This PDF file contains the front matter associated with SPIE Proceedings Volume 9488 including the Title Page, Copyright information, Table of Contents, Introduction, and Conference Committee listing.

In order to secure food safety, high throughput pathogen screening technique that can quickly identify and isolate unsafe contaminated foods has been long-desired. Recently, magnetoelastic (ME) free-standing biosensors have been investigated as a label-free wireless biosensor system for real-time pathogen detection. ME biosensor is composed of a ME resonator coated with a bio-molecular recognition element that binds specifically with a target pathogen. Once the biosensor comes into contact with the target pathogen, binding occurs, resulting in a decrease of the sensor's resonant frequency. Interrogated through magnetic signals, large amount of ME sensors can be deployed and monitored wirelessly. ME biosensors have been investigated to detect foodborne pathogens in cultures and liquid foods. Recently, it has been demonstrated that phage-based ME biosensors are able to directly detect Salmonella Typhimurium on food surfaces without the requirement of pre-analysis culture preparation. This paper will review the novel ME biosensor technique, including the detection principle, the characterization of the sensor performance, the deployment of multiple sensor detection and their applications, especially for food safety analysis. The ME biosensor technique has the potential to be a powerful pathogen screening tool for detecting contaminated food, identifying critical hazard points, and tracking contamination sources along the entire food supply chain.

Real-time in-situ detection of pathogenic bacteria on fresh food surfaces was accomplished with phage-based magnetoelastic (ME) biosensors. The ME biosensor is constructed of a small rectangular strip of ME material that is coated with a biomolecular recognition element (phage, antibodies or proteins, etc.) that is specific to the target pathogen. This mass-sensitive ME biosensor is wirelessly actuated into mechanical resonance by an externally applied time-varying magnetic field. When the biosensor binds with target bacteria, the mass of the sensor increases, resulting in a decrease in the sensor's resonant frequency. In order to compensate for nonspecific binding, control biosensors without phage were used in this experiment. In previous research, the biosensors were measured one by one. However, the simultaneous measurement of multiple sensors was accomplished in this research, and promises to greatly shorten the analysis time for bacterial detection. Additionally, the use of multiple biosensors enables the possibility of simultaneous detection of different pathogenic bacteria. This paper presents results of experiments in which multiple phage-based ME biosensors were simultaneously monitored. The E2 phage and JRB7 phage from a landscape phage library served as the bio-recognition element that have the capability of binding specifically with Salmonella typhimurium and B. anthracis spores, respectively. Real-time in-situ detection of Salmonella typhimurium and B. anthracis spores on food surfaces are presented.

Nondestructive in situ determination of the antioxidant lycopene of fresh tomato fruits is of large interest for the growers, willing to optimize the harvest time for high quality products. For this, we developed a portable LED-based colorimeter which was able to measure reflectance spectra of whole tomatoes in the 400-750 nm range. The tomato skins from the same samples were then frozen in liquid nitrogen, extracted with an acetone/ethanol/hexane mixture and analyzed by means of a spectrophotometer for their lycopene content. Concentration of lycopene was varying between 70 and 550 mg/Kg fresh weight skin. Partial Least Square regression was used to correlate spectral data to the tomato lycopene content. The multivariate processing of the reflectance data showed that lycopene content could be nicely predicted with a coefficient of determination R²=0.945 and a root mean square error of cross-validation RMSECV=57 mg/Kg skin fresh weight. These results suggest that portable, low-cost and compact LED-based sensors appear to be promising instruments for the nondestructive assessment of tomato lycopene even in the field.

This paper describes the results of a research project to investigate magnetoelastic (ME) biosensors actuated with a pulse excitation to measure the concentration of Salmonella Typhimurium of globe fruits. The ME biosensors are based on an acoustic wave resonator platform that is a freestanding (free-free) thin ribbon of magnetostrictive material with a lengthto- width ratio of 5:1. A biorecognition probe coated on the surface of the resonator platform binds with a targeted pathogen, i.e. E2 phage that binds with S. Typhimurium. The biosensor was actuated to vibrate longitudinally such that the resonant frequency depended primarily on the length of sensor and its overall mass. A pulsed excitation and measurement system was used to actuate micron scale ME biosensors to vibrate. The biosensor responds in a ring-down manner, a damped decay of the resonance amplitude, from which the resonant frequency was measured. An increase in mass due to the binding of the target pathogen resulted in a decrease in the resonant frequency. The pulsed excitation and measurement system that was developed under this effort and the characterization of its performance on the measurement of Salmonella concentrations on globe fruits is described.

