Near-infrared Transmission Spectroscopy Research Articles

Near-infrared transmittance spectroscopy for radiochemical ageing of EPDM.

The feasibility of using near-infrared spectroscopy as a sensitive technique to follow the influence of gamma-irradiation upon ageing of different EPDM (ethylene propylene diene monomer) elastomers has been evaluated. Although identification is difficult, differences can be observed between the non-irradiated and irradiated materials for total integrated doses from 50 to 450 kGy using a dose rate of 1 kGy h(-1) under an oxygen flow. The decrease in intensity of bands at 7040, 4610 and 4910 cm(-1) are linked to the disappearance of additives present in the elastomer such as excess of vulcanising or antioxidant agents and occur for the lowest irradiation dose. This disappearance is confirmed by TGA (thermogravimetric analysis). The increase in band intensities assigned to the formation of hydroxyl and carbonyl groups (5100, 4860 and 4670 cm(-1)) irradiation indicates an increase in oxidation with irradiation in the presence of oxygen. No bands linked to the presence of C=C from the diene have been detected, probably owing to the low concentration in the material and the weak intensity in near-infrared region. For strong irradiation doses (450 kGy), the three formulations studied show no difference in their NIR spectra, which is confirmed by the TGA of these irradiated materials. PCA performed at 5000-4600 cm(-1) or 7090-6980 cm(-1) shows efficient discrimination.

Determination of Maximum Viscosity of Milled Rice Flours Using Near-Infrared Transmittance Spectroscopy.

The objective of this study was to develop a partial least squares regression (PLS) calibration method of maximum viscosity determination of Japanese milled rice flours using near-infrared transmittance (NIT) spectroscopy. The diversity of spectra and maximum viscosity of wide ranging of rice subfamilies were much more than those of japonica type rices. The variations of spectra and maximum viscosity were found to influence PLS loading weights. C-H and O-H in ROH and H2O absorbances presented by the loading weights were significant in the 8th loading of the PLS model for japonica type rices. The performance of this PLS calibration model (11 components) for maximum viscosity of a rapid visco analyser (RVA) was the standard error of prediction (SEP) of 17.7, square of regression coefficient (R2) of 0.75 and the ratio of the SEP to the standard deviation of the original data (RPD) of 1.9. This method can be applied to the determination of maximum viscosity of japonica type rices.

Variation within and between F2-derived families for grain yield and barley malting quality

Variation within and between F2-derived families for grain yield and malting quality was investigated using F4 breeding lines derived from F2 families of 4 barley crosses. The variation between F2-derived families was greater than within F2-derived families for grain yield and all malting quality attributes. Superior segregates almost exclusively came from the best performing families. The greater similarity of lines eventually drawn from an F2-derived family has significant implications for selection strategies in barley breeding programs as it facilitates the early discard of F2-derived families. To maximise the exploitation of genetic variation as early as possible, selection for malting quality could start in the F2 generation using near infrared transmittance (NIT) spectroscopy and for grain yield in the F3 generation.

Near-Infrared Turbidity of β-FeOOH Particle Suspensions

Near-infrared transmission spectroscopy can be complicated by the light scattering from heterogeneous materials. For the examination of an evolving system exhibiting such light scattering, transmission spectra near wavenumber v = 104cm−1were obtained during the hydrolysis of FeCl3solutions. At first, the resulting turbid suspension of cigar-shaped β-FeOOH particles exhibits single-particle scattering, including a Rayleigh regime (attenuation ∞ v4). At later times, the scattering increases strongly as the particles aggregate, and becomes proportional to vα, with α ≈ 2, consistent with scattering models that interpret the structure of aggregates in terms of a fractal dimension dfroughly equal to 2. In all cases investigated, the attenuation due to scattering is spectrally smooth and increases monotonically with wavenumber. It can be written in the simple form vα with 1 ≤ α ≤ 4. While over limited spectral ranges α may be taken independent of v, over wide ranges it decreases with increasing v. This behavior is consistent with the theoretical limits of α = 4 at v = 0, and α = 0 at v = ∞. Overall, the results suggest that a useful form for simulating scattering backgrounds in near-infrared spectroscopy is Avα, with A and α fitted constants.

