Monitoring polymorphic transformations by using in situ Raman hyperspectral imaging and image multiset analysis

Polymorphism is often encountered in many crystalline compounds. To control the quality of the products, it is relevant knowing the potential presence of polymorph transformations induced by different agents, such as light exposure or temperature changes. Raman images offer a great potential to identify polymorphs involved in a process and to accurately describe this kind of solid-state transformation in the surface scanned.     

A new regression method based on independent component analysis

Based on independent component analysis (ICA), a new regression method, independent component regression (ICR), was developed to build the model of NIR spectra and the routine components of plant samples. It is found that ICR and principal component regression (PCR) are completely equivalent when they are applied in quantitative prediction. However, independent components (ICs) can give more chemical explanation than principal components (PCs) because independence is a high-order statistic that is a much stronger condition than orthogonality. Three ICs are obtained by ICA from the NIR spectra of plant samples; it is found that they are strongly correlated to the NIR spectra of water, hydrocarbons and organonitrogen compounds, respectively. Therefore, ICA may be a promising tool to retrieve both quantitative and qualitative information from complex chemical data sets.     

Fiber array based hyperspectral Raman imaging for chemical selective analysis of malaria-infected red blood cells

A new setup for Raman spectroscopic wide-field imaging is presented. It combines the advantages of a fiber array based spectral translator with a tailor-made laser illumination system for high-quality Raman chemical imaging of sensitive biological samples.     

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.     

Identifying constituents in commercial gasoline using Fourier transform-infrared spectroscopy and independent component analysis

A new method is proposed that enables the identification of five refinery fractions present in commercial gasoline mixtures using infrared spectroscopic analysis. The data analysis and interpretation was carried out based on independent component analysis (ICA) and spectral similarity techniques. The FT-IR spectra of the gasoline constituents were determined using the ICA method, exclusively based on the spectra of their mixtures as a blind separation procedure, i.e. assuming unknown the spectra of the constituents. The identity of the constituents was subsequently determined using similarity measures commonly employed in spectra library searches against the spectra of the constituent components. The high correlation scores that were obtained in the identification of the constituents indicates that the developed method can be employed as a rapid and effective tool in quality control, fingerprinting or forensic applications, where gasoline constituents are suspected.     

Cross validation and uncertainty estimates in independent component analysis

A data analysis tool, known as independent component analysis (ICA), is the main focus of this paper. The theory of ICA is briefly reviewed, and the underlying statistical assumptions and a practical algorithm are described. This paper introduces cross validation/jack-knifing and significance tests to ICA. Jack-knifing is applied to estimate uncertainties for the ICA loadings, which also serve as a basis for significance tests. These tests are shown to improve ICA performance, indicating how many components are mixed in the observed data, and also which parts of the extracted sources that contain significant information. We address the issue of stability for the ICA model through uncertainty plots. The ICA performance is compared to principal component analysis (PCA) for two selected applications, a simulated experiment and a real world application.     

基于高光谱和化学计量学方法的除草剂残留分类识别研究

本研究基于高光谱技术和化学计量学方法,对薄荷叶上的异丙甲草胺、烟嘧(莠去津、敌草胺和砜嘧(精喹4类除草剂残留进行种类判别.高光谱成像光谱范围为450nm～950nm的可见-近红外区域.为降低噪音对光谱数据的干扰、提升判别精度,采用SG平滑和多元散射校正对高光谱曲线进行处理.利用主成分分析算法(PCA)对原始数据进行降维...     

