/j.jhazmat.2013.06.027. [11] G. Renella, L. Landi, P. Nannipieri, Degradation of low molecular weight organic acids complexed with heavy metals in soil, Geoderma. 122 (2004) 311–315. https://doi.org/10.1016/j.geoderma.2004.01.018. [12] R. Adeleke, C. Nwangburuka, B. Oboirien, Origins, roles and fate of organic acids in soils: A review, South African J. Bot. 108 (2017) 393–406. https://doi.org/10.1016/j.sajb.2016.09.002. [13] D.L. Jones, Organic acids in the rhizosphere - A critical review, Plant Soil. 205 (1998) 40 25–44. https://doi.org/10.1023/A:1004356007312. [14] C. Ash, V. Tejnecký, L. Borůvka, O. Drábek, Different low-molecular-mass organic acids specifically control leaching of arsenic and lead from contaminated soil, J. Contam. Hydrol. 187 (2016) 18–30. https://doi.org/10.1016/j.jconhyd.2016.01.009. [15] J. Jiang, M. Yang, Y. Gao, J. Wang, D. Li, T. Li, Removal of toxic metals from vanadium-contaminated soils using a washing method: Reagent selection and parameter optimization, Chemosphere. 180 (2017) 295–301. https://doi.org/10.1016/j.chemosphere.2017.03.116. [16] B. Chakraborty, P.N. Singh, S. Kumar, P.C. Srivastava, Uptake and Distribution of Iron from Different Iron Sources Applied as Foliar Sprays to Chlorotic Leaves of Low- Chill Peach Cultivars, Agric. Res. 3 (2014) 293–301. https://doi.org/10.1007/s40003- 014-0128-4. [17] S. Sharma, H. Malhotra, P. Borah, M.K. Meena, P. Bindraban, S. Chandra, V. Pande, R. Pandey, Foliar application of organic and inorganic iron formulation induces differential detoxification response to improve growth and biofortification in soybean, Plant Physiol. Reports. 24 (2019) 119–128. https://doi.org/10.1007/s40502-018-0412- 6. [18] J.M. Garcia-Mina, Stability, solubility and maximum metal binding capacity in metal- humic complexes involving humic substances extracted from peat and organic compost, Org. Geochem. 37 (2006) 1960–1972. https://doi.org/10.1016/j.orggeochem.2006.07.027. [19] I. Guardado, O. Urrutia, J.M. García-Mina, Size distribution, complexing capacity, and stability of phosphate-metal-humic complexes, J. Agric. Food Chem. 55 (2007) 408– 413. https://doi.org/10.1021/jf062894y. [20] G. Joshi-Tope, A.J. Francis, Mechanisms of biodegradation of metal-citrate complexes by Pseudomonas fluorescens, J. Bacteriol. 177 (1995) 1989–1993. https://doi.org/10.1128/jb.177.8.1989-1993.1995. [21] J.J. Lensbouer, A. Patel, J.P. Sirianni, R.P. Doyle, Functional characterization and metal ion specificity of the metal-citrate complex transporter from Streptomyces coelicolor, J. Bacteriol. 190 (2008) 5616–5623. https://doi.org/10.1128/JB.00456-08. 41 [22] J.M. García-Mina, M.C. Antolín, M. Sanchez-Diaz, Metal-humic complexes and plant micronutrient uptake: A study based on different plant species cultivated in diverse soil types, Plant Soil. 258 (2004) 57–68. https://doi.org/10.1023/B:PLSO.0000016509.56780.40. [23] S. Muthaiah, A. Bhatia, M. Kannan, Stability of Metal Complexes, Stab. Coord. Compd. [Working Title]. (2020). https://doi.org/10.5772/intechopen.90894. [24] W.L. Lindsay, Chemical reactions in soils that affect iron availability to plants. A quantative approach, Iron Nutr. Soils Plants. (1995) 7–14. https://doi.org/10.1007/978- 94-011-0503-3_2. [25] A.K. Pandey, S.D. Pandey, V. Misra, Stability constants of metal-humic acid complexes and its role in environmental detoxification, Ecotoxicol. Environ. Saf. 47 (2000) 195–200. https://doi.org/10.1006/eesa.2000.1947. [26] F. Arnaud-Neu, R. Delgado, S. Chaves, Critical evaluation of stability constants and thermodynamic functions of metal complexes of crown ethers: (IUPAC technical report), Pure Appl. Chem. 75 (2003) 71–102. https://doi.org/10.1351/pac200375010071. [27] A.C. Bertoli, R. Carvalho, M.P. Freitas, T.C. Ramalho, D.T. Mancini, M.C. Oliveira, A. De Varennes, A. Dias, Theoretical spectroscopic studies and identification of metal- citrate (Cd and Pb) complexes by ESI-MS in aqueous solution, Spectrochim. Acta - Part A Mol. Biomol. Spectrosc. 