We present low-cost bioenabled surface-enhanced Raman scattering (SERS) substrates that can be massively produced in sustainable and eco-friendly methods with significant commercial potentials for the detection of food contamination and drinking water pollution. The sensors are based on diatom frustules with integrated plasmonic nanoparticles. The ultra-high sensitivity of the SERS substrates comes from the coupling between the diatom frustules and Ag nanoparticles to achieve dramatically increased local optical field to enhance the light-matter interactions for SERS sensing. We successfully applied the bioenabled SERS substrates to detect melamine in milk and aromatic compounds in water with sensitivity down to 1μg/L.

This work describes a biosensor based on magnetic resonance relaxation switching. The method leverages a large body of work involving nanoscale contrast agents employed in nuclear magnetic resonance (NMR) imaging. The aim was to develop a detection approach that mimics the human immune response to an invading pathogen, the release of 10⁹ to 10¹² specific antigens to guarantee quick contact with the pathogen. The technique employs magnetic nanoparticle contrast agents conjugated with specific capture agents to achieve a similar contact goal. Detection of the species involves monitoring the average relaxation time (T2) of water protons in the solution, which is highly sensitive to the concentration and distribution of the magnetic nanoparticles present. With multiple nanoparticles attaching to each individual target species their distribution will be altered, and correspondingly, the average proton relaxation time will change

This paper presents an investigation into magnetoelastic (ME) biosentinels that capture and detect low-concentration pathogenic bacteria in stagnant liquids. The ME biosentinels are designed to mimic a variety of white blood cell types, known as the main defensive mechanism in the human body against different pathogenic invaders. The ME biosentinels are composed of a freestanding ME resonator coated with an engineered phage that specifically binds with the pathogens of interest. These biosentinels are ferromagnetic and thus can be moved through a liquid by externally applied magnetic fields. In addition, when a time-varying magnetic field is applied, the ME biosentinels can be placed into mechanical resonance by magnetostriction. As soon as the biosentinels bind with the target pathogen through the phage-based biomolecular recognition, a change in the biosentinel’s resonant frequency occurs, and thereby the presence of the target pathogen can be detected. Detection of Bacillus anthracis spores under stagnant flow conditions was demonstrated.

Addition of edible and inedible chemical contaminants in food powders for purposes of economic benefit has become a recurring trend. In recent years, severe health issues have been reported due to consumption of food powders contaminated with chemical substances. This study examines the effect of spatial resolution used during spectral collection to select the optimal spatial resolution for detecting melamine in milk powder. Sample depth of 2mm, laser intensity of 200mw, and exposure time of 0.1s were previously determined as optimal experimental parameters for Raman imaging. Spatial resolution of 0.25mm was determined as the optimal resolution for acquiring spectral signal of melamine particles from a milk-melamine mixture sample. Using the optimal resolution of 0.25mm, sample depth of 2mm and laser intensity of 200mw obtained from previous study, spectral signal from 5 different concentration of milk-melamine mixture (1%, 0.5%, 0.1%, 0.05%, and 0.025%) were acquired to study the relationship between number of detected melamine pixels and corresponding sample concentration. The result shows that melamine concentration has a linear relation with detected number of melamine pixels with correlation coefficient of 0.99. It can be concluded that the quantitative analysis of powder mixture is dependent on many factors including physical characteristics of mixture, experimental parameters, and sample depth. The results obtained in this study are promising. We plan to apply the result obtained from this study to develop quantitative detection model for rapid screening of melamine in milk powder. This methodology can also be used for detection of other chemical contaminants in milk powders.

