The rising environmental concerns due to the excessive usage of synthetic, non-renewable, recalcitrant petroleum-based materials have driven the research for sustainable, renewable and biodegradable products. Cellulose is one of the most abundant natural polymers with attractive properties such as renewability, potentially low cost, biocompatibility, and biodegradability. It is abundantly available as part of lignocellulosic biomass, such as forest residues, agricultural residues, and energy crops. A great deal of research is focused on cellulose extraction from such sustainable sources and its applications in various eco-friendly products (Satlewal et al., 2018a; Satlewal et al., 2018b).

With the evolution in the field of nanotechnology and the emergence of new and advanced nanomaterials, the traditional research practices and many industrial applications are now getting reformed. Nanocellulose has the desirable properties of cellulose such as biodegradability, low density and non-toxicity; along with certain special properties such as great mechanical and thermal stability, self-assembly in aqueous dispersion media, reinforcing capabilities and long-lasting nature (Thomas et al., 2018). It has been applied in various applications such as food packaging, chemical industry as viscosity modifier, biomedicine, mechanical reinforcement of matrices, and membrane filtration (Squinca et al., 2020). With the decrease in size from micrometres such as in cellulose polymer chains to nanometres such as in nanocellulose, various parameters such as effective surface area, crystallinity, and mechanical strength change. These variations can be analysed using different characterization techniques.

As per the Technical Association of the Pulp and Paper Industry (TAPPI), nanocellulose could be broadly categorized into: 1) cellulose nanocrystals (CNCs) and 2) cellulose nanofibrils (CNFs) (Shak et al., 2018). Nanocellulose extraction primarily involves a combination of enzymatic, chemical or mechanical treatments to attain purified cellulose. This cellulose is then subjected to chemical processes typically by acid hydrolysis to separate amorphous domains to obtain CNCs, whereas CNFs are obtained via mechanical treatments (Eapen et al., 2020). A detailed comparison of the properties of CNCs and CNFs isolated from different feedstocks and methods is discussed in literature (Borjesson and Westman, 2015; Shak et al., 2018; Trache et al., 2020) and in the present review.

In this review, various sources and types of nanocellulose with emphasis on different production methods such as pre-treatment, mechanical and chemical methods with recent literature are presented to build an understanding of how the cellulose fibers disintegrate into primary constituents nanocellulose (Agrawal et al., 2021a; Agrawal et al., 2021b). This review also discusses the unique properties of nanocellulose and general awareness for eco-friendly products that have directed both research and industrial attention towards nanocellulose and its applications. Moreover, nanocellulose (both CNCs and CNFs) are presently mostly confined to lab-scale due to scalability issues, high energy demand, time consuming process, and high production costs (Trache et al., 2020). Environmental impacts associated with higher energy demands and chemical usage pose an additional limitation (Arvidsson et al., 2015). Performing life cycle assessment (LCA) and techno-economic assessment (TEA) at early development stages of a production routes could address these challenges. Such assessment studies help in identification of the process improvement opportunities to improve the material and energy efficiency. Even though a few LCA studies are available in public domain, TEA studies for CNCs and particularly CNFs are quite limited (de Assis et al., 2017; de Assis et al., 2017; Bondancia et al., 2020; Blair et al., 2021) This work reviews these existing studies and summarizes the key learning and research gaps.

Cellulose is the most abundant renewable compound found in the biosphere with an estimated production 1,012 tonnes annually by photosynthesis (Sundarraj and Ranganathan, 2018). It is an unbranched polysaccharide composed of D-glucose units linked via β (1, 4) glycosidic linkage. Every monomer unit in chain is twisted at 180° compared to its neighbours. The repeating unit of cellulose is a disaccharide cellobiose (Gardner and Blackwell, 1974). Cellobiose units are linked together to produce long homo polymer cellulose chains. Adjacent cellulose chains are linked by hydrogen bonding and van der Waals forces during biosynthesis promote stacking of the chains into elementary fibrils. Numerous elementary fibrils are further bundled together to produce micro-fibrils, which in turn are bundled to form macro-fibrils or cellulosic fibers. Figure 1 shows the hierarchical structure of cellulose from bundle of cellulose fibers to single microfibril and its microstructure. Native cellulose fibers are composed of both ordered (crystalline) and disordered (amorphous) domains. Its crystallinity can range from 40 to 70% varying on the basis of the source and the extraction procedure. The amorphous regions are preferably broken down by physical, chemical, physicochemical and enzymatic treatments. The degree of polymerization of native cellulose varies from 1,500 to 3,500 depending on source of cellulose and the treatment used for separation (Hallac and Ragauskas, 2011).

