Implementation of enhanced correlation maps in near infrared chemical images: Application in pharmaceutical research

Recent developments in Hyperspectral Imaging equipment have made possible the use of this analytical technique for fast scanning of sample surfaces. This technique has turned out to be especially useful in Pharmacy, where information about the distribution of the components in the surface of a tablet can be obtained. One particular application of Hyperspectral Chemical Imaging is the search for singularities inside pharmaceutical tablets, e.g. coating defects. Nevertheless, one problem has to be faced: how to analyze a sample without any previous knowledge about it, or having only the minimum information about the tablet.In this work a new methodology, based on correlation coefficients, is introduced to obtain valuable information about one Hyperspectral Image (detection of defects, punctual contaminants, etc.) without any previous knowledge. The methodology combines Principal Component Analysis (PCA), correlation coefficient between one specific pixel included in the image and the rest of the image; and a new enhanced contrast function to obtain more selective chemical and spatial information about the image. To illustrate the applicability of the proposed methodology, real tablets of ibuprofen have been studied.The proposed methodology is presented as a control technique to detect batch variability, defects in final tablets and punctual contaminants, being a potential supplementary tool for quality controls. In addition, the usefulness of the proposed methodology is not exclusive to NIR-CI devices, but to any hyperspectral and multivariate image system.     

Development of a rapid process monitoring method for dry-coated tableting process by using near-infrared spectroscopy

A nondestructive transmittance near-infrared (NIR) method for detecting off-centered cores in dry-coated (DC) tablets was developed as a monitoring system in the DC tableting process. Caffeine anhydrate was used as a core active pharmaceutical ingredient (API), and DC tablets were made by the direct compression method. NIR spectra were obtained from these intact DC tablets using the transmittance method. The reference assay was performed with HPLC. Calibration models were generated by partial least squares (PLS) regression and principal component regression (PCR) utilizing external validations. Hierarchical cluster analysis (HCA) of the results confirmed that NIR spectroscopy correctly detected off-centered cores in DC tablets. We formulated and used the Centering Index (CI) to evaluate the precision of core alignment and generated an NIR calibration model that could correctly predict this index. The principal component (PC) 1 loading vector of the final calibration model indicated that it could specifically detect the misalignment of tablet cores. The model also had good linearity and accuracy. The CIs of unknown sample tablets predicted by the final calibration model and those calculated through the HPLC analysis were closely parallel with each other. These results demonstrate the validity of the final calibration model and the utility of the transmittance NIR spectroscopic method developed in this study as a monitoring system in DC tableting process.     

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     

Application to classification of mulberry leaves using multivariate analysis of proton NMR metabolomic data

Recently, NMR-based metabolomic analysis has been used to acquire information based on differentiation among biological samples. In the present study, we examined whether multivariate analysis was able to be applied to natural products and/or material field. Each extraction of 24 leaf samples, divided into six locations from the tip of the stem in each of four strains, was analyzed by pattern recognition methods, known as Principal Component Analysis (PCA) and Soft Independent Modeling of Class Analogy (SIMCA). Twenty-four extracts from mulberry leaf showed independent spectra by 1H NMR. The separation of leaf extraction data due to the difference at six locations was achieved in the PCA score plot as correlation PC1 (86.1%) and PC3 (4.6%) and showed two loading plots, suggesting classification by leaf position as an independent variable in the loading plot. Moreover, the difference among six locations clarified the seven highest discrimination powers by the SIMCA method. Meanwhile, the PCA score plot obtained classification by the variety of mulberry strains with three loading plots, but the SIMCA method did not give a peak by classification.     

Theoretical analysis of tablet hardness prediction using chemoinformetric near-infrared spectroscopy.