Independent component analysis as a pretreatment method for parallel factor analysis to eliminate artefacts from multiway data

Parallel factor analysis (PARAFAC) has successfully been used in many applications for the analysis of excitation-emission fluorescence data. However, some measurement “artefacts”, such as Rayleigh or Raman scattering, can pose a problem for the extraction of the PARAFAC components and their interpretation. Replacing the spectral zones corresponding to these signals by missing values in the data is not necessarily a method of choice in the cases where informative signals lie in the same wavelength regions. In this article, independent component analysis (ICA) is used on the unfolded cubic array, and the independent components related to the Rayleigh and Raman scattering are identified and removed prior to the reconstruction of the excitation-emission fluorescence data cube. PARAFAC is then applied on these data reconstructed after selective artefact removal, and satisfactory models can be obtained. This procedure, although particularly useful for 3D fluorescence data, may be applied to other types of data as well.     

Photostable lysosomal imaging of living cell with hyperspectral stimulated Raman scattering microscopy using a probe based on bisarylbutadiyne

Lysosome-selective Raman probe, Lyso-BADY, was created with improved Raman signal intensity, stability as well as specificity under the visualization of hyperspectral stimulated Raman scattering (SRS) microscopy. Prolonged exposure to irradiation further proved its photostability and potentiality for continuous imaging.     

Investigation of sport supplements quality by Raman spectroscopy and principal component analysis

In this work, sport supplements were investigated by Raman spectroscopy. Samples were obtained from health foods shops, gyms and sports centers covering a wide range of available supplement powders. A systematic comparison of Raman spectra of the analyzed supplements allowed identifying the supplement type through the characteristic vibrational modes of carbohydrates and proteins. The protein supplements were identified by Raman bands at 1650, 1250 and 1004 cm⁻¹, while the spectral range between 1200 and 800 cm⁻¹ was useful to identify the carbohydrate supplements. Due to the diversity in composition of sport supplements, a chemometric tool such as principal component analysis (PCA) was employed to assist in the interpretation of Raman spectra, allowing also the identification of compounds present in sport supplements. Especially, the Raman scattering of aromatic and aliphatic amino acids residues contributes to the existence of bands characteristic for the different types of proteins. This kind of information is very important for the quality control of these products, for detecting the presence of fraud or a sample composition in disagreement with the label, thus ensuring the provenance of the supplements.     

Determination of cocaine contamination on banknotes using tandem mass spectrometry and pattern recognition

Tandem mass spectrometry is used to monitor the contamination of banknotes by cocaine. By introducing a series of banknotes into an instrument a distribution of contamination can be obtained. The distribution of samples arising from defendants where the banknotes have been in close proximity to cocaine should differ from the distribution from the general background population. Peak picking and integration is used to produce a series of intensity readings for a batch of banknotes. By visually inspecting these distribution, and applying a variety of chemometric methods (principal components analysis, cluster analysis and class modelling via Mahalanobis distance) it is possible to discriminate effectively between the two classes of distribution (7157 background notes and 4826 case notes alleged to be from drug dealers). By calculating the Mahalonobis distance over 100 bootstrap iterations, background samples were correctly classified 96.48% of the time, while case samples were correctly classified 89.37% of the time.     

Raman imaging analysis of pharmaceutical tablets by two-dimensional (2D) correlation spectroscopy

Chemical properties of active substances and insoluble excipient within tablets such as crystalline structures can be seen as an important index for solubility of ingredients. Spectroscopic imaging can potentially be a solid solution to understanding mechanisms at the molecular level and it may bring useful insight in terms of process analytical technique. In the present study, generalized two-dimensional (2D) correlation spectroscopy is utilized for the Raman image analysis of pharmaceutical tablets to reveal molecular interactions between chemical components. By using a spatial distance as a perturbation variable in 2D correlation scheme, synchronous and asynchronous correlation analysis becomes possible. Two kinds of pharmaceutical tablets, pentoxifylline (PTX) as an active substance and palmitic acid (PA) as an insoluble excipient, are prepared with different grinding times, 0.5 and 45 min. The 2D correlation analysis of Raman images of the tablets clearly reveals both physical and chemical effects of grinding process on the properties of the tablets. Asynchronous correlations indicate that a specific molecular structural change of PTX related to the crystallinity is induced by the grinding process. Namely, the crystallinity of PTX based on CH₂ structure is a key factor to control the solubility of the tablets. Some properties of pharmaceutical tablets, i.e. solubility or distribution of components in turn may become possible by the simple grinding process. Detailed analysis of Raman images becomes possible by the 2D correlation spectroscopy.     