Nanocellulose is defined as the cellulose molecules with at least one dimension in nanoscale (1–100 nm). Development in the emerging field of nanotechnology has made nanocellulose a more attractive research subject than its cellulose counterpart. The main reason for popularity of nanocellulose is that a uniform material with enhanced physicochemical and mechanical properties and crystallinity can be formed by simply reducing the size of cellulose fiber (Suryanegara et al, 2010). Each elementary units of cellulose is strong; when assembled into a fiber, the defects in the assembly make the fiber weaker and theoretical strength of combination of elements in a fiber is poor. So when nanoelements are extracted they show better inherent properties than parent fiber.

Generally, nanocellulose is extracted from cellulose, which is present in various sources such as hardwood, soft wood, agricultural biomass, bacteria, algae, tunicates etc., (Jonoobi et. al., 2015). The cellulose composition is same in all sources, however the structure might show slight variations. The cellulose extracted from plants species have about 13–22 μm diameter, and about 44–65% crystallinity (Ververis et al., 2004), while the cellulose synthetised by bacteria called bacterial cellulose, are naturally nanoscale fibrils, with a diameter of about 10–100 nm and with a much higher crystallinity of about 90% (Gorgieva and Trček, 2019).

Typically, cellulose nanostructured materials refer to three major classes: 1) cellulose nanofibers 2) cellulose nanocrystals/cellulose whiskers 3) bacterial nanocellulose. Cellulose nanofibrils are nanoscale cellulose with diameter of about 5–60 nm and length in many microns (Habibi et al., 2010). They are often extracted from wood pulp by TEMPO-mediated oxidation method (Jonasson et al., 2020). The fibrils consist of both crystalline and amorphous sections. On the other hand, cellulose nanocrystals are needle shaped and highly crystalline with some residual amorphous regions. The acid hydrolysis method used for obtaining the nanocrystals preferably dissolves the amorphous part, yielding cellulose nanocrystals with 5 nm diameter and 20–100 nm length (Phanthong et al., 2018). The bacterial nanocellulose is ribbon shaped with about 20–100 nm size and length in many micrometers (Jonoobi et al., 2015). Bacterial nanocellulose is secreted as an extracellular product during bacterial fermentation by many species of the genera Rhizobium, Agrobacterium, Gluconacetobacter and Sarcina (Gupta and Shukla, 2020). The main advantage of bacterial nanocellulose is that it does not have other unwanted polymers linked to it, unlike cellulose from plant sources. The type of nanocellulose, its size, morphology and other properties depend on the cellulose source, processing conditions during isolation and any pre- or post-treatments. Table 1 shows the comparison of different types of nanocellulose.

Nanocellulose is an interesting material owing to its unique inherent properties like high mechanical strength, thermal stability, high surface area (250 m g⁻¹), high aspect ratio, interesting optical and rheological properties and ease of surface modification (Barbash et al., 2017; Daud and Lee 2017). Various analytical techniques are used for characterization and better understanding of nanocellulose properties as shown in Figure 2.

The surface charge of nanocellulose can be determined by zeta potential value (Yang et al., 2017). The negative values of zeta potential indicate the negatively charged hydroxyl groups (˗OH) on nanocellulose surface. The dispersion stability can also be determined from the zeta potential values. Usually the absolute value less than 15 mV represents the onset of agglomeration, while the absolute value greater than 30 mV results in a stable nanocellulose suspension (Mohaiyiddin et al., 2016). Yang et al. (2017) showed that nanocellulose derived by TEMPO oxidisation (Zeta potential = -69 mV) and acid hydrolysis (Zeta potential = −34 mV) formed stable suspension, while high intensity ultra sonication derived nanocellulose (Zeta potential = −24 mV) formed agglomeration in suspension.

The hydroxyl groups on the surface of nanocellulose have strong interacting ability and contribute to self-association. This property makes nanocellulose a good reinforcing material in composite matrix of polymers with ten-fold to hundred-fold mechanical strength (Luo et al., 2017), and enhanced optical transparency (Su et al., 2017) as compared to polymer matrix without nanocellulose loading.