137 (2015) 271–280. https://doi.org/10.1016/j.saa.2014.08.053. [28] A.C. Bertoli, R. Carvalho, M.P. Freitas, T.C. Ramalho, D.T. Mancini, M.C. Oliveira, A. De Varennes, A. Dias, Theoretical and experimental investigation of complex structures citrate of zinc (II), Inorganica Chim. Acta. 425 (2015) 164–168. https://doi.org/10.1016/j.ica.2014.10.025. [29] M. Villén, J.J. Lucena, M.C. Cartagena, R. Bravo, J. García-Mina, M.I.M. De La Hinojosa, Comparison of two analytical methods for the evaluation of the complexed metal in fertilizers and the complexing capacity of complexing agents, J. Agric. Food Chem. 55 (2007) 5746–5753. https://doi.org/10.1021/jf070422t. [30] L. Weng, E.J.M. Temminghoff, S. Lofts, E. Tipping, W.H. Van Riemsdijk, Complexation with dissolved organic matter and solubility control of heavy metals in a 42 sandy soil, Environ. Sci. Technol. 36 (2002) 4804–4810. https://doi.org/10.1021/es0200084. [31] C. Catrouillet, M. Davranche, A. Dia, M. Bouhnik-Le Coz, R. Marsac, O. Pourret, G. Gruau, Geochemical modeling of Fe(II) binding to humic and fulvic acids, Chem. Geol. 372 (2014) 109–118. https://doi.org/10.1016/j.chemgeo.2014.02.019. [32] S. Goldberg, Chemical modeling of boron adsorption by humic materials using the constant capacitance model, Soil Sci. 179 (2014) 561–567. https://doi.org/10.1097/SS.0000000000000098. [33] S.K. Papageorgiou, E.P. Kouvelos, E.P. Favvas, A.A. Sapalidis, G.E. Romanos, F.K. Katsaros, Metal-carboxylate interactions in metal-alginate complexes studied with FTIR spectroscopy, Carbohydr. Res. 345 (2010) 469–473. https://doi.org/10.1016/j.carres.2009.12.010. [34] M.C. D’Antonio, A. Wladimirsky, D. Palacios, L. Coggiola, A.C. González-Baró, E.J. Baran, R.C. Mercader, Spectroscopic investigations of iron(II) and iron(III) oxalates, J. Braz. Chem. Soc. 20 (2009) 445–450. https://doi.org/10.1590/S0103- 50532009000300006. [35] V. Zeleňák, Z. Vargová, K. Györyová, Correlation of infrared spectra of zinc(II) carboxylates with their structures, Spectrochim. Acta - Part A Mol. Biomol. Spectrosc. 66 (2007) 262–272. https://doi.org/10.1016/j.saa.2006.02.050. [36] A. Gredilla, S. Fdez-Ortiz de Vallejuelo, N. Elejoste, A. de Diego, J.M. Madariaga, Non-destructive Spectroscopy combined with chemometrics as a tool for Green Chemical Analysis of environmental samples: A review, TrAC - Trends Anal. Chem. 76 (2016) 30–39. https://doi.org/10.1016/j.trac.2015.11.011. [37] G.W. Small, Chemometrics and near-infrared spectroscopy: Avoiding the pitfalls, TrAC - Trends Anal. Chem. 25 (2006) 1057–1066. https://doi.org/10.1016/j.trac.2006.09.004. [38] N. Zhao, Z.S. Wu, Q. Zhang, X.Y. Shi, Q. Ma, Y.J. Qiao, Optimization of Parameter Selection for Partial Least Squares Model Development, Sci. Rep. 5 (2015) 1–10. https://doi.org/10.1038/srep11647. [39] T.A. Tolentino, A.C. Bertoli, M. Dos Santos Pires, R. Carvalho, C.R.G. Labory, J.S. 43 Nunes, A.R.R. Bastos, M.P. De Freitas, Applications in environmental bioinorganic: Nutritional and ultrastructural evaluation and calculus of thermodynamic and structural properties of metal-oxalate complexes, Spectrochim. Acta - Part A Mol. Biomol. Spectrosc. 150 (2015) 750–757. https://doi.org/10.1016/j.saa.2015.06.016. [40] C. Gardana, A. Scialpi, C. Fachechi, P. Simonetti, Near-infrared spectroscopy and chemometrics for the routine detection of bilberry extract adulteration and quantitative determination of the anthocyanins, J. Spectrosc. 