A newly developed line-scan Raman imaging system using a 785 nm line laser was used to authenticate powdered foods and ingredients. The system was used to collect hyperspectral Raman images in a wavenumber range of 102–2865 cm^–1 from three representative food powders mixed with selected adulterants with a concentration of 0.5%, including milk and melamine, flour and benzoyl peroxide, and starch and maleic anhydride. An acoustic mixer was used to create food adulterant mixtures. All the mixed samples were placed in sample holders with a surface area of 50 mm×50 mm. Spectral and image processing algorithms were developed based on single-band images at unique Raman peaks of the individual adulterants. Chemical images were created to show identification, spatial distribution, and morphological features of the adulterant particles mixed in the food powders. The potential of estimating mass concentrations of the adulterants using the percentages of the adulterant pixels in the chemical images was also demonstrated.

In this work we present Shifted Excitation Raman Difference Spectroscopy (SERDS) as a potential spectroscopic tool for outdoor measurements in precision agriculture. A dual-wavelength diode laser at 785 nm is used as an excitation light source which provides an optical power up to 100 mW in cw-operation. Both emission lines for SERDS show single mode operation with a spectral width of ≤ 11 pm and a spectral distance of about 10 cm^-1 over the whole power range. Raman experiments on apples are carried out and show Raman signals from wax layer and β-carotene. Raman investigations under daylight conditions are performed to simulate outdoor measurements. Here, polystyrene (PS) is used as test sample. A broadband signal together with narrow absorption lines of water vapor and Fraunhofer lines of singly ionized calcium (Ca II) mask the Raman lines of PS. Only the strong Raman signal at 999 cm^-1 is visible. SERDS efficiently separates the Raman signals of PS from the background signals and a 14-fold improvement of the signal-tobackground noise ratio is achieved.

This paper reports the chemometric analysis of near-infrared spectra drawn from hyperspectral images to develop, evaluate, and compare statistical models for the detection of beef in fish meal. There were 40 pure-fish meal samples, 15 pure-beef meal samples, and 127 fish/beef mixture meal samples prepared for hyperspectral line-scan imaging by a machine vision system. Spectral data for 3600 pixels per sample, in which individual spectra was obtain, were retrieved from the region of interest (ROI) in every sample image. The spectral data spanning 969 nm to 1551 nm (across 176 spectral bands) were analyzed. Statistical models were built using the principal component analysis (PCA) and the partial least squares regression (PLSR) methods. The models were created and developed using the spectral data from the purefish meal and pure-beef meal samples, and were tested and evaluated using the data from the ROI in the mixture meal samples. The results showed that, with a ROI as large as 3600 pixels to cover sufficient area of a mixture meal sample, the success detection rate of beef in fish meal could be satisfactory 99.2% by PCA and 98.4% by PLSR.

Aflatoxins are secondary metabolites produced by certain fungal species of the Aspergillus genus. Aflatoxin contamination remains a problem in agricultural products due to its toxic and carcinogenic properties. Conventional chemical methods for aflatoxin detection are time-consuming and destructive. This study employed fluorescence and reflectance visible near-infrared (VNIR) hyperspectral images to classify aflatoxin contaminated corn kernels rapidly and non-destructively. Corn ears were artificially inoculated in the field with toxigenic A. flavus spores at the early dough stage of kernel development. After harvest, a total of 300 kernels were collected from the inoculated ears. Fluorescence hyperspectral imagery with UV excitation and reflectance hyperspectral imagery with halogen illumination were acquired on both endosperm and germ sides of kernels. All kernels were then subjected to chemical analysis individually to determine aflatoxin concentrations. A region of interest (ROI) was created for each kernel to extract averaged spectra. Compared with healthy kernels, fluorescence spectral peaks for contaminated kernels shifted to longer wavelengths with lower intensity, and reflectance values for contaminated kernels were lower with a different spectral shape in 700-800 nm region. Principal component analysis was applied for data compression before classifying kernels into contaminated and healthy based on a 20 ppb threshold utilizing the K-nearest neighbors algorithm. The best overall accuracy achieved was 92.67% for germ side in the fluorescence data analysis. The germ side generally performed better than endosperm side. Fluorescence and reflectance image data achieved similar accuracy.