Crystallinity index of nanocellulose determines its microstructural and mechanical properties. It is measured using X-ray diffractometer and expressed as crystallinity index (CrI), which is calculated according to Eq. 1 (Segal et al., 1959). CrI % = [ I   002 − I   a m ] I   002   ×   100   % (1) where, I002 is the maximum diffraction intensity (the 002 plane), located around 20 = 22.5° and corresponds to the crystalline region; Iam is the minimum diffraction intensity located around 20 = 18° and corresponds to the amorphous material.

Microscopic techniques such as Field Emission Scanning Electron Microscopy (FESEM), Transmission Electron Microscopy (TEM) and Atomic Force Microscopy (AFM) can be used for dimensional analysis of nanocellulose. The magnified microscopic images captured using AFM, cryo SEM and TEM showed that all the nanocellulose fibrils and crystals possess right handled chirality at fibril level and at bundle level (Usov et al., 2015). Atomic Force Microscopy (AFM) can be used to probe the topography (Barbash et al., 2019), morphology (Alonso-Lerma et al., 2020), and mechanical properties of nanocellulose. The transverse elastic modulus of wood derived CNCs calculated using AFM was in the range of 18–50 GPa (Lahiji et al., 2010) and that of CNF was 6.9 ± 0.4 GPa (Parvej et al., 2020). It is a favoured technique among researchers due to the attainable resolution and high degree of adaptability. It can be employed to investigate the exact same area of a sample in different environments to see how the local sample changes when exposed to different environments. For instance, the quantitative height difference of cotton cellulose nanofibers in dry and wet state recorded using AFM was used to determine swelling and shrinking, indicated by positive and negative values, respectively (Ottesen and Syverud, 2021).

The hydrodynamic size can be determined by Dynamic Light Scattering (DLS). It gives information about the apparent diameter, and though it is not very accurate for non-spherical particles, it can be used in combination with other techniques to track CNC consistency and change in size over due time (Marway, 2017). The density and porosity of nanocellulose can be measured using a helium pycnometer. Daicho et al. (2020) reported the density of pure native celluloses, such cotton and tunicin, are approximately 1.60 g/cm³ using helium pycnometer. In order to study the elemental composition on nanocellulose surface, X-ray photoelectron spectroscopy (XPS) technique could be used (Jordan et al., 2019). In cellulosic materials, four categories of carbon bonds can be identified by XPS: C1 carbons bonded to other carbons or hydrogen (C–C, C–H); C2 carbons bonded to one oxygen atom (C–O); C3 carbons attached to two oxygen atoms or a carbonyl group (C=O, O–C–O) and C4 carbons from carboxyl groups (O-C=O) (Bastidas et al, 2005). Dasong (2015) studied the contribution of each carbon bond in hemp yarn. C1, C2, C3 and C4 were reported 35.76, 35.69, 18.99, 9.56%, respectively; while for the nanocellulose, the contribution of C1, C2, C3 and C4 were 32.66%, 28.67%, 18.73%, 19.94%. For nanocellulose, the significant decrease of C2 (C-O) and significant increase of C4 (O-C=O) indicated that parts of the hydroxyl groups in cellulose chains were oxidized into carboxyl groups during the oxidation/sonication process. XPS is also a useful technique to study the various functional groups inserted onto nanocellulose surface during different methods of extraction (Khili et al., 2019; Xu and Gu, 2020).

For determining the thermal stability of nanocellulose, various techniques such as thermogravimetric analysis (TGA), differential scanning calorimetry (DSC) and differential thermogravimetry (DTG) can be used (Jordan et al., 2019). TGA of microcrystalline cellulose (MCC) and nanocrystalline cellulose (NCC) obtained from almond and peanut shells, indicated stable materials with a degradation temperature reaching 240 and 200°C for MCC and NCC, respectively (Khili et al., 2019). Thermal decomposition temperature for nanocellulose is usually around 200–300°C, however, thermal stability can be increased by physical or chemical modifications such as desulfation of cellulose nanocrystals extracted by sulphuric acid hydrolysis (Gan et al., 2020).