In order to clarify the theoretical basis of the variability in the measurement of tablet hardness by compression pressure, NIR spectroscopic methods were used to predict tablet hardness of the formulations. Tablets (200 mg, 8 mm in diameter) consisting of berberine chloride, lactose, and potato starch were formed at various compression pressures (59, 78, 98, 127, 195 MPa). The hardness and the distribution of micropores were measured. The reflectance NIR spectra of various compressed tablets were used as a calibration set to establish a calibration model to predict tablet hardness by principal component regression (PCR) analysis. The distribution of micropores was shifted to a smaller pore size with increasing compression pressure. The total pore volume of tablets decreased as the compression pressure increased. The hardness increased as the compression pressure increased. The hardness could be predicted using a calibration model consisting of 7 principal components (PCs) obtained by PCR. The relationship between the predicted and the actual hardness values exhibited a straight line, an R(2) of 0.925. In order to understand the theoretical analysis (scientific background) of calibration models used to evaluate tablet hardness, the standard error of cross validation (SEV) values, the loading vectors of each PC and the regression vector were investigated. The result obtained with the calibration models for hardness suggested that the regression vector might involve physical and chemical factors. In contrast, the porosity could be predicted using a calibration model composed of 2 PCs. The relationship between the predicted and the actual total pore volume showed a straight line with R(2) = 0.801. The regression vector of the total pore volume might be due to physical factors.     

独立分量分析预处理法提高苹果糖度模型预测精度研究

为了提高苹果近红外光谱糖度预测模型精度,利用独立分量分析方法(ICA)对苹果近红外光谱进行了预处理,并且建立了糖度的偏最小二乘(PLS)预测模型。结果表明,独立分量分析不但能分离出噪声信号,而且所分离出来的光谱信号也比原始光谱信号光滑。在预处理后的最佳PLS糖度模型校正时的相关系数rc和标准偏差SEC分别为0.9549和0.3361,用于预测时的相关系数rp和标准偏差SEP分别为0.9071和0.4355。与普通的平均处理法的PLS模型相比,其精度有所提高,且模型更加简洁。     

Generalized Dice's coefficient‐based multi‐block principal component analysis with Bayesian inference for plant‐wide process monitoring

Plant‐wide process monitoring is challenging because of the complex relationships among numerous variables in modern industrial processes. The multi‐block process monitoring method is an efficient approach applied to plant‐wide processes. However, dividing the original space into subspaces remains an open issue. The loading matrix generated by principal component analysis (PCA) describes the correlation between original variables and extracted components and reveals the internal relations within the plant‐wide process. Thus, a multi‐block PCA method that constructs principal component (PC) sub‐blocks according to the generalized Dice coefficient of the loading matrix is proposed. The PCs corresponding to similar loading vectors are divided within the same sub‐block. Thus, the PCs in the same sub‐block share similar variational behavior for certain faults. This behavior improves the sensitivity of process monitoring in the sub‐block. A monitoring statistic T² corresponding to each sub‐block is produced and is integrated into the final probability index based on Bayesian inference. A corresponding contribution plot is also developed to identify the root cause. The superiority of the proposed method is demonstrated by two case studies: a numerical example and the Tennessee Eastman benchmark. Comparisons with other PCA‐based methods are also provided. Copyright © 2014 John Wiley & Sons, Ltd.     

Near infrared and Raman spectroscopy as Process Analytical Technology tools for the manufacturing of silicone-based drug reservoirs

