The first two decades of the twenty first century witnessed the emergence of sustainable plant-based lifestyle, with food scientists focussing their attention on creating meat and dairy alternates that supplant animal-based ingredients with sustainable and inexpensive plant-based sources (Lam, Can Karaca, Tyler, & Nickerson, 2018; Pietrysiak, Smith, Smith, & Ganjyal, 2018; Stone, Karalash, Tyler, Warkentin, & Nickerson, 2015). Flavour is one of the most important attributes that determines whether foods are accepted or rejected by consumers (Su et al., 2020), yet flavour perception is complex and is influenced by many factors such as flavour type, concentration and release. However, plant-based proteins tend to have an undesirable beany, green or grassy off-flavour (Shi and others, 2021; Schindler et al., 2012) that limit their application especially in products with a bland taste profile, such as non-dairy beverages. Poor water solubility of plant-based proteins is another major issue that influences both protein digestibility and functional properties such as emulsification, foaming, gelation, water and oil-holding capacities (Sashikala, Sreerama, Pratape, & Narasimha, 2015; Aryee, Agyei, & Udenigwe, 2018; Hou and others 2018; Kudre, Bejjanki, Kanwate, & Sakhare, 2018; Lan, Chen, & Rao, 2018). These properties influence food texture, processability and consumer acceptance (Can Karaca, Nickerson, & Low, 2011). These poor organoleptic and functional properties necessitate energy-intensive thermal treatments such as direct steam injection for ameliorating the flavour and functionality of plant-based proteins (Chao, Jung, & Aluko, 2018; Lan et al., 2018; Saint-Eve, Granda, Legay, Cuvelier, & Delarue, 2019). Exploration of less energy intensive hybrid methods to improve plant protein flavour and functionality is necessary to expand the global plant-based protein market.

The food matrix plays a large role in flavour perception as various constituents such as proteins, lipids and carbohydrates can interact with flavour compounds (Wang & Arntfield, 2017). Pea proteins, that are increasingly being used as a hypoallergenic plant-based source of proteins, are known to have a very notable off-flavour, characterized by beany, vegetative, green or grassy notes (Schindler et al., 2012; Damodaran and Arora, 2013). These flavour perceptions are known to be influenced by flavour type, concentration and release, and are detected by humans via the orthonasal (nose) and retronasal (mouth) olfaction (Roland, Pouvreau, Curran, Velde, & Kok, 2017; Trikusuma, Paravisin, & Peterson, 2020; Wang & Arntfield, 2017). n-hexanal is the principle odour-active compound that attributes this undesirable grassy or green off-flavour to pea protein (Yen and Pratap-Singh, 2020; Shi and others, 2021; Schindler et al., 2012; Murat, Bard, Dhalleine, & Cayot, 2013; Ma, Boye, Azarnia, & Simpson, 2016; Roland et al., 2017; Fahmi, Ryland, Sopiwnyk, & Aliani, 2019; Zha, Dong, Rao, & Chen, 2019; Trikusuma et al., 2020). Other than n-hexanal, volatile compounds such as 1-pentanol, 1-hexanol, 1-octen-3-ol and 3,5-octadien-2-ones have also been reported to contribute to the overall aroma profile of pea proteins (Yen and Pratap-Singh, 2020; Shi and others, 2021; Lan, Xu, Ohm, Chen, & Rao, 2019; Ma et al., 2016; Murat et al., 2013; Roland et al., 2017; Schindler et al., 2012; Trikusuma et al., 2020; Zha et al., 2019). n-hexanal and n-hexanol are known to augment lipid oxidation (oxidation of linoleic acid) that plays an important role in the changes in overall flavour and aroma of pea protein incorporated food products (Schindler et al., 2012). Although, methods relying on volatile content measurement and odour activity values, like those developed by Singh et al. (2021), could be used to study the flavour perception using gas chromatography coupled with mass spectrometry (GC–MS); however, these methods have limitations. First, not all volatile compounds may be odour-active (Murat et al., 2013). Even if a volatile compound is odour-active, it may not be detectable, let alone recognizable, depending on the threshold (Rowe 2004). Volatile compound data may not be representative to how a person perceives the product; for instance, the rate of volatile release, matrix effect and flavour balance are factors that may affect perception (Baek and others 1999). Therefore these findings must be validated with sensory evaluation (Stone and Sidel 2004). In sensory evaluation, descriptive analysis (DA) is a powerful tool where trained panelists determine the magnitude of perceived sensory attributes in food products using methods such quantitative descriptive analysis (QDA) or another widely used SpectrumTM method (Aguiar, Melo, & Lacerda de Oliveira, 2019; Stone, Bleibaum, & Thomas, 2012).

Vacuum microwave dehydration (VMD) is an emerging technology that uses a combination of vacuum and microwaves for dehydration of food products (Mohammadi et al., 2020; Landymore et al., 2019; Mandal and Pratap-Singh, 2021). Microwaves produce volumetric heating that results in lower energy requirements and shorter processing times with low thermal damage to food products (Drulyte & Orlien, 2019). The reduced pressure facilitates the removal of volatile compounds at lower temperatures due to the reduction of boiling points under vacuum (Michailidis & Krokida, 2014). Such a method could be used to remove volatiles from pea protein isolates with little effect on its' functionality (Yen and Pratap-Singh, 2020). Although microwaves are known to impact hydrophobicity and disulphide bonds that influence a protein's folding pattern and tertiary structure stability (Liu et al., 2016), a low microwave power could be used to minimize such adversary effects. Pratap-Singh et al. (2020) have demonstrated the application of microwave-vacuum dehydration as a sustainable technology to stabilize food industry waste from craft brewing industries. To the best of our knowledge, no study has investigated the functionality or volatile composition removal efficiency of vacuum microwave dehydrated proteins along with sensory analysis of the treated proteins, to investigate the potential of VMD as a plant-protein pre-treatment technology.

Earlier, Yen and Pratap-Singh (2020) reported significant reduction of the volatile content at the cost of the protein solubility in their preliminary work on employing VMD as a potential protein off-flavour removal technique. Yet, their work did not include a comprehensive analysis of the effect of VMD on protein functionality and sensory perception. The overall objective of this research was to comprehensively understand how various VMD parameters affect protein functionality, volatile composition and sensory perception. The overall hypothesis of the work is that VMD will help in reducing volatile composition to improve sensory perception while producing minimum effects on protein functionality. First, the VMD process was optimized based on the volatile profile and protein functionality. Subsequently, pea protein based dairy alternatives formulated with optimized VMD treatment were subjected to descriptive sensory analysis.