2018 (2018). https://doi.org/10.1155/2018/4751247. [41] M.A. Jiménez-González, A.M. Álvarez, P. Carral, G. Almendros, Chemometric assessment of soil organic matter storage and quality from humic acid infrared spectra, Sci. Total Environ. 685 (2019) 1160–1168. https://doi.org/10.1016/j.scitotenv.2019.06.231. [42] F.S. Higashikawa, C.A. Silva, C.A.Ô. Nunes, M.A. Sánchez-Monedero, Fourier transform infrared spectroscopy and partial least square regression for the prediction of substrate maturity indexes, Sci. Total Environ. 470–471 (2014) 536–542. https://doi.org/10.1016/j.scitotenv.2013.09.065. [43] A.C. Bertoli, R. Carvalho, M.P. Freitas, T.C. Ramalho, D.T. Mancini, M.C. Oliveira, A. De Varennes, A. Dias, Structural determination of Cu and Fe-Citrate complexes: Theoretical investigation and analysis by ESI-MS, J. Inorg. Biochem. 144 (2015) 31– 37. https://doi.org/10.1016/j.jinorgbio.2014.12.008. [44] J. Tomasi, B. Mennucci, R. Cammi, Quantum mechanical continuum solvation models, Chem. Rev. 105 (2005) 2999–3093. https://doi.org/10.1021/cr9904009. [45] R. Kiralj, M.M.C. Ferreira, Basic validation procedures for regression models in QSAR and QSPR studies: Theory and application, J. Braz. Chem. Soc. 20 (2009) 770–787. https://doi.org/10.1590/S0103-50532009000400021. [46] C.A. Nunes, Vibrational spectroscopy and chemometrics to assess authenticity, adulteration and intrinsic quality parameters of edible oils and fats, Food Res. Int. 60 (2014) 255–261. https://doi.org/10.1016/j.foodres.2013.08.041. [47] C.A. Nunes, M.P. Freitas, A.C.M. Pinheiro, S.C. Bastos, Chemoface: A novel free user-friendly interface for chemometrics, J. Braz. Chem. Soc. 23 (2012) 2003–2010. https://doi.org/10.1590/S0103-50532012005000073. 44 [48] O.C. Gagné, F.C. Hawthorne, Empirical Lewis acid strengths for 135 cations bonded to oxygen, Acta Crystallogr. Sect. B Struct. Sci. Cryst. Eng. Mater. 73 (2017) 956–961. https://doi.org/10.1107/S2052520617010988. [49] R.G. Pearson, D.H. Busch, Hard and Soft Acids and Bases, 1963. [50] N.S. Hosmane, Review of Bonding Theories for d -Block Metal Complexes, Adv. Inorg. Chem. (2017) 89–114. https://doi.org/10.1016/b978-0-12-801982-5.00006-0. [51] N.S. Hosmane, Ligands and d -Block Metal Complexes, Adv. Inorg. Chem. (2017) 75– 87. https://doi.org/10.1016/b978-0-12-801982-5.00005-9. [52] M. Dudev, J. Wang, T. Dudev, C. Lim, Factors governing the metal coordination number in metal complexes from cambridge structural database analyses, J. Phys. Chem. B. 110 (2006) 1889–1895. https://doi.org/10.1021/jp054975n. [53] A.F. Lopez-Millan, F. Morales, A. Abadia, J. Abadia, Effects of iron deficiency on the composition of the leaf apoplastic fluid and xylem sap in sugar beet. Implications for iron and carbon transport, Plant Physiol. 124 (2000) 873–884. https://doi.org/10.1104/pp.124.2.873. [54] R.M. Ferreira, M. Motta, A. Batagin-Neto, C.F. De Oliveira Graeff, P.N. Lisboa-Filho, F.C. Lavarda, Theoretical investigation of geometric configurations and vibrational spectra in citric acid complexes, Mater. Res. 17 (2014) 550–556. https://doi.org/10.1590/S1516-14392014005000056. [55] E.J. Baran, Review: Natural oxalates and their analogous synthetic complexes, J. Coord. Chem. 67 (2014) 3734–3768. https://doi.org/10.1080/00958972.2014.937340. [56] Q. Gao, Y.B. Xie, D. Wang, Synthesis and crystal structure of the hydrated nickel(II) complex of tartrate, J. Chem. Crystallogr. 38 (2008) 587–590. https://doi.org/10.1007/s10870-008-9347-5. [57] G. Socrates, Infrared and Raman characteristic group frequencies. Tables and charts, 2001. https://doi.org/10.1002/jrs.1238. [58] C.C.R. Sutton, G. Da Silva, G. V. Franks, Modeling the IR spectra of aqueous metal carboxylate complexes: Correlation between bonding geometry and stretching mode wavenumber shifts, Chem. - A Eur. J. 21 (2015) 6801–6805. https://doi.org/10.1002/chem.201406516. 