NMR can be used to evaluate the purity and molecular structure of nanocellulose. Solid state 13C NMR spectroscopy can provide information about the conformation of the constituent carbon atoms of cellulose. In the NMR spectra, the C4 and C6 carbon atoms are distinguishable into crystalline and noncrystalline signals. The signal ratio of C4 crystalline and noncrystalline phases can be used for calculation of crystallinity index (CI) value of nanocellulose (Daicho et al., 2020). However, nanocellulose tends to agglomerate while drying, therefore solid-state characterization may reflect certain structural changes from drying and not necessarily the precise nature of the individual nanocellulose (Jiang et al., 2014). Therefore, solution-state NMR could be a useful tool for surface characterization of individual nanocellulose in never-dried state. TEMPO oxidation liberates cellulose nanofibrils containing some surface carboxyls converted from C6 primary hydroxyls. These structural modifications were displayed by a weakened signal at C6 (δ 60) and a new peak at δ175, in the solution-state 13C NMR spectra (Jiang et al., 2014). However, almost no C6 and H6 peaks were detected indicating that only the surface nuclei could be detected by NMR of the nanocellulose suspensions (Jiang et al., 2014).

The study of rheological properties of nanocellulose aqueous suspension is important as it can have important consequences on the preparation, processing and combination with other materials for industrial applications. The shape, size and surface composition of nanocellulose mainly affects the rheological properties of nanocellulose (Shafiei-Sabet et al., 2013). measured the shear viscosity of CNC dispersions as a function of the shear rate. At a concentration of 1 wt% and shear rates between 0.1 and 1 s, the measured viscosity was found to be 10–100 mPas. Common ways to access the viscosity of nanocellulose suspension is by using Couette-type rheometer, extensional viscometer, cone and plate rotational viscometer and rolling ball viscometer; or alternate methods that avoid wall depletion effects such as serrated plate system and vane system (Hubbe et al., 2017).

These broad and versatile properties of nanocellulose originate because of the size and chemistry introduced by scaling down to nanometer (Thomas et al., 2018). Due to these attractive properties, nanocellulose can have diverse range of applications that can solve major challenges of sustainable society.

In this section, various methods of nanocellulose preparation are summarized.

Nanocellulose is an alternative biomaterial to synthetic materials that are environmentally friendly, and biodegradable. Recently, cellulose nanocomposites are utilized in medical, automotive, electronics, packaging, construction, and wastewater treatment applications (Figure 6).

Environmental impacts associated with nanocellulose production vary widely based on feedstock, technology, and end products. Bio-based raw material is not a sufficient condition to ensure that the nanocellulose production process is sustainable. Current technologies for nanocellulose production constitute multiple steps that are energy-intensive and employ harsh operating conditions (Dhali et al., 2021). For example, most research studies have employed bleached pulp obtained from chemical pulping and purification as feedstock to produce nanocellulose (Jiang et al., 2021). However, chemical pulping uses harsh chemicals and large quantities of water that will contribute adversely to environmental impacts. These impacts are not captured in conventional environmental impact assessment (EIA) studies. Life cycle assessment (LCA) is a holistic approach that quantifies the direct and indirect environmental impacts by taking a cradle-to-grave scope. The LCA methodology is formalized through the ISO 14040 framework and is being increasingly used to assess novel products and processes. Here, important insights gained from LCA studies in literature are summarized.

For CNCs, energy demand was reported to significantly affect the overall impacts, greenhouse gas (GHG) emissions, in particular. Energy demands of CNCs production from different feedstocks and production routes varied from 457 to 15,943 MJ, and corresponding global warming potential (GWP) ranged from 12.9 to 1,086 kg CO² eq./kg CNC (Zargar et al., 2021). De Figueirêdo et al. (2012) employed acid hydrolysis of two different feedstocks i.e., cotton fibers and coconut fibers for CNC production. Energy consumption for coconut fibers was reported to be an order of magnitude higher at 16,000 J/kg CNC compared to that for cotton fibers (1800 J/kg CNC production). Consequently, associated GHG impacts for coconut fibers and cotton fibers were 1,086 and 122 kg CO² eq./kg CNC, respectively. In addition to being highly energy intensive, CNCs isolation involves chemicals that lead to other environmental impacts (Trache et al., 2020). For CNCs, although sulfuric acid is conventionally utilized for acid hydrolysis, developing cost-competitive production routes that employ organic acids can be a way forward (Bondancia et al., 2020). However, the weak nature and higher pKa value mean that organic acids are required in larger quantities as compared to mineral acids. This can completely offset the benefit. Optimizing the acid selection, quantity and designing the acid/chemical recovery system is important. Additionally, enzyme-mediated isolation that requires mild conditions and a potential environmentally benign option should be further explored (Arantes et al., 2020). The influence of feedstock on energy demand and GHG impacts was demonstrated in the by De Figueirêdo et al. (2012).