Near-infrared transmittance spectroscopy: a potential tool for non-destructive determination of oil content in groundnuts

The oil content of 64 samples of groundnut kernels, representing 47 genotypes, was determined by the conventional Soxhlet extraction procedure (OilSOX). The values of OilSOX ranged from 403 to 536 g kg−1. The optical densities (ODs) of these samples were determined at 12 wavelengths (918, 928, 940, 950, 968, 975, 985, 998, 1010, 1023, 1037 and 1045 nm) in the near-infrared (NIR) region using a food composition analyser (essentially a filter-based NIR spectrophotometer). The instrument also recorded the temperatures of the sample (TempS) and the air (TempA) surrounding it. A sample holder (75 mm × 150 mm; optical path length 25 mm) was used for optical density measurement. The data obtained were subjected to multiple linear regression analysis using the ODs at 12 wavelengths, TempS and TempA as the independent (predictor) variables, and OilSOX as the dependent variable. The multiple linear regression equation comprising 14 predictors showed a significant relationship between predicted values of oil content (OilNIR) and OilSOX. The standard error of calibration and the coefficient of determination for calibration (R) were 3.54 and 0.821 respectively, while the standard error of prediction and the coefficient of determination for prediction were 5.82 and 0.865 respectively. The ratio of standard error of calibration and standard error of prediction was 0.608. The differences between OilSOX and OilNIR were less than ± 20 g kg−1 for samples having oil contents in the range from 480 to 510 g kg−1. However, for samples having OilSOX lower than 480 g kg−1 or higher than 510 g kg−1, differences greater than ± 20 g kg−1 were observed. There exists scope for further refining the regression equation by using a larger number of samples for generating optical data. The results demonstrated the potential of NIR transmittance spectroscopy for determining the oil content of groundnuts in a non-destructive manner. © 2000 Society of Chemical Industry

Barley malting quality and yield interrelationships and the effect on yield distribution of selection for malting quality in the early generations

Relationships between malting quality attributes and grain yield in segregating populations can profoundly influence the intensity and sequence of trait selection. Consequently, the interrelationships between malting quality parameters predicted by near infrared transmittance (NIT) spectroscopy, grain weight, and grain yield in unselected populations of F2, F3, and F4 breeding lines from 4 barley crosses were examined. The simple and partial correlations between malt extract, protein content, and diastatic power were similar to those reported in previous studies except for a positive correlation between malt extract and diastatic power in the F2 and F3 generations. This positive relationship should enhance selection for improved malting quality in breeding programs. There were no relationships between grain yield and malting quality attributes, which would have an adverse impact on the intensity and sequence of trait selection. The effect of F2 and F3 selection for malting quality on F3 and F4 yield distributions was estimated by comparing the F3 and F4 yield distributions of the entire unselected population with those for the selected populations. Individual selection and sequential independent selection in the F2 and F3 generation for malting quality parameters predicted by NIT spectroscopy and grain weight in 4 crosses generally did not alter the nature of the subsequent yield distributions, yet 78–90% of lines could be discarded and there still existed adequate genetic gain for grain yield in the retained population of potentially good malting quality lines. These results indicate that barley breeders could use NIT spectroscopy to efficiently select in the early generations for malting quality prior to the conduct of yield testing and obtain good genetic gain for both malting quality and grain yield.

Near-Infrared Spectroscopy for in situ Cure Monitoring of Dimethacrylate-Based Networks

In this study, near-infrared transmission spectroscopy is investigated as an alternative technique for the real-time cure monitoring of thermosets systems. Two polymer systems based on dimethacrylate/styrene and dimethacrylate/divinylbenzene monomers were studied in order to check the influence of comonomer functionality. A new device utilizing ITO (indium-tin-oxide)-coated glass plates was developed to take into account the specificity of chain polymerization: oxygen inhibition, bulk polymerization, etc. Two absorption bands at 6166 and 6135 cm−1 suitable for in situ monitoring of methacrylate and vinyl double bonds were identified, and the molar absorption coefficient was determined. A quantitative analysis allows us to calculate the double-bond conversion during the polymerization in isothermal conditions between 60 and 80 °C.