45 [59] M. Matzapetakis, N. Karligiano, A. Bino, M. Dakanali, C.P. Raptopoulou, V. Tangoulis, A. Terzis, J. Giapintzakis, A. Salifoglou, Manganese citrate chemistry: Syntheses, spectroscopic studies, and structural characterizations of novel mononuclear, water-soluble manganese citrate complexes, Inorg. Chem. 39 (2000) 4044–4051. https://doi.org/10.1021/ic9912631. [60] B.H. Stuart, Infrared Spectroscopy: Fundamentals and Applications, 2005. https://doi.org/10.1002/0470011149. [61] E.G. Palacios, G. Juárez-López, A.J. Monhemius, Infrared spectroscopy of metal carboxylates: II. Analysis of Fe(III), Ni and Zn carboxylate solutions, Hydrometallurgy. 72 (2004) 139–148. https://doi.org/10.1016/S0304-386X(03)00137- 3. [62] M. Clausén, L.O. Öhman, P. Persson, Spectroscopic studies of aqueous gallium(III) and aluminum(III) citrate complexes, J. Inorg. Biochem. 99 (2005) 716–726. https://doi.org/10.1016/j.jinorgbio.2004.12.007. [63] M. Nara, M. Tanokura, Infrared spectroscopic study of the metal-coordination structures of calcium-binding proteins, Biochem. Biophys. Res. Commun. 369 (2008) 225–239. https://doi.org/10.1016/j.bbrc.2007.11.188. [64] K.C. Lanigan, K. Pidsosny, Reflectance FTIR spectroscopic analysis of metal complexation to EDTA and EDDS, Vib. Spectrosc. 45 (2007) 2–9. https://doi.org/10.1016/j.vibspec.2007.03.003. [65] P. Boguta, Z. Sokołowska, Interactions of Zn(II) Ions with Humic Acids Isolated from Various Type of Soils. Effect of pH, Zn Concentrations and Humic Acids Chemical Properties, PLoS One. 11 (2016) 1–20. https://doi.org/10.1371/journal.pone.0153626. [66] G.B. Deacon, R.J. Phillips, Deacon G Phillips R, Coord. Chem. Rev. 33 (1980) 227– 250. [67] M. Nara, H. Morii, M. Tanokura, Coordination to divalent cations by calcium-binding proteins studied by FTIR spectroscopy, Biochim. Biophys. Acta - Biomembr. 1828 (2013) 2319–2327. https://doi.org/10.1016/j.bbamem.2012.11.025. [68] S.C. Edington, A. Gonzalez, T.R. Middendorf, D.B. Halling, R.W. Aldrich, C.R. Baiz, Coordination to lanthanide ions distorts binding site conformation in calmodulin, Proc. 46 Natl. Acad. Sci. U. S. A. 115 (2018) E3126–E3134. https://doi.org/10.1073/pnas.1722042115. [69] D.L.J. Alexander, A. Tropsha, D.A. Winkler, Beware of R2: Simple, Unambiguous Assessment of the Prediction Accuracy of QSAR and QSPR Models, J. Chem. Inf. Model. 55 (2015) 1316–1322. https://doi.org/10.1021/acs.jcim.5b00206. [70] L. Liu, W. Li, W. Song, M. Guo, Remediation techniques for heavy metal- contaminated soils: Principles and applicability, Sci. Total Environ. 633 (2018) 206– 219. https://doi.org/10.1016/j.scitoenv.2018.03.161. 47 ARTICLE 2 - Influence of complex stability on iron accumulation and redistribution for foliar applied iron-organic acid complexes Marina Justi† *; Carlos Alberto Silva†; Everton Geraldo de Morais†; Josué Mariani Silla§; Matheus Puggina de Freitas§ † Department of Soil Science, Federal University of Lavras/UFLA, Post office box 3037, 37200-000, Lavras, MG, Brazil §Department of Chemistry, Federal University of Lavras, Post office box 3037, 37200-000, Lavras, MG, Brazil *Corresponding author Email address: marina.justi@gmail.com This article was prepared in line with the guidelines of the journal ‘Archives of Agronomy and Soil Science’ Abstract The relation between chemical attributes of complexed Fe sources and the efficiency of foliar fertilization is still unclear. This study aimed to investigate the influence of Fe-complex stability on iron accumulation and redistribution in maize by foliar application. The iron complexing agents used were biodegradable citric acid (CA), malic acid (MA), tartaric acid (TA), and oxalic acid (OA). The prepared FeSO4 based complexes in the foliar spray solutions had “Fe: organic acid” stoichiometric ratios of 1:1 and 