Multispectral optical imaging combined in situ with XPS or ToFSIMS and principal component analysis

All X‐ray photoelectron spectroscopy (XPS) and time‐of‐flight secondary ion mass spectrometry (ToF‐SIMS) instruments have optical cameras to image the specimen under analysis, and often to image the sample holder as it enters the system too. These cameras help the user find the appropriate points for analysis of specimens. However they seldom give as good images as stand‐alone bench optical microscopes, because of the limited geometry, source/analyser solid angle and ultra‐high‐vacuum (UHV) design compromises. This often means that the images displayed to the user necessarily have low contrast, low resolution and poor depth‐of‐field. To help identify the different regions of the samples present we have found it useful to perform multispectral imaging by illuminating the sample with narrow‐wavelength‐range light emitting diodes (LEDs). By taking an image under the illumination of these LEDs in turn, each at a successively longer wavelength, one can build up a set of registered images that contain more information than a simple Red–Green–Blue image under white‐light illumination. We show that this type of multispectral imaging is easy and inexpensive to fit to common XPS and ToF‐SIMS instruments, using LEDs that are widely available. In our system we typically use 14 LEDs including one emitting in the ultraviolet (so as to allow fluorescent imaging) and three in the near infra‐red. The design considerations of this system are discussed in detail, including the design of the drive and control electronics, and three practical examples are presented where this multispectral imaging was extremely useful. Copyright © 2016 The Authors Surface and Interface Analysis Published by John Wiley & Sons Ltd.     

Multivariate data analysis for Raman imaging of a model pharmaceutical tablet

Spectroscopic imaging techniques provide spatial and spectral information about a sample simultaneously and are finding ever-increasing application in the pharmaceutical industry. Effective extraction of chemical information from imaging data sets is a crucial step during the application of imaging techniques. Multivariate imaging data analysis methods have been reported but few applications of these methods for pharmaceutical samples have been demonstrated. In this study, a bilayer model tablet consisting of avicel, lactose, sodium benzoate, magnesium stearate and red dye was prepared using custom press tooling, and Raman mapping data were collected from a 400 μm × 400 μm area of the tablet surface. Several representative multivariate methods were selected and used in the analysis of the data. Multivariate data analysis methods investigated include principal component analysis (PCA), cluster analysis, direct classical least squares (DCLS) and multivariate curve resolution (MCR). The relative merits and drawbacks of each technique for this application were evaluated. In addition, some practical issues associated with the use of these methods were addressed including data preprocessing, determination of the optimal number of clusters in cluster analysis and the optimization of window size in second derivative calculation.     

Detection of illicit substances in fingerprints by infrared spectral imaging

FTIR and Raman spectral imaging can be used to simultaneously image a latent fingerprint and detect exogenous substances deposited within it. These substances might include drugs of abuse or traces of explosives or gunshot residue. In this work, spectral searching algorithms were tested for their efficacy in finding targeted substances deposited within fingerprints. “Reverse” library searching, where a large number of possibly poor-quality spectra from a spectral image are searched against a small number of high-quality reference spectra, poses problems for common search algorithms as they are usually implemented. Out of a range of algorithms which included conventional Euclidean distance searching, the spectral angle mapper (SAM) and correlation algorithms gave the best results when used with second-derivative image and reference spectra. All methods tested gave poorer performances with first derivative and undifferentiated spectra. In a search against a caffeine reference, the SAM and correlation methods were able to correctly rank a set of 40 confirmed but poor-quality caffeine spectra at the top of a dataset which also contained 4,096 spectra from an image of an uncontaminated latent fingerprint. These methods also successfully and individually detected aspirin, diazepam and caffeine that had been deposited together in another fingerprint, and they did not indicate any of these substances as a match in a search for another substance which was known not to be present. The SAM was used to successfully locate explosive components in fingerprints deposited on silicon windows. The potential of other spectral searching algorithms used in the field of remote sensing is considered, and the applicability of the methods tested in this work to other modes of spectral imaging is discussed.     