Using near infrared (NIR) and Raman spectroscopy as PAT tools, 3 critical quality attributes of a silicone-based drug reservoir were studied. First, the Active Pharmaceutical Ingredient (API) homogeneity in the reservoir was evaluated using Raman spectroscopy (mapping): the API distribution within the industrial drug reservoirs was found to be homogeneous while API aggregates were detected in laboratory scale samples manufactured with a non optimal mixing process. Second, the crosslinking process of the reservoirs was monitored at different temperatures with NIR spectroscopy. Conformity tests and Principal Component Analysis (PCA) were performed on the collected data to find out the relation between the temperature and the time necessary to reach the crosslinking endpoints. An agreement was found between the conformity test results and the PCA results. Compared to the conformity test method, PCA had the advantage to discriminate the heating effect from the crosslinking effect occurring together during the monitored process. Therefore the 2 approaches were found to be complementary. Third, based on the HPLC reference method, a NIR model able to quantify the API in the drug reservoir was developed and thoroughly validated. Partial Least Squares (PLS) regression on the calibration set was performed to build prediction models of which the ability to quantify accurately was tested with the external validation set. The 1.2% Root Mean Squared Error of Prediction (RMSEP) of the NIR model indicated the global accuracy of the model. The accuracy profile based on tolerance intervals was used to generate a complete validation report. The 95% tolerance interval calculated on the validation results indicated that each future result will have a relative error below ±5% with a probability of at least 95%. In conclusion, 3 critical quality attributes of silicone-based drug reservoirs were quickly and efficiently evaluated by NIR and Raman spectroscopy.     

Quantitative Analysis of Uncoated Eszopiclone Tablets by Near-Infrared Spectroscopy

This study employed near-infrared (NIR) spectroscopy to analyze content uniformity, moisture content, compression force, tablet hardness, average particle size, and particle-size distribution. The content uniformity, moisture content, compression force, tablet hardness, and average particle size models yielded high correlation coefficients (R²) of 0.99582, 0.99725, 0.99620, 0.96294, and 0.98421, respectively, whereas the particle size distribution models showed good predictive ability. Conventional criteria such as R², root-mean-square error of calibration, and the root-mean-square error of prediction were used to evaluate the accuracy and precision of the model. To ensure the accuracy and predictability of the content model for low-dose tablets, additional validation and reliability evaluations were performed using 70%, 80%, 100%, 120%, and 130% drug concentrations as well as 90% and 110% active content formulations. Near-infrared spectroscopy with multivariate modeling is a rapid, nondestructive technique for the characterization of the manufacturing process.     

Comparison of near infrared and microwave resonance sensors for at-line moisture determination in powders and tablets

In this paper we demonstrate the feasibility of replacing KF for water content testing in bulk powders and tablets with at-line near infrared (NIR) or microwave resonance (MR) methods. Accurate NIR and MR prediction models were developed with a minimalistic approach to calibration. The NIR method can accurately predict water content in bulk powders in the range of 0.5-5% w/w. Results from this method were compared to a MR method. We demonstrated excellent agreement of both NIR and MR methods for powders vs. the reference KF method. These methods are applicable to in-process control or quality control environments. One of the aims of this study was to determine if a calibration developed for a particular product could be used to predict the water content of another product (with related composition) but containing a different active pharmaceutical ingredient (API). We demonstrated that, contrary to the NIR method, a general MR method can be used to predict water content in two different types of blends. Finally, we demonstrated that a MR method can be developed for at-line moisture determination in tablets.     

Determination of biodiesel content when blended with mineral diesel fuel using infrared spectroscopy and multivariate calibration

In this work, multivariable calibration models based on middle- and near-infrared spectroscopy were developed in order to determine the content of biodiesel in diesel fuel blends, considering the presence of raw vegetable oil. Soybean, castor and used frying oils and their corresponding esters were used to prepare the blends with conventional diesel. Results indicated that partial least squares (PLS) models based on MID or NIR infrared spectra were proven suitable as practical analytical methods for predicting biodiesel content in conventional diesel blends in the volume fraction range from 0% to 5%. PLS models were validated by independent prediction set and the RMSEPs were estimated as 0.25 and 0.18 (%, v/v). Linear correlations were observed for predicted vs. observed values plots with correlation coefficient (R) of 0.986 and 0.994 for the MID and NIR models, respectively. Additionally, principal component analysis (PCA) in the MID region 1700 to 1800 cm^− 1 was suitable for identifying raw vegetable oil contaminations and illegal blends of petrodiesel containing the raw vegetable oil instead of ester.     