Evaluation of Measurement Sites for Noninvasive Blood Glucose Sensing with Near-Infrared Transmission Spectroscopy

Six putative measurement sites were evaluated for noninvasive sensing of blood glucose by first-overtone near-infrared spectroscopy. The cheek, lower lip, upper lip, nasal septum, tongue, and webbing tissue between the thumb and forefinger were examined. These sites were evaluated on the basis of their chemical and physical properties as they pertain to the noninvasive measurement of glucose. Critical features included the effective optical pathlength of aqueous material within the tissue and the percentage of body fat within the optical path. Aqueous optical paths of 5 mm are required to measure clinically relevant concentrations of glucose in the first-overtone region. All of the tested sites met this requirement. The percentage of body fat affects the signal-to-noise ratio of the measurement and must be minimized for reliable glucose sensing. The webbing tissue contains a considerable amount of fat tissue and is clearly the worse measurement site. All other sites possess substantially less fat, with the least amount of fat in tongue tissue. For this reason, the tongue provides spectra with the highest signal-to-noise ratio and is, therefore, the site of choice on the basis of spectral quality.

Prediction of Starch Amylose Content Versus Total Grain Amylose Content in Corn by Near‐Infrared Transmittance Spectroscopy

ABSTRACTA study was conducted to investigate methods of improving a near‐infrared transmittance spectroscopy (NITS) amylose calibration that could serve as a rapid, nondestructive alternative to traditional methods for determining amylose content in corn. Calibrations were developed using a set of genotypes possessing endosperm mutations in single‐ and double‐mutant combinations ranging in starch‐amylose content (SAC) from ‐8.5 to 76%, relative to a standard curve. The influence of three factors were examined including comparing calibrations made against SAC versus grain amylose content (GAC), developing calibrations using partial least squares (PLS) analysis versus artificial neural networking (ANN), and using all samples in the calibrations set versus using progressively narrower ranges of SAC or GAC in the calibration set. Grain samples were divided into calibration and validation sets for PLS analysis while samples used in ANN were assigned to a training set, test set, and validation set. Performance statistics of the validation sets that were considered were the coefficient of determination (R), the standard error of prediction (SEP), and the ratio of the standard deviation of amylose values to the SEP (RPD), which were used to compare all NITS models. The study revealed an NITS prediction model for SAC (R = 0.96, SEP = 5.1%, RDP = 3.8) of similar precision to the best GAC model (R = 0.96, SEP = 2.7%, RPD = 3.5). Narrowing the amylose range of the calibration set generally did not improve performance statistics except for PLS models for SAC in which a decrease in SEP values was observed. In one model, the SEP improved while R and RPD remained constant (R = 0.94, SEP = 4.2%, RPD = 2.8) when samples with SAC values <20% were removed from the calibration set. Although the NITS amylose calibrations in this study are of limited precision, they may be useful when a rough screening method is needed for SAC. For example, NITS may be useful to detect severe contamination during transport and storage of specialty grains or to aid breeders when selecting for amylose content from large numbers of grain samples.

Near-infrared analysis of fat, protein, and casein in cow's milk.

Fat, crude protein, true protein, and casein were determined in cow milks by near-infrared transmission spectroscopy (NIR). Partial and overall PLS calibrations were performed on two sets of samples: partial calibration included 76 unhomogenized samples, whereas overall calibration used 96 homogenized and unhomogenized samples. Standard errors of calibration were 0.12% for fat, 0.06% for crude protein, 0.04% for true protein, and 0.05% for casein in the overall calibration. Validation of the overall calibration with an independent set of samples gave standard errors of prediction of 0. 07% for fat, 0.06% for crude protein and casein, and 0.05% for true protein. Except for fat, all of the statistical parameters were better with overall than with partial calibrations, which indicates that homogenization has an effect on NIR fat determination. Despite the relatively small number of samples included in the calibration model, NIR transmission was found to be a reliable method for the determination of fat and nitrogenous constituents in milk.

Determination of Apparent Amylose Content in Japanese Milled Rice Using Near-Infrared Transmittance Spectroscopy.

