Analysis of the oxypropylation process of a lignocellulosic material, almond shell, using the response surface methodology (RSM)

Highlights

• Almond shell, was oxypropylated for the first time to yield liquid polyols.

• RSM was used for the first time to optimize the oxypropylation process.

• RSM with restricted responses was used to target polyols with specific properties.

• RSM is a useful decision-making tool in complex processes, e.g. oxypropylation.

Abstract

Developing polyols from abundant and renewable biomass resources is an important topic for polymer synthesis. In this work, the Response Surface Methodology (RSM) was applied to a novel oxypropylation case study, almond shell (AS), an agroindustry lignocellulosic by-product. Mathematical models were developed to determine responses maximizing the reaction efficiency to yield polyols with specific technical requirements (polyols suitable for rigid polyurethane foams; hydroxyl index between 300 and 800 mgKOH/g, and viscosity below 300 Pa*s). In a general way, the properties of the obtained polyols were within the range of the ones currently used commercially, reinforcing the interest to exploit lignocellulosic bio-residues for polyol synthesis. For simultaneous minimization of homopolymer content and unreacted biomass, values of 14.0% and 14.1% were achieved, respectively. This was attained using a formulation with an AS/PO ratio of 20.1/79.9 g/ml and a catalyst content of 3.14%, giving rise to a polyol with an hydroxyl index of 392.1 KOH/g and a viscosity of 107.4 Pa*s. Overall, the advantages of using RSM to better understand complex reactive systems and the interest to use these statistical approaches as decision-making tools was demonstrated.

Introduction

The almond is a drupe, quite produced in Portugal, in particular in the regions of Trás-os-Montes e Alto Douro, Alentejo and Algarve. In Portugal, and according to INE (Instituto Nacional de Estatística), in 2011, the almond harvested area was 26,900 ha, and the annual production was 7700 tonnes. Thence, the production increased reaching, in 2018, 39,642 ha of cultivated area and 21,642 tonnes of annual production (INE, 2018). The world production is led by Northern America (1,872,500 tonnes), with a total contribution of 59%, followed by Europe with 15% (475,513 tonnes). Spain, with a world production share of 11% (339,033 tonnes), stands out as the dominant European producer (FAOSTAT, 2020). In this context, and considering that almond fruit accounts with 80% (w/w) of residues (hull, shell, and skin) (Esfahlan et al., 2010; Estevinho et al., 2008), approximately 2,546,321 tonnes of these by-products can be generated, and are potentially available to develop novel applications.

Fig. 1 gives an overview of the typical operations carried out in the almond processing industry, associated primary and secondary products, as well as the generated by-products and potential applications. Some are traditional ones, but others are proposed in light of some existing studies (Holtman et al., 2015; Rosa et al., 2015; Valdés et al., 2015), and having in view the integrated valorisation of the almond processing industry value-chain. Aside from the hull, which is discarded in harvesting, shell and skin residues can be recovered as by-products of the almond processing industry. Moreover, the shell can be used as energy source being commercialized at the cost of 0.11 €/kg (data provided by the local producer Amendouro - Comércio e Indústria de Frutos Secos, Lda, headquartered in Alfândega da Fé, Portugal).

Face to the status of the almond cultivation and processing industry where residues can be recovered concentrated, and the fact that they achieve no, or only modest values, the proposal of new applications, namely in the field of polymeric materials are of particular interest (Rosa et al., 2015). One strategy is the use of the oxypropylation process to obtain polyols, both liquid (total oxypropylation) or biphasic (partial oxypropylation). These polyols can be used to produce polyurethanes and polyesters (Bernardini et al., 2015; Cateto et al., 2013) and one-source composites, (De Menezes et al., 2007; Rosa et al., 2014), respectively.

During the past 40 years, oxypropylation has been extensively applied. Several types of biomass residues have been tested to add value to some industrial by-products. Examples include lignin (Cateto et al., 2009; Nadji et al., 2005), sodium lignosulfonates (Oliveira et al., 2015), sugar beet pulp (Pavier and Gandini, 2000), cork powder residues (Evtiouguina et al., 2002), olive stone (Matos et al., 2010), rapeseed cake (Serrano et al., 2010), and soy hulls (Rosa et al., 2015). Other biomasses, namely gambier tannin (Arbenz and Avérous, 2015), chitin and chitosan (Fernandes et al., 2007), bark (D’Souza et al., 2015), date seeds (Briones et al., 2011), and Brazilian pine fruit shell (de Rezende et al., 2018), were also studied. All these examples reported the synthesis of the polyols directly from the selected biomass, or by using a biomass pre-treatment to facilitate the impregnation of the catalyst. More recently the strategy to apply biomass fractionation was also attempted (Li et al., 2020). Another common feature is the absence of any optimization studies, only the work of de Rezende et al. (2018) provided a systematic analysis based on Principal Components Analysis (PCA), which in a context of both biomass and polyol variability is of high interest. Apart from a better understanding the oxypropylation process itself, the optimal conditions to achieve polyols with target properties can be determined, which is of industrial relevance.

The efficiency of the oxypropylation reaction varies according to several variables and operating conditions (Aniceto et al., 2012; Cateto et al., 2009). The optimal reaction conditions may not be generalized to all biomasses due to the diverse nature of their constituents (e.g. cork residues versus sugar beet pulp). Moreover, lignin, cellulose, and hemicellulose contents, and the crystalline organization, can differ, which may limit reagent’s access. Another critical factor is the hydroxyl index (IOH, mg KOH/g); a biomass with a high IOH value requires higher amounts of reactants, namely catalyst content and more severe reaction conditions, to favour the reaction of the hydroxyls entrapped inside the biomass structure with propylene oxide. Therefore, owing to biomass variability, the selection, and optimization of the operating conditions for the oxypropylation reaction is an important topic to be studied.

In this context, one-factor-at-a-time approaches, classical methods in process optimization, can be used. Nevertheless, these simple methods cannot either accurately predict the optimal operating conditions and interactions between variables, nor provide a complete understanding of variable’s patterns. One-factor-at-a-time approaches are laborious methods that, frequently, do not guarantee the reaching of the optimal conditions (Box and Hunter, 1957; Intergovernmental Panel on Climate Change, 2005). Also, carrying out experiments with all variable’s combinations is impractical due to a large number of required trials (De Lean et al., 1978; Heleno et al., 2016). These limitations can be surpassed by employing the response surface methodology (RSM), which is a method specially designed to optimize responses as a function of multiple variables by requiring minimum data.

In this work, the production of biopolyols from the almond shell, a lignocellulosic agro-industrial residue, was studied, and the RSM methodology used to optimize the oxypropylation process. The chosen independent variables were the almond shell to propylene oxide ratio and the catalyst content. In terms of response variables, the hydroxyl index, homopolymer content, unreacted biomass, and viscosity were chosen. Polyol requirements were defined as the ones needed for rigid polyurethane foam preparation (IOH between 300 and 800 mgKOH/g and viscosity below 300 Pa*s).