Study of the temperature-dependent near-infrared spectra of water by two-dimensional correlation spectroscopy and principal components analysis

Near-infrared spectroscopy (NIR) is an important analytical tool in monitoring properties of systems for that water is a major constituent. For such objects of analysis a quality of information extracted from the NIR spectra depends essentially on used methods of analysis of a massive absorbance of water. Correctly chosen method should be able to identified rational number of overlapped components hidden under the broad band of water. The resolved components have to be justified on grounds of the structure of water and by relation to the properties a hydrogen-bonded network of water molecules. The interest in the correlation is imposed by a fact that hydrogen bonds of water around nonpolar group are significantly strengthened and weakened around polar groups. Intensity variations classified in this context could be valuable source of information on varying properties of the solute molecules embedded in water environment. Therefore, there is a big interest in methods that have a power for detailed analysis of the intensity changes in the broad NIR spectra. Two-dimensional correlation spectroscopy (2DCOS) and principal component analysis (PCA) are our proposition. In the analysis of the temperature-dependent NIR spectra of water by means of the two methods we have focused on the interpretation of the 2DCOS results through the concept of linear and nonlinear relationships. Moreover, a cascaded curve fitting procedure has been employed. Presented approach should be very instructive of how to interpret the features of the 2D results that frequently is not a straightforward task.     

Quality assessment of cortex cinnamomi by HPLC chemical fingerprint, principle component analysis and cluster analysis

HPLC fingerprint analysis, principle component analysis (PCA), and cluster analysis were introduced for quality assessment of Cortex cinnamomi (CC). The fingerprint of CC was developed and validated by analyzing 30 samples of CC from different species and geographic locations. Seventeen chromatographic peaks were selected as characteristic peaks and their relative peak areas (RPA) were calculated for quantitative expression of the HPLC fingerprints. The correlation coefficients of similarity in chromatograms were higher than 0.95 for the same species while much lower than 0.6 for different species. Besides, two principal components (PCs) have been extracted by PCA. PC1 separated Cinnamomum cassia from other species, capturing 56.75% of variance while PC2 contributed for their further separation, capturing 19.08% variance. The scores of the samples showed that the samples could be clustered reasonably into different groups corresponding to different species and different regions. The scores and loading plots together revealed different chemical properties of each group clearly. The cluster analysis confirmed the results of PCA analysis. Therefore, HPLC fingerprint in combination with chemometric techniques provide a very flexible and reliable method for quality assessment of traditional Chinese medicines.     

废旧涤/棉混纺织物近红外定量分析模型的建立及预测

将中红外光谱筛选出的598个纯涤、纯棉及涤/棉混纺样本采用GB/T 2910.11-2009法测定其涤、棉准确含量,其中校正集样本252个,验证集样本346个。使用便携式近红外光谱仪获取样本的原始近红外光谱(NIRS)。校正集样本依据回归系数的分布趋势和范围选取最佳建模谱区,并采用差分一阶导、S-G平滑和均值中心化相结合的方法对原始光谱进行预处理,利用偏最小二乘法(PLS)建立涤/棉混纺织物中涤含量的近红外(NIR)定量分析模型。同时分析了样本颜色对NIRS的影响,探讨了斜线光谱样本、奇异样本和不同组织结构织物对模型预测效果的影响。结果表明:利用PLS法建立的涤/棉混纺织物定量分析模型最优组合包含1个光谱区间和9个主成分因子,校正集相关系数(RC)为0.998,标准偏差(SEC)为0.908。为验证所建模型的有效性和实用性,对346个未参与建模的涤棉样本进行了预测,并将预测结果与国标法测定值进行方差分析,两种方法结果无显著差异,预测正确率达97%以上。模型的建立为废旧涤/棉混纺织物快速、无损分拣提供了基础数据库。     