The objective of the present study was to develop a method to analyze apparent amylose content (AAC) of Japanese milled rices using near-infrared transmittance spectroscopy (NIT). Samples (n=110, varieties=37), harvested in 1996, were collected at various sites throughout Japan. Whole-grain milled rice was scanned using a near-infrared range (833-1050 nm with 8 nm steps and 27 wavelengths) transmittance filter type spectrometer. The AACs of samples were in the range of 13.1% to 20.7% (SD: 1.53). The wide range AAC (0-35.3%) partial least squares (PLS) model was found to be inadequate for accurate prediction of the narrow AAC range (13.2-20.7%) of the rice samples. The statistical performance of PLS modeling (11 factors) for narrow range AAC analyses gave a standard error of cross-validation (SECv) of 0.78 and square of regression coefficient (R2) of 0.74. The AAC model was applied to 20 unknown samples of products from different crop year (1997), and gave a standard error of prediction (SEP) of 1.25, R2 of 0.49 on the validation set. These results suggested that this model based on NIT spectroscopy could be applied for rapid and nondestructive measurement of narrow range AAC of Japanese milled rices.

DETERMINATION OF UNDRIED ROUGH RICE CONSTITUENT CONTENT USING NEAR-INFRARED TRANSMISSION SPECTROSCOPY

Near-infrared transmission (NIRT) spectroscopy was used in an attempt to predict moisture content, proteincontent and amylose content from undried whole-grain rough rice spectra. One hundred-fifty undried rough rice sampleswere collected. Using partial least squares calibration models obtained from undried whole-grain rough rice spectra, thecoefficient of determination (r 2) and the standard error of prediction (SEP) of the validation set were r 2 = 0.96 andSEP = 0.70 for rough rice moisture content, r 2 = 0.70 and SEP = 0.24 for brown rice protein content, r 2 = 0.76 andSEP = 0.22 for milled rice protein content, and r 2 = 0.00 and SEP = 0.27 for milled rice amylose content. The results ofthe validation indicated that NIRT could be used to determine moisture content and protein content. Thus, NIRTtechnology may be used to classify undried rough rice into qualitative groups such as high protein content rice and lowprotein content rice upon arrival at a rice-drying facility after harvesting.

Environmental and genetic variation for grain yield and barley malting quality attributes

Genotype, environment, and genotype × environment interaction effects for malting quality attributes and grain yield were investigated using breeding lines from the F2, F3, and F4 generations and the parental varieties of 4 barley crosses. There were significant differences between the parental varieties for all attributes studied. Both malting quality and grain yield exhibited a wide range among progenies in all generations. While performance of the parental varieties and progeny for malting quality and grain yield were greatly influenced by environment, performance in one environment was predictive of that in other environments. Only for grain protein content was there evidence of crossover G × E interaction. Heritability was generally higher for F3 to F4 than for F2 to F3 for all malting quality attributes. F3 on F2 regression per cent heritability estimates for protein content, potential malt extract and grain weight were all highly significant with values generally medium in magnitude. Genetic gain was obtained from selection in both the F2 and F3 generations. Heritability and genetic gain varied from cross to cross for diastatic power. Progress from selection for the other quality attributes attests to the potential value of NIT (near infrared transmittance) spectroscopy for predicting potential malting quality. Heritability for F2 to F3 for grain yield was not significant in any cross, indicating selection for yield on the basis of individual F2 plant yield was ineffective. Heritability for grain yield from F3 to F4 was highly significant and medium in magnitude for 3 of the 4 crosses. The results of this study indicate that good genetic gain could be expected from early generation selection for potential malting quality using NIT spectroscopy and for grain yield using F3 progeny testing.

Measurement of glucose and other analytes in undiluted human serum with near-infrared transmission spectroscopy

Near-infrared calibration models are described for the measurement of total protein, albumin protein, globulin protein, triglycerides, cholesterol, urea, glucose, and lactate. Spectra are collected in triplicate over the 5000–4000 cm −1 spectral range with a 2.5 mm optical path length for 242 undiluted human serum samples. Calibration models are generated for each analyte by performing partial least-squares (PLS) regression on raw and digitally filtered spectra. Models are optimized individually for each analyte by considering spectral range, number of model factors and width/position of a Gaussian shaped filter response function for a digital Fourier filter. Accurate measurements are demonstrated for each analyte except lactate which is below the detection limit under our experimental conditions. Standard error of prediction (SEP) for glucose is 23.3 mg/dl (1.29 mM). Relevance and stability of the glucose calibration model are examined by evaluating the accuracy of glucose predictions from 50 human serum samples collected on a modified spectrometer nineteen months after the calibration data were collected. In addition, these 50 subsequent samples were treated in a blind manner by withholding their glucose values until all predictions were complete. Results indicate a slight positive bias in the predictions corresponding to a minor instability in the model. This instability is likely due to changes in the spectrometer hardware. Nevertheless, the strong correlation between predicted and actual glucose levels in these blind samples strongly suggests that this calibration model is based on information particular to glucose.