1:2. Additionally, Fe-EDTA and FeSO4 were also tested. The iron complexed fraction and the main complexed chemical species in the foliar spray solutions were determined using the Minteq software. Molecular modeling helped to obtain probable complex structures and stabilities. The parameters, such as maize dry matter, shoot N accumulation, SPAD index, Fe-shoot, and Fe-root accumulation were determined. The complexes with TA and CA prepared at the stoichiometric ratio 1:2 showed the most favorable results in plant growth, and iron nutrition. The stability and solubility of complexes affected iron accumulation and plant growth. The results indicate that iron redistribution from leaf to the root increased in tandem with complex stability. Additionally, shoot iron accumulation was more significant for complexes with high solubility and low stability. Keywords: carboxylic acids, iron deficiency, iron complexes, Gibbs free energy. 48 1. Introduction Iron is the most required micronutrient by plants, integrating proteins involved in photosynthesis processes (Broadley et al. 2011). Iron availability in soils is controlled by Fe- oxides, which have low solubility, mainly in high soil pH conditions (pH > 7) (Marschner & Rengel 2011), such as calcareous and weathered over-limed soils. Therefore, in alkalinized soils, the concentration of soluble iron is below the requirement for plentiful plant growth (Lindsay 1995). Iron uptake by plant roots is also restricted in conditions of low water availability conditions because diffusion is the primary process involved in iron transport from soil solution to the roots (Marschner & Rengel 2011). Leaf chlorosis, a common symptom of Fe-deficiency, is caused by the severe reduction of photosynthetic pigments, affecting crop development and yield (Abadía et al. 2002). Foliar fertilization is a widely used method to overcome Fe chlorosis (Eichert & Fernández 2011; Malhotra et al. 2020) and increase cereal grain iron concentration in crops (He et al. 2013; Sharma et al. 2019a) It is common to use chelated or complexed forms of iron in foliar sprays because iron is easily oxidized and precipitated in aqueous solution. Additionally, the use of complexed forms of iron increased the mobility of iron from the leaves to non- fertilized plant tissue regions, along with efficient root redistribution (He et al. 2013). The most common chelators used are synthetic aminopolycarboxylic acids, such as EDTA (Eichert & Fernández 2011). Iron chelated forms are highly stable, but in several foliar fertilization studies, their efficiency does not usually exceed that of inorganic iron forms (Shenker & Chen 2005; Rodríguez-Lucena et al. 2010). The high stability is a crucial factor in determining the efficiency of Fe-chelates applied to the soil and nutrient solution (Rodríguez-Lucena et al. 2010). However, the influence of this factor on iron accumulation and translocation for foliar- applied sources is still unclear. Despite the increased use of foliar sprays in crops, the current knowledge about foliar penetration and translocation of free and complexed nutrient forms is still limited (Fernandez & Eichert 2009; Fernández & Brown 2013). Synthetic Fe-chelates are expensive and mainly used for high-value crops. Furthermore, aminopolicarboxylic acids are hardly degraded in the environment, and could influence metal availability and mobility in soils (Nowack 2002). Therefore, the development of environmental-friendly fertilizers is an essential research theme on plant nutrition with iron (Abadía et al. 2011). Soil solution and plant systems naturally contain organic acids, such as citrate, malate, and oxalate (Adeleke et al. 2017). Carboxylic and hydroxylic groups of organic acids can 49 complex cationic elements, acting