Loading and release of ibuprofen in multi- and monofilament surgical sutures

The preparation of mono- and multifilament sutures incorporating ibuprofen as an anti-inflammatory agent is considered. Poly(p-dioxanone) monofilament samples can be loaded by a molecular diffusion process using a swelling agent such as dichloromethane. The mechanical properties have been measured and have not shown a significant change for the ibuprofen loaded samples in knot tensile assays. The kinetics of both the loading process and the release in a S?rensen's medium at 37 degrees C have been investigated. Diffusion coefficients have also been estimated from film and slab poly(p-dioxanone) samples containing ibuprofen and their release behavior compared to that shown by monofilaments. Release from a coating copolymer based on lactide, epsilon-caprolactone and trimethylene carbonate (PLA/PCA/PTMC 10/60/30) has also been studied. This coating solubilizes ibuprofen molecules well and can be used for braided sutures or when a rapid dose of ibuprofen is preferred.     

Near-IR Fourier transform Raman spectroscopy and PLS-2 modeling for improved understanding of near-IR spectra

The recent development of non-destructive near-IR (NIR) Raman techniques, which have the capability of providing fundamental vibrational information for bulk materials, has opened up a great possibility of understanding the non-destructive NIR spectra of such materials better, through statistical correlation of the two spectral methods. In this work, the use of NIR-FT-Raman spectroscopy and PLS-2 modeling to improve the understanding of the NIR spectroscopy of polyurethane elastomers is demonstrated. The use of this procedure resulted in improved assignments of the NIR bands corresponding to aromatic, urethane and urea groups in the elastomers, and an improved understanding of the NIR spectral effect that corresponds to the nitrogen void content in the elastomers.     

Maize kernel hardness classification by near infrared (NIR) hyperspectral imaging and multivariate data analysis

The use of near infrared (NIR) hyperspectral imaging and hyperspectral image analysis for distinguishing between hard, intermediate and soft maize kernels from inbred lines was evaluated. NIR hyperspectral images of two sets (12 and 24 kernels) of whole maize kernels were acquired using a Spectral Dimensions MatrixNIR camera with a spectral range of 960-1662 nm and a sisuChema SWIR (short wave infrared) hyperspectral pushbroom imaging system with a spectral range of 1000-2498 nm. Exploratory principal component analysis (PCA) was used on absorbance images to remove background, bad pixels and shading. On the cleaned images, PCA could be used effectively to find histological classes including glassy (hard) and floury (soft) endosperm. PCA illustrated a distinct difference between glassy and floury endosperm along principal component (PC) three on the MatrixNIR and PC two on the sisuChema with two distinguishable clusters. Subsequently partial least squares discriminant analysis (PLS-DA) was applied to build a classification model. The PLS-DA model from the MatrixNIR image (12 kernels) resulted in root mean square error of prediction (RMSEP) value of 0.18. This was repeated on the MatrixNIR image of the 24 kernels which resulted in RMSEP of 0.18. The sisuChema image yielded RMSEP value of 0.29. The reproducible results obtained with the different data sets indicate that the method proposed in this paper has a real potential for future classification uses.     

Nondestructive discrimination of ivories and prediction of their specific gravity by Fourier-transform Raman spectroscopy and chemometrics

Fourier-transform (FF) Raman spectroscopy and chemometrics were used for nondestructive analysis of ivories. The discrimination of five kinds of ivories, two subspecies of African elephant, mammoth, hippopotamus, and sperm whale, was investigated, and a calibration model for predicting their specific gravity was developed. FT-Raman spectra were measured in situ for them and chemometrics analyses were carried out for the 3050-350 cm(-1) region. The five kinds of ivories were clearly discriminated from each other on the scores plots of two or three principal components (PCs) obtained by principal component analysis (PCA). The loadings plot for PC 1 shows that the discrimination relies on the content ratio of organic collagenous protein and inorganic hydroxyapatite of ivories. The loadings plot for PC 2 shows that bands due to the CH3 and CH2 stretching modes of the protein also play a role in the discrimination. Using partial least squares regression (PLSR), we developed a calibration model that predicts the specific gravity of the ivories from the FT-Raman spectra. The correlation coefficient and root mean square error of cross validation (RMSECV) of this model were 0.980 and 0.024, respectively.