Application of near-infrared transmittance spectroscopy in the determination of fat, protein and dry matter in Atlantic halibut fillet

Near infrared transmittance (NIT) spectroscopy (850–1048 nm) has been applied to the measurement of fat, protein and dry matter in wet homogenized Atlantic halibut ( Hippoglossus hippoglossus) fillet. A total of 155 fillet cuts (five sample sites) were selected from the start of the investigation (survey I) and 98 fillet cuts (one sample site) were also selected from eight fish groups six months later (survey II). The fish groups originated from a slaughtering quality experiment, where fish size and fat content in the feed had been varied. The multivariate calibration models were tested separately for each survey and after mixing the survey samples. The models were also tested after inclusion of sample temperature or fish weight, or after reduction to spectra of eight wavelengths. The selected fillet cuts from survey II gave calibration models with the best prediction abilities for fat and dry matter, whereas the combination of samples from both surveys gave the best prediction abilities for protein. The constituent ranges (w.w.) were: 10–123 g/kg fat, 165–274 g/kg protein and 234–335 g/kg dry matter. The partial least squares regression resulted in the following prediction errors, expressed as root-mean-square error of cross validation: 2.7 g/kg fat, 5.2 g/kg protein and 4.2 g/kg dry matter. The temperature adjustment had positive influence on the test set validation of protein and dry matter, whereas the fish-weight adjustment had positive influence on the test set validation of fat. Calibration models based on eight wavelengths were not optimal, but showed promising results, applicable in a simpler instrument for future online quality monitoring of Atlantic halibut fillet.

Discrimination of ethylene/vinyl acetate copolymers with different composition and prediction of the content of vinyl acetate in the copolymers and their melting points by near-infrared spectroscopy and chemometrics

Near-infrared (NIR) diffuse reflectance spectra have been measured by use of a rotating drawer for pellets of 12 kinds of ethylene/vinyl acetate (EVA) copolymers with vinyl acetate (VA, the comonomer) varying in the 7–44 wt % range. They are unambiguously discriminated from one another by a score plot of the principal component analysis (PCA) Factor 1 and 2, based upon the NIR spectra pretreated by multiplicative scatter correction (MSC). Principal component (PC) weight loadings for Factor 1 show that the discrimination relies largely upon bands due to the overtone and combination modes arising from the VA unit. We have found one “outlier” in the score plot and elucidated its spectral characteristics based upon PC weight loadings for Factor 2. Partial least-squares (PLS) regression has been applied to propose calibration models which predict the VA content in EVA. The models have been prepared for three kinds of pretreatment, the first derivative, the second derivative, and MSC; and four kinds of wavelength regions. The NIR spectra in the 1100–2200 nm region after the MSC treatment has given the best correlation coefficient and standard error of prediction (SEP) of 0.998 and 0.70%, respectively. The calibration models, prepared by NIR diffuse reflectance spectroscopy for the pellet samples, are compared with previously reported models by NIR transmission spectroscopy for the flowing molten samples, and with those by Raman spectroscopy for the pellet samples. PLS regression has also allowed us to predict melting points of the copolymers with the correlation coefficient and SEP of 0.997 and 0.78°C, respectively. © 1998 John Wiley & Sons, Inc. J Polym Sci B: Polym Phys 36: 1529–1537, 1998

Discrimination of ethylene/vinyl acetate copolymers with different composition and prediction of the content of vinyl acetate in the copolymers and their melting points by near‐infrared spectroscopy and chemometrics