For CNFs, energy demands in the literature ranged from 87 to 19,000 MJ, and the GWP was reported to be between 0.79 and 800 kg CO² eq./kg CNF (Zargar et al., 2021). High energy demands of CNF production from mechanical treatments pose a major concern, even though these methods are suitable for scaling. In this context, chemical and enzymatic pretreatment have been proposed to overcome the critical challenges of employing mechanical treatment alone (Negi et al., 2019). However, a scenario with chemical pretreatment resulted in higher impacts in the categories of global warming potential, water depletion, cumulative energy demand, and terrestrial acidification (Arvidsson et al., 2015). In comparison, enzymatic pretreatment reduced the energy usage by 64% compared to a scenario without any pretreatment. Emissions associated with enzymatic pretreatment reduced to 0.79 kg CO² eq./kg CNF compared to chemical pretreatment with 99 kg CO² eq./kg CNF. Another potential alternative could be employing mild treatments to produce lignin-containing cellulose nanomaterials (Ewulonu et al., 2019; Zargar et al., 2021).

Due to the lack of pilot plant or industrial-scale data, lab-scale data of nanocelluloses production has often been used in the environmental assessment of nanocelluloses. Even though early assessment of lab-scale data is essential to identify the process hotspots, results from such studies cannot be competitive with the nanocellulose counterparts produced at an industrial scale with optimized and well-established processes (Piccinno et al., 2018). Particularly for the production of CNFs that is energy-intensive, several authors reported significant overestimation of electricity consumed for lab-scale (Bartolozzi et al., 2020; Gallo Stampino et al., 2021). Sensitivity analysis of electricity consumption revealed that though results did not vary qualitatively, impacts were drastically affected from a quantitative perspective. Therefore, in the absence of pilot-scale data, a scale-up framework (Piccinno et al., 2015; Piccinno et al., 2018; Bartolozzi et al., 2020) has to be adopted to model the industrial scale nanocellulose production.

Many LCA studies in the literature overlooked the environmental and biological toxicity impacts of nanocellulose (Li et al., 2013; Jiang et al., 2021). Scope of these LCA studies should consider the environmental release and human exposure of nanocelluloses in the inventory modeling for completeness. However, lack of appropriate characterization factors specific to nanomaterials that determine toxicity-related potential impacts poses an additional challenge (Salieri et al., 2018). To this end, environmental risk assessment studies (Lazarevic and Finnveden, 2013; Jiang et al., 2021) and development of specific characterisation factors for ecotoxicity impacts should be at the forefront of future studies. Thus, extending the scope of LCA from “cradle to gate” approach to “cradle to grave” can provide a better estimation of the overall impacts associated with nanocellulose based materials. A comprehensive LCA study should also take credit for avoided impacts. In this approach, the impacts for the production of conventional product that the nanocellulose is replacing are deducted, giving a net difference.

