Postharvest Biology and Technology 45 (2007) 38–45 Calibration transfer between NIR diode array and FT-NIR spectrophotometers for measuring the soluble solids contents of apple M a Carmen Alamar a,∗ , Els Bobelyn b , Jeroen Lammertyn c , Bart M. Nicola¨ ı b,c , Enrique Molt ´ o a a Instituto Valenciano de Investigaciones Agrarias, Centro de Agroingenier´ ıa, Ctra. Moncada-N´ aquera, Km. 4.5, 46113 Moncada-Valencia, Spain b Flanders Centre of Postharvest Technology, W. de Croylaan 42, B-3001 Leuven, Belgium c BIOSYST-MeBioS, Katholieke Universiteit Leuven, W. de Croylaan 42, B-3001 Leuven, Belgium Received 29 June 2006; accepted 20 January 2007 Abstract Prediction models of soluble solids contents of fruit obtained with a specific near-infrared spectrophotometer cannot be directly transferred to another spectrophotometer. In this research, a piecewise direct standardization method has been used for this task. As a result, a calibration model for soluble solids contents of apple developed on a Fourier transform based spectrophotometer has been successfully transferred to a diode array (DA) spectrophotometer. The standardization procedure was performed on a data set of 477 Jonagored apples and root mean squared error of prediction of 0.85 ◦ Brix was obtained. Additionally, the necessity of calibration transfer procedures between two DA spectrophotometers of the same type and model has also been found. © 2007 Elsevier B.V. All rights reserved. Keywords: Apple; Quality; Soluble solids content; NIR; Spectroscopy; Calibration transfer; PDS 1. Introduction The potential to measure quality attributes such as sugar content and firmness of intact fruit such as nectarines and peaches, apples and kiwifruit has been evaluated extensively dur- ing the last 10 years (Slaughter, 1995; Lammertyn et al., 1998; McGlone and Kawano, 1998; Ventura et al., 1998; Schaare and Fraser, 2000; McGlone et al., 2002). In NIR spectroscopy, NIR radiation is guided into the product, and some of the backscat- tered radiation is captured and related to variables of interest via multivariate statistical techniques. Because of their high acquisition speed, diode array systems are more suitable than Fourier Transform instruments to be mounted on high speed grading lines in such a way that, grading based on internal quality attributes rather than external appearance comes within reach. The inability of a calibration model developed on one instru- ment to be directly used on another one, even between two ∗ Corresponding author. Tel.: +34 96 342 40 00; fax: +34 96 342 40 01. devices of the same type, is a significant limitation of this technique. This seriously hampers widespread application as this would mean that the calibration model has to be con- structed again for every spectrophotometer. As other authors have pointed out before (Bouveresse and Massart, 1996a; Fearn, 2001; Feudale et al., 2002b) there are three main causes which introduce variation in newly recorded spectra that have not been considered in the calibration of the equipment: (i) changes in physical and/or chemical compounds of the samples, (ii) changes in the instrumental response function (different instrument, age- ing of sources, replacement of some parts, etc.) and (iii) changes in the environment of the instrument over time (temperature, humidity). In order to avoid repetition of the whole calibration proce- dure, which is normally expensive and time consuming, different calibration transfer techniques (also known as instrumental stan- dardization) have been developed over the past years. The term standardization indicates the set of mathematical and statistical methods used to transfer the calibration model developed on a master (or primary) instrument to another instrument (slave or secondary). 0925-5214/$ – see front matter © 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.postharvbio.2007.01.008