Near-infrared (NIR) diffuse reflectance spectra have been measured by use of a rotating drawer for pellets of 12 kinds of ethylene/vinyl acetate (EVA) copolymers with vinyl acetate (VA, the comonomer) varying in the 7–44 wt % range. They are unambiguously discriminated from one another by a score plot of the principal component analysis (PCA) Factor 1 and 2, based upon the NIR spectra pretreated by multiplicative scatter correction (MSC). Principal component (PC) weight loadings for Factor 1 show that the discrimination relies largely upon bands due to the overtone and combination modes arising from the VA unit. We have found one “outlier” in the score plot and elucidated its spectral characteristics based upon PC weight loadings for Factor 2. Partial least-squares (PLS) regression has been applied to propose calibration models which predict the VA content in EVA. The models have been prepared for three kinds of pretreatment, the first derivative, the second derivative, and MSC; and four kinds of wavelength regions. The NIR spectra in the 1100–2200 nm region after the MSC treatment has given the best correlation coefficient and standard error of prediction (SEP) of 0.998 and 0.70%, respectively. The calibration models, prepared by NIR diffuse reflectance spectroscopy for the pellet samples, are compared with previously reported models by NIR transmission spectroscopy for the flowing molten samples, and with those by Raman spectroscopy for the pellet samples. PLS regression has also allowed us to predict melting points of the copolymers with the correlation coefficient and SEP of 0.997 and 0.78°C, respectively. © 1998 John Wiley & Sons, Inc. J Polym Sci B: Polym Phys 36: 1529–1537, 1998

Classification of High Acid Satsuma Mandarins by Near Infrared Transmittance Spectroscopy.

On-line measurement of sugar content in satsuma mandarins was already achieved using NIR transmittance spectroscopy. The feasibility of simultaneous measurement of citric acid content was investigated. The citric acid content determined by titration with NaOH and the second derivative absorption values autoscaled in the 710-930 nm region were analyzed by PLS1 using Unscrambler software. The spectra of peeled fruits measured by an NIR Systems Model 6250 were analyzed, and the model composed of 12 factors indicated the highest accuracy; R was 0.93, the mean residual (Bias) was -0.013% and the standard error of prediction (SEP) was 0.146%. The citric acid content in intact satsuma mandarins was regressed by the same method using an on-line instrument. The model composed of 5 factors calibrated from 689 samples showed the highest accuracy; R=0.83, Bias=0.024% and SEP=0.147% as a prediction result from 548 samples. It was possible to classify nondestructively the high acid fruits using near infrared (NIR) transmittance spectroscopy at about 20% error rate.

Evaluating Techniques for Rice Grain Quality Using near Infrared Transmission Spectroscopy

The development of advanced evaluation techniques for rice quality has been a desire of the Japanese rice industry (breeding, distribution and processing). The objective of the present study is to develop novel techniques for evaluating rice grain quality. A reliable determination method for amylose in whole grain rice using near infrared transmission (NIT) is proposed, using Partial Least Squares (PLS) regression analysis. It was suggested from results based on two different validation methods that the PLS models have possibilities for determination of apparent amylose content using NIT spectroscopy. PLS modelling for constituents important in rice quality indicates that reasonably accurate models are attainable for moisture content and protein content in whole grain rice. However our PLS models were not sufficiently accurate for physical rice quality (head rice ratio, apparent density, whiteness) using NIT spectroscopy.

Whole Grain Amylose Analysis in Maize Using Near‐Infrared Transmittance Spectroscopy

ABSTRACTThe development of genetically modified starches has relied on the use of maize (Zea mays L.) endosperm mutant alleles that alter starch structural and physical properties. A rapid method for predicting amylose content would benefit breeders and commercial handlers of specialty starch corn. For this reason, a study was conducted to investigate the use of near‐infrared transmittance spectroscopy (NITS) as a rapid and nondestructive technique for predicting grain amylose content (GAC) in maize. Many single‐ and double‐mutant inbreds and hybrids were used to create a calibration set for the development of a predictive model using partial least squares analysis. A validation set composed of similar genetic material was used to test the prediction model. A coefficient of correlation (r) of 0.94 was observed between GAC values determined colorimetrically and those predicted by NITS; however, the predicted values were associated with a large standard error of prediction (SEP = 3.5). Overall, NITS discriminated well among high amylose and waxy genotypes. The NITS calibration was used to determine levels of contamination by normal kernels in waxy and high‐amylose (Amy VII) grain samples intended for wet milling. In both cases, a 5% contaminated sample could be detected from pure samples according to predicted NITS values.