Independent component analysis of nanomechanical responses of cantilever arrays

The ability to detect and identify chemical and biological elements in air or liquid environments is of far reaching importance. Performing this task using technology that minimally impacts the perceived environment is the ultimate goal. The development of functionalized cantilever arrays with nanomechanical sensing is an important step towards this goal. This report couples the feature extraction abilities of independent component analysis (ICA) and the classification techniques of neural networks to analyze the signals produced by microcantilever-array-based nanomechanical sensors. The unique capabilities of this analysis unleash the potential of this sensing technology to accurately identify chemical mixtures and concentrations. Furthermore, it is demonstrated that the knowledge of how the sensor array reacts to individual analytes in isolation is sufficient information to decode mixtures of analytes—a substantial benefit, significantly increasing the analytical utility of these sensing devices.     

Prediction of some quality attributes of lamb meat using near-infrared hyperspectral imaging and multivariate analysis

The goal of this study was to explore the potential of near-infrared (NIR) hyperspectral imaging in combination with multivariate analysis for the prediction of some quality attributes of lamb meat. In this study, samples from three different muscles (semitendinosus (ST), semimembranosus (SM), longissimus dorsi (LD)) originated from Texel, Suffolk, Scottish Blackface and Charollais breeds were collected and used for image acquisition and quality measurements. Hyperspectral images were acquired using a pushbroom NIR hyperspectral imaging system in the spectral range of 900–1700 nm. A partial least-squares (PLS) regression, as a multivariate calibration method, was used to correlate the NIR reflectance spectra with quality values of the tested muscles. The models performed well for predicting pH, colour and drip loss with the coefficient of determination (R²) of 0.65, 0.91 and 0.77, respectively. Image processing algorithm was also developed to transfer the predictive model in every pixel to generate prediction maps that visualize the spatial distribution of quality parameter in the imaged lamb samples. In addition, textural analysis based on gray level co-occurrence matrix (GLCM) was also conducted to determine the correlation between textural features and drip loss. The results clearly indicated that NIR hyperspectral imaging technique has the potential as a fast and non-invasive method for predicting quality attributes of lamb meat.     

Passive standoff detection of RDX residues on metal surfaces via infrared hyperspectral imaging

Hyperspectral images of galvanized steel plates, each containing a stain of cyclotrimethylenetrinitramine (RDX), were recorded using a commercial long-wave infrared imaging spectrometer. Demonstrations of passive RDX chemical detection at areal dosages between 16 and 90 μg/cm² were carried out over practical standoff ranges between 14 and 50 m. Anomaly and target detection algorithms were applied to the images to determine the effect of areal dosage and sensing distance on detection performance for target RDX. The anomaly detection algorithms included principal component analysis, maximum autocorrelation factors, and principal autocorrelation factors. Maximum difference factors and principal difference factors are novel multivariate edge detection techniques that were examined for their utility in detection of the RDX stains in the images. A target detection algorithm based on generalized least squares was applied to the images, as well, to see if the algorithm can identify the compound in the stains on the plates using laboratory reflection spectra of RDX, cyclotetramethylenetetranitramine (HMX), and 2,4,6-trinitrotoluene (TNT) as the target spectra. The algorithm could easily distinguish between the nitroaromatic (TNT) compound and the nitramine (RDX, HMX) compounds, and, though the distinction between RDX and HMX was less clear, the mean weighted residuals identified RDX as the stain on the plate. Improvements that can be made in this detection technique are discussed in detail. As expected, it was found that detection was best for short distances and higher areal dosages. However, the target was easily detected at all distances and areal dosages used in this study. Electronic supplementary material  The online version of this article (doi:) contains supplementary material, which is available to authorized users.     

