Rapid screening of apple juice quality using ultraviolet, visible, and near infrared spectroscopy and chemometrics: A comparative study

https://doi.org/10.1016/j.microc.2021.106051Get rights and content

Highlights

  • Chemometric models for predicting quality of apple juices from spectra.

  • Partial least square (PLS) regression for data analysis.

  • Model optimization using pre-processing and variable selection.

  • SSC, TA, pH predicted using NIR spectra of fruit (R2P > 0.8).

  • SSC, TA, SSC/TA, pH, TPC predicted from spectra of juice (R2P > 0.9).

Abstract

The aim of this study was to evaluate and compare potential of different optical spectroscopic techniques for quality assessment of apple juices. The calibration partial least squares (PLS) regression models were developed and optimized for determination of quality parameters of apple juice non-destructively from the NIR spectra of fruit or directly from the UV-VIS-NIR spectra of juices. The external models validation was performed using the test sample set. Calibration models based on the NIR spectra of fruit were characterized by high predictive ability for soluble solids content, SSC (R2P = 0.93, RPDP = 3.5), moderate for titratable acidity, TA (R2P = 0.81, RPDP = 2.1), pH (R2P = 0.82, RPDP = 2.1) and ratio of SSC to TA, SSC/TA (R2P = 0.72, RPDP = 1.8), and low for total phenolic content, TPC (R2P = 0.44, RPDP = 1.3) and total flavonoid content, TFC (R2P = 0.49, RPDP = 1.2). The performance of PLS models based on the juice spectra was dependent on the spectral region used for analysis. The models with the highest predictive ability for SSC (R2P = 0.95, RPDP = 5.0) and TA (R2P = 0.96, RPDP = 4.6) were obtained using UV-VIS-NIR spectra of juice. The models with the highest predictive ability for SSC/TA (R2P = 0.89, RPDP = 3.0) and pH (R2P = 0.91, RPDP = 3.3) were obtained using VIS-NIR spectra. The optimal model for TPC (R2P = 0.95, RPDP = 4.1) was obtained using UV-VIS spectra, and for TFC using UV spectra of juices (R2P = 0.80, RPDP = 2.5). The results may contribute to the development of method for fast quality control of juices.

Introduction

The methods used to routinely assess the quality of fruit and fruit products should enable rapid measurements of a large number of samples, preferably in a non-destructive manner, and in the place where they occur. In order to ensure effective control of the quality of fruits and their products, traditional analytical methods are replaced by spectroscopic techniques coupled with chemometric analysis [1]. The main advantage of these methods is the ability to perform quick measurements of spectra in a non-destructive way.

Spectroscopic techniques are based on the study of the interaction of electromagnetic radiation with matter. Optical spectroscopy involving infrared, visible and ultraviolet radiation is often used in food studies [2]. Using different ranges of electromagnetic radiation provides information about the different properties of samples.

Near infrared (NIR) spectroscopy (780–2500 nm) most often used in practical applications is vibrational spectroscopy [1]. The spectra in this range contain bands coming from overtones and combination tones of basic vibrations of groups having a hydrogen atom, mainly C–H, O–H and N–H chemical bonds. Thus, the NIR spectra enables study of organic major and minor food constituents, containing these structures. The NIR spectra may be measured directly for samples in different forms without the need for dilution or other preparation.

The electronic spectroscopy in the UV (200–380 nm) and VIS (380–780 nm) range is associated with the excitation of electrons forming chemical bonds in the molecule [3]. The bands in the UV-VIS spectra correspond to different chromophores presented in foods such as carbonyl groups, nitro groups, double and triple bonds, conjugated double bonds, etc. The absorption in VIS region, responsible for food color, is due to the presence of pigments with conjugated double bonds systems.

The use of these spectroscopic techniques for quality assessment is based on the development of multivariate calibration models. Such models describe the relationship between spectra and determined properties of samples. Partial least squares (PLS) regression is most commonly used for models development [1]. They are developed based on the analysis of training (calibration) samples for which the spectra and properties of interest are known. Optimized and validated models are used in routine analysis to determine the chemical or physical properties of new samples. Based on the measurement of one spectrum, it is possible to determine all the sample parameters for which calibration models have been developed.

Apples are very popular and widely consumed fruit, being a relevant source of dietary fibre minerals, vitamins, phytochemicals, including phenolics. The consumption of apples and apple products might be linked to a reduced risk of several forms of cancer, cardiovascular disease, and asthma [4].

Many studies have demonstrated the applicability of NIR or NIR combined with VIS spectroscopy for non-destructive determination of quality of apple fruit. A lot of studies concerned optimization of a prediction of a soluble solids content [5], [6], [7], [8]. The quantitative calibration models with acceptable performance predictive ability were reported for determination of parameters, such as: total acidity, total sugars, pH, ascorbic acid contents, total polyphenols, starch index, chlorophyll content, firmness, mealiness, and taste attributes [9], [10], [11], [12].

The use of NIR spectroscopy for whole fruit testing is particularly attractive due to the possibility of non-destructive measurements. This technique can also be useful for testing juice. Several studies have shown the possibility of using NIR spectroscopy for apple juice analysis, particularly to predict sugar content in apple juice [13], detect adulteration [14], differentiate between the juices based on heat treatment and apple varieties [15], detect deterioration of apple juice during storage and heating [16] and determine quality parameters of commercial apple juice [17]. Only few papers reported application of UV or UV-VIS spectroscopy for determination quality parameters of apple juice [18], [19].