as natural complexing agents for micronutrients in the soil (Jones 1998). Citric and malic acids are complexing agents for iron transported in the xylem (Kobayashi & Nishizawa 2012; Malhota et al. 2019). Citric acid has also already be tested as Fe-complexing agent in foliar application, with appropriate agronomic value for iron supplying. However, the iron: ligand stoichiometric ratio in foliar fertilization studies with organic acids is not clearly defined (Álvarez-Fernández et al. 2004; Chakraborty et al. 2014). The stoichiometric ratio between metals and ligands is an important variable that affects the stability and agronomic efficiency of the micronutrient complexed forms (Garcia-Mina 2006). Studies in the field of theoretical chemistry and electrospray mass spectrometry indicate that iron and organic acids, such as citrate and oxalate, are prone to form chemical species with 1:1 and 1:2 (metal: organic acid molar ratio) stoichiometric ratios, respectively (Bertoli et al. 2015a; Tolentino et al. 2015). Besides, Larbi et al. (2010) show that the apoplastic fluid of Fe- deficient plants contains Fe-citrate complexes of molar ratios 1:1 ([Fe-CitOH]-) and 1: 2 ([FeCit2] -3). Polynucleate Fe-citrate complexes at molar ratio 1:1 (Fe2Cit2 and Fe3Cit3) were also found in xylem sap of Fe deficient and Fe sufficient leaves (Rellán-Álvarez et al. 2010). The role played by natural complexes, a low-cost source, in delivering iron to plants is less studied than synthetic chelates. In the same way, the relation among complex stability, complexed fraction and iron foliar fertilization efficiency for both complexed and chelated sources is still unexplored (Abadía et al. 2011). Therefore, this study intended to investigate some chemical properties of Fe-complexes and verify if the complex stability and iron complexed fraction are associated with the efficiency of the foliar iron nutrition. Furthermore, this study explores if organic acids could be successfully used as iron complex agents in foliar fertilization, at two specific stoichiometric ratios (1:1 and 1:2 Fe: organic acid molar ratio). A chemical speciation software was employed to verify the Fe-complexed fraction and corresponding Fe-complex species for each tested ligand (EDTA, citric acid, malic acid, tartaric acid, and oxalic acid). Sequentially, theoretical calculations helped to estimate the chemical structure and the thermodynamic stability of the iron chemical species. Lastly, after the formulation of Fe-complexes in laboratory conditions, a greenhouse experiment was conducted to test the Fe-complexes capacity for iron supplying and redistribution in the foliar application, using maize as plant test. We hypothesized that the stability is a crucial factor determining the accumulation and redistribution of foliar-applied Fe-complexed sources. 2. Material and Methods 2.1 Foliar sprays solution preparation and analysis 50 The organic acids used as complexing agents for iron in the foliar sprays were: tricarboxylic citric acid (CA), dicarboxylic malic acid (MA), dicarboxylic tartaric acid (TA), and dicarboxylic oxalic acid (OA). The iron source used for Fe-complexes and Fe-EDTA preparation was FeSO4 (Fe II). All the compounds used were of analytical grade. Stock solutions of all Fe-organic acids and Fe-EDTA in 0.01 mol L-1 concentration were prepared. Ligands were dissolved in water and KOH (0.05 mol L-1) was used to adjust the solutions to pH 5.0-6.0. Subsequently, an amount of iron (as FeSO4), calculated to be 2% above the molar amount of the ligand, was then slowly added while maintaining the pH between 5.0 and 6.0 (Rodríguez-Lucena et al. 2010; Martín-Fernández et al. 2017). The stoichiometric molar ratios (SR) used for organic acid reactions with iron were 1:1 and 