Semi-quantitative trace analysis of nuclear fast red by surface enhanced resonance Raman scattering

The dye nuclear fast red has been detected and determined semi-quantitatively by means of surface enhanced resonance Raman scattering (SERRS) and surface enhanced Raman scattering (SERS), using laser exciting wavelengths of 514.5 and 632.8 nm, respectively, by employing a citrate-reduced silver colloid. A good linear correlation is observed for the dependence of the intensities of the SERRS bands at 989 cm⁻¹ (R=0.9897) and 1278 cm⁻¹ (R=0.9872) on dye concentration over the range 10⁻⁹ to 10⁻⁷ M, when using an exciting wavelength of 514.5 nm. At dye concentrations above 10⁻⁷ M, the concentration dependence of the SERRS signals is non-linear. This is almost certainly due to the coverage of the colloidal silver particles being in excess of a full monolayer of the dye. A linear correlation is also observed for the dependence of the intensities of the SERS bands at 989 cm⁻¹ (R=0.9739) and 1278 cm⁻¹ (R=0.9838) on the dye concentration over the range 10⁻⁸ to 10⁻⁶ M when using an exciting wavelength of 632.8 nm. Strong fluorescence prevented collection of resonance Raman scattering (RRS) spectra from powdered samples or aqueous solutions of the dye using an exciting wavelength of 514.5 nm, but weak bands were observed in the spectra obtained from both powdered and aqueous samples of the dye using an exciting wavelength of 632.8 nm. A study of the pH dependence of SERRS/SERS and UV–VIS absorption spectra revealed the presence of different ionisation states of the dye. The limits of detection for nuclear fast red by SERRS (514.5 nm), SERS (632.8 nm) and visible spectroscopy (535 nm) are 9, 89 and 1000 ng ml⁻¹, respectively.     

Innovative high-performance liquid chromatography method development for the screening of 19 antimalarial drugs based on a generic approach, using design of experiments, independent component analysis and design space

An innovative methodology based on design of experiments (DoE), independent component analysis (ICA) and design space (DS) was developed in previous works and was tested out with a mixture of 19 antimalarial drugs. This global LC method development methodology (i.e. DoE-ICA-DS) was used to optimize the separation of 19 antimalarial drugs to obtain a screening method. DoE-ICA-DS methodology is fully compliant with the current trend of quality by design. DoE was used to define the set of experiments to model the retention times at the beginning, the apex and the end of each peak. Furthermore, ICA was used to numerically separate coeluting peaks and estimate their unbiased retention times. Gradient time, temperature and pH were selected as the factors of a full factorial design. These retention times were modelled by stepwise multiple linear regressions. A recently introduced critical quality attribute, namely the separation criterion (S), was also used to assess the quality of separations rather than using the resolution. Furthermore, the resulting mathematical models were also studied from a chromatographic point of view to understand and investigate the chromatographic behaviour of each compound. Good adequacies were found between the mathematical models and the expected chromatographic behaviours predicted by chromatographic theory. Finally, focusing at quality risk management, the DS was computed as the multidimensional subspace where the probability for the separation criterion to lie in acceptance limits was higher than a defined quality level. The DS was computed propagating the prediction error from the modelled responses to the quality criterion using Monte Carlo simulations. DoE-ICA-DS allowed encountering optimal operating conditions to obtain a robust screening method for the 19 considered antimalarial drugs in the framework of the fight against counterfeit medicines. Moreover and only on the basis of the same data set, a dedicated method for the determination of three antimalarial compounds in a pharmaceutical formulation was optimized to demonstrate both the efficiency and flexibility of the methodology proposed in the present study.