In routine analysis of fruit juices, usually only selected parameters are controlled, including soluble solids content (SSC) and titratable acidity (TA), related to the sensory attributes and stability of juice. The ratio of soluble solids content to titratable acidity (SSC/TA) is also determined as an index of sensory acceptability of the fruit taste [10]. Important constituents of apple juice are phenolic compounds. They affect organoleptic properties of juice: their color, flavor, and astringency [20]. Phenolics due to their antioxidant properties account for the health benefits to humans related to consumption of apple juice [4].

In traditional analysis, the determination of each of these parameters requires the use of a separate analytical method. Some of these methods are time consuming, require sample preparation and use of reagents. Thus, in routine quality control, application of optical spectroscopic techniques, which enable fast and simple, simultaneous determination of several parameters may be valuable alternative to traditional, laboratory methods.

The objective of this study was to explore the feasibility and compare performance of different optical spectroscopic techniques to determine the quality parameters of apple juices from fruit and juice spectra. The calibration models were developed and optimized based on NIR spectra of intact fruit and individual UV, VIS, and NIR spectra recorded for juices and their fusion (UV-VIS, VIS-NIR, UV-NIR, UV-VIS-NIR).

Section snippets

Apples and apple juice

The apples (Malus × domestica Borkh.) were obtained during one season, from orchards located in Greater Poland region of Poland. In total, 105 fruit samples from nine cultivars were studied. The sample set consisted of 5 samples of Bankroft, 5 samples of Cortland, 5 samples of Empire, 20 samples of Idared, 5 samples of Jonagold, 20 samples of Jonathan, 20 samples of Lobo, 5 samples of McIntosh, and 20 samples of Zimnoje cultivars.

The apples were washed under running water and dried. The juice

NIR spectra of apple fruit

The spectra of fruit and juices were recorded for 105 apple samples originated from nine apple cultivars. Firstly, the NIR spectra were recorded for each of the fruit. Next, the juice was squeezed from each fruit. The UV-VIS-NIR spectra were measured for all juice samples and chemical parameters were determined using standard laboratory methods.

The mean spectra of apple fruit from various cultivars studied are shown in Fig. 2A.

The NIR spectra of the entire set of apple fruits show absorption

Conclusions

The calibration models for the determination of the several quality parameters of apple juice were developed and optimized. In most cases selection of variables improved the predictive ability of models. The models based on analysis of NIR spectra of whole fruit measured in non-destructive way, allowed determination of SSC with good prediction accuracy and determination of TA and pH with moderate accuracy. Juice quality parameters may be determined directly from their spectra. In this case

CRediT authorship contribution statement

Katarzyna Włodarska: Conceptualization, Methodology, Investigation. Paweł Piasecki: Software, Formal analysis. Ana Lobo-Prieto: Investigation. Katarzyna Pawlak-Lemańska: Methodology, Investigation. Tomasz Górecki: Validation. Ewa Sikorska: Conceptualization, Methodology, Writing - original draft, Supervision, Funding acquisition.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgement

Grant 2016/23/B/NZ9/03591 from the National Science Centre, Poland, is gratefully acknowledged.

References (40)

  • E. Borràs et al.

    Data fusion methodologies for food and beverage authentication and quality assessment – A review

    Anal. Chim. Acta.

    (2015)
  • J. Engel et al.

    Breaking with trends in pre-processing?

    TrAC Trends Anal. Chem.

    (2013)
  • R.P. Feliciano et al.

    Characterization of traditional and exotic apple varieties from Portugal. Part 1 – Nutritional, phytochemical and sensory evaluation

    J. Funct. Foods.

    (2010)
  • J. Wu et al.

    Chemical compositional characterization of some apple cultivars

    Food Chem.

    (2007)
  • E. de Rijke et al.

    Analytical separation and detection methods for flavonoids

    J. Chromatogr. A.

    (2006)
  • S. Abasi et al.

    Rapid measurement of apple quality parameters using wavelet de-noising transform with Vis/NIR analysis

    Sci. Hortic.

    (2019)
  • G. Giovanelli et al.

    NIR spectroscopy for the optimization of postharvest apple management

    Postharvest Biol. Technol.

    (2014)
  • M. Schmutzler et al.

    Simultaneous detection of total antioxidant capacity and total soluble solids content by Fourier transform near-infrared (FT-NIR) spectroscopy: A quick and sensitive method for on-site analyses of apples

    Food Control.

    (2016)
  • H. Wang et al.

    Fruit quality evaluation using spectroscopy technology: A review

    Sensors.

    (2015)
  • M.H. Penner, Ultraviolet, visible, and fluorescence spectroscopy, in: S. Nielsen (Ed.), Food Anal., Springer, New York...
  • Cited by (11)

    • Determination of total phenolic content and selected phenolic compounds in sweet wines by fluorescence spectroscopy and multivariate calibration

      2022, Microchemical Journal
      Citation Excerpt :

      The best RPD value was 3.8 for Tokaj selection wine (Table 2). Other authors reported RPD values of 2.3, 2.7 and 3.8 for grape vinegar [17], apple juice [18] and wine juice [16]. The PLS models using all variables (1022 or 1000) of unfolded EEM fluorescence spectra provided a good predictive accuracy expressed by RPD values > 2.5, which was a better result compared to the full PLS models of UV–vis spectra.

    • Quantitative analysis of multi-optical length NIR spectroscopy based on quaternion parallel feature extraction method

      2021, Infrared Physics and Technology
      Citation Excerpt :

      Near infrared spectroscopy (NIR) is one of the most powerful and non-destructive analytical tools, which has been widely applied to quantitative analysis in many fields [1,2].

    View all citing articles on Scopus
    View full text