1:2 (Fe: organic acid). Fe-EDTA was prepared in the recommended 1:1 SR (Fernández et al. 2006) Solutions were aerated and left to stand overnight to allow the excess iron to precipitate (Steiner & Van Winden 1970). All chelate/complex solutions preparation and storage were carried out in the dark to avoid potential photodecomposition of chelates (Steiner & Van Winden 1970). In order to obtain a 0.005 mol L-1 Fe solutions for foliar spray (Rodríguez-Lucena et al. 2010), aliquots of Fe-organic acids or Fe-EDTA stock solutions were mixed with distilled water was added to complete 200 mL. After adjusting the solutions’ pH to 5.0 (Rodríguez-Lucena et al., 2010; Eichert & Fernández, 2011), the non-ionic Agral® surfactant (Syngenta) was added at the recommended rate of 0.5 mL L-1 to enhance the the foliar spray adherence. As a control, an uncomplexed source solution using FeSO4 was prepared few minutes before the application, to avoid Fe2+ oxidation and precipitation. Additional solutions of 0.005 mol L-1 Fe were made in triplicate to measure the concentration of soluble iron. The solutions were mixed and centrifuged for precipitation of the insoluble material. The supernatant was collected and the iron concentration was determined using the atomic absorption technique. The measured soluble iron concentration was then divided by the total iron concentration (0.005 mol L-1) to determine soluble iron fraction. For Fourier Transform Infra-Red (FTIR) analysis, a separate batch of the iron complex foliar spray solutions was freeze-dried by lyophilization to obtain solid samples. The samples were analyzed by a Cary 630 Agilent® equipment containing a ZnSe crystal in an ATR (Attenuated Total Reflectance) configuration. The scans were recorded between 650 and 4000 cm-1 and at 4 cm-1 resolution. 2.2 Fe chemical species in the foliar spray solutions 51 The software Visual Minteq version 3.1 was used to estimate the proportion of iron complexed by the organic acids in the foliar spray solutions. The software determined the ionic strength after pH and temperature had been set at 5.0 and 25ºC, respectively. It was considered that only Fe(III) complexes were formed in the foliar spray solutions, due to the exposure to aeration and the fact that Fe(III)-organic acids complexes are more stable and preferably formed (Gomathl 2000; D’Antonio et al. 2009; Rodríguez-Lucena et al. 2010). The chemical species of the Minteq outputs were classified as Free-Fe or Complexed-Fe. Free-Fe represented the sum of ionic and non-complexed forms, such as Fe3+, FeOH2+, Fe2+, and FeSO4 (aq). Complexed-Fe represented the sum of all other forms of iron associated with the ligands (complexes), with negative, positive and neutral charges. Free-Fe and Complexed-Fe fractions were expressed as percentage of the total iron amount in the foliar spray solution. 2.3. Fe-complexes computational modeling The complexed species identified in the Minteq were used to computational geometry optimization and thermodynamic calculations. The geometry optimization of reagents (organic acids and iron) and complexes formed between iron and organic acids was obtained using the Gaussian 09W software. Initially, the probable geometries were input for both the ligands (with the farthest carboxyl groups) and the complexes (with three groups for CA; or two carboxyl groups in SR 1:1 for MA and TA; or four carboxyl groups in SR 1:2 for OA). Subsequently, using the B3LYP method and the LanL2DZ basis set, the geometries were optimized by the Density Functional Theory (DFT) (Bertoli et al. 2015b). The B3LYP/LanL2DZ is a suitable effective core potential for post-third row atoms, such as the metal iron (Bertoli et al. 2015b). The calculations were performed considering both the gas phase and water