Production cost depends on the pretreatment method, acid used for hydrolysis, and the feedstock cost. Production costs reported in literature varies widely from 2,000 to 10,000 USD/ton (de Assis et al., 2017; Jiang et al., 2018 ; Blair et al., 2021). CNC production via conventional sulfuric acid hydrolysis mandates acid-resistant equipment that increases the capital expenditure (CAPEX). Moreover, acid recovery via neutralization and effluent treatment further increases the CAPEX. A study on CNCs reported that the manufacturing costs for sulfuric acid hydrolysis varied from 3,632 to 4,420 USD/dry ton (de Assis et al., 2017). In all scenarios under consideration, dissolving pulp was reported as the primary cost driver contributing to 38–45% of manufacturing cost. For a configuration with no acid recovery, acid and lime costs accounted for 24–28% of manufacturing cost. Upon acid recovery, this contribution drastically decreased to 3–4% of manufacturing cost. The minimum selling price (MSP) of CNC was majorly influenced by CAPEX and the cost required for dissolving pulp. An increase in CAPEX due to acid recovery increased the MSP from 4,829 to 5,125 USD/dry ton (de Assis et al., 2017). A recent TEA study reported equipment cost contributed to 65% of CAPEX, with chemical recovery section specifically accounting for 22% of CAPEX (Blair et al., 2021). This study reported a relatively lower MSP at 1,656 USD/ton, explained by the lower feedstock handling cost. This production route benefitted from the less energy-intensive mechanical processing coupled with SO2-ethanol-water pulping. Additionally, employing wood chips as feedstock rather than chemical pulp can benefit the process economics due to co-product credits. In another study that reported CNCs obtained via citric acid hydrolysis, a better alternative than sulfuric acid from an environmental perspective, the MSP was relatively higher at 16,460 USD/dry ton (Bondancia et al., 2020). CNC production using citric acid hydrolysis displayed a relatively lower capital cost at seven million USD. Whereas, operating costs were higher at 8.25 million USD, with citric acid alone contributing to 40% of this cost even at 89% acid recovery. It was estimated that for a 79% decrease in citric acid market price i.e., at 290 USD/ton, MSP of CNC produced using citric acid and sulfuric acid would be equivalent. Upon observing the MSP with a chemical recovery section and citric acid hydrolysis, it is noteworthy that there is a trade-off between economic and environmental aspects of nanocellulose production.

Currently, nanocelluloses are not cost-effective compared to their alternatives due to low production capacities. Shen et al. (2020) discussed the factors inhibiting the commercialization of nanocellulose reinforced polymer composites. Authors reported that pretreatment costs and low production capacities associated with nanocelluloses resulted in substantially higher production costs compared to the alternatives such as nano-montmorillonite and nanosilica. In this context, it is essential to achieve mass production through process optimization for nanocelluloses commercialization. Moreover, the lack of standardized product index due to inconsistent properties of nanocelluloses produced from comprehensive selection of sources and production routes hinders the nanocelluloses marketization.

To effectively address both the environmental and economic concerns, enzyme-mediated production of nanocelluloses can offer a potential solution. However, it is to be noted that commercial enzymes are often optimized to convert cellulose and hemicellulose fractions to fermentable sugars. These issues demand further research on optimizing enzyme cocktails for nanocellulose production (Arantes et al., 2020) and the on-site production of enzymes to increase the economic viability (Squinca et al., 2020). Most studies on nanocellulose production employed bleached chemical pulp or purified cellulose that already lost 50–60% of the biomass content for nanocellulose production (Jiang et al., 2021). This practice results in less efficient utilization of biomass due to reduced nanocellulose yields with respect to original biomass. In comparison, mechanical pulp can serve as an efficient alternative and needs further investigation. Furthermore, research efforts have to be focused on optimizing the reaction yield that will improve the overall economics of the production route. Moreover, developing technologies that are feedstock agnostic can play a crucial role in enhancing the production scale. In such a scenario, economies of scale can be leveraged to achieve cost-competitive production. To achieve sustainable nanocellulose production, routes with mild reaction conditions and single-step extraction processes should be the focus of future research studies.

Nano cellulose is renewable, non-toxic, environmentally friendly material and recent technologies make it suitable for creating alternatives options in the field of packaging products, biomedical, composites, paper making, water purification, and cosmetics.

The effective design of nano cellulose products suitable for different fields depends on the identification of problems and resolving them. Further, this showed that extensive research is required on the utilization of nano cellulose in various other fields as well. The commercial manufacturer of nanocelluloses is dependent on the different end-user requirements and implementation of the product would have encouraging future in terms of expanding global technology development.

With the emerging market for nanocellulose, nanocellulose production such as CNCs and CNFs can benefit from the early-stage environmental and economic assessments. These studies can be instrumental in identifying the hotspots and opportunities for optimization of process energy and chemicals consumption. It is mandatory to address the challenges posed by conventional methods such as the energy-intensive nature of mechanical pretreatment, effluent handling and recovery system for acid hydrolysis, and higher reaction times for enzyme-mediated treatment. In this context, research gaps to be addressed by LCA practitioners and technology developers have been suggested. In addition to eco-toxicology studies of nanocellulose, pretreatment and isolation methods that are scalable, environmentally friendly, and cost-effective have to be investigated further.