Rice is a dominant food crop of Bangladesh accounting about 75 percent of agricultural land use for rice cultivation and currently Bangladesh is the world’s fourth largest rice producing country. Rice provides about two-third of total calorie supply and about one-half of the agricultural GDP and one-sixth of the national income in Bangladesh. Aus is one of the main rice varieties in Bangladesh. Crop production, especially rice, the main food staple, is the most susceptible to climate change and variability. Any change in climate will, thus, increase uncertainty regarding rice production as climate is major cause year-to-year variability in rice productivity. This paper shows the application of remote sensing data for estimating Aus rice yield in Bangladesh using official statistics of rice yield with real time acquired satellite data from Advanced Very High Resolution Radiometer (AVHRR) sensor and Principal Component Regression (PCR) method was used to construct a model. The simulated result was compared with official agricultural statistics showing that the error of estimation of Aus rice yield was less than 10%. Remote sensing, therefore, is a valuable tool for estimating crop yields well in advance of harvest, and at a low cost.

In some fruits and vegetables, it is difficult to visually identify the ones which are pest infested. This particular aspect is important for quarantine and commercial operations. In this article, we propose to present the results of a novel technique using thermal imaging camera to detect the nature and extent of pest damage in fruits and vegetables, besides indicating the level of maturity and often the presence of the pest. Our key idea relies on the fact that there is a difference in the heat capacity of normal and damaged ones and also observed the change in surface temperature over time that is slower in damaged ones. This paper presents the concept of non-destructive evaluation using thermal imaging technique for identifying pest damage levels of fruits and vegetables based on investigations carried out on random samples collected from a local market.

Conventional microbiological detection and enumeration methods are time-consuming, labor-intensive, and giving retrospective information. The objectives of the present work are to study the capability of surface enhanced Raman scattering (SERS) to detect Escherichia coli (E. coli) using the presented silver colloidal substrate. The obtained results showed that the adaptive iteratively reweighed Penalized Least Squares (airPLS) algorithm could effectively remove the fluorescent background from original Raman spectra, and Raman characteristic peaks of 558, 682, 726, 1128, 1210 and 1328 cm^-1 could be observed stably in the baseline corrected SERS spectra of all studied bacterial concentrations. The detection limit of SERS could be determined to be as low as 0.73 log CFU/ml for E. coli with the prepared silver colloidal substrate. The quantitative prediction results using the intensity values of characteristic peaks were not good, with the correlation coefficients of calibration set and cross validation set of 0.99 and 0.64, respectively.

Residual pesticides in fruits and vegetables have become one of the major food safety concerns around the world. At present, routine analytical methods used for the determination of pesticide residue on the surface of fruits and vegetables are destructive, complex, time-consuming, high cost and not environmentally friendly. In this study, a novel Surface Enhanced Raman Spectroscopy (SERS) method with silver colloid was developed for fast and sensitive nondestructive detection of residual pesticides in fruits and vegetables by using a self-developed Raman system. SERS technology is a combination of Raman spectroscopy and nanotechnology. SERS can greatly enhance the Raman signal intensity, achieve single-molecule detection, and has a simple sample pre-treatment characteristic of high sensitivity and no damage; in recent years it has begun to be used in food safety testing research. In this study a rapid and sensitive method was developed to identify and analyze mixed pesticides of chlorpyrifos, deltamethrin and acetamiprid in apple samples by SERS. Silver colloid was used for SERS measurement by hydroxylamine hydrochloride reduced. The advantages of this method are seen in its fast preparation at room temperature, good reproducibility and immediate applicability. Raman spectrum is highly interfered by noise signals and fluorescence background, which make it too complex to get good result. In this study the noise signals and fluorescence background were removed by Savitzky-Golay filter and min-max signal adaptive zooming method. Under optimal conditions, pesticide residues in apple samples can be detected by SERS at 0.005 μg/cm² and 0.002 μg/cm² for individual acetamiprid and thiram, respectively. When mixing the two pesticides at low concentrations, their characteristic peaks can still be identified from the SERS spectrum of the mixture. Based on the synthesized material and its application in SERS operation, the method represents an ultrasensitive SERS performance in apple samples detection without sample pre-treatment, which indicates that it could be served as a useful means in monitoring pesticide residues.

Benzoyl peroxide is a common flour additive that improves the whiteness of flour and the storage properties of flour products. However, benzoyl peroxide adversely affects the nutritional content of flour, and excess consumption causes nausea, dizziness, other poisoning, and serious liver damage. This study was focus on detection of the benzoyl peroxide added in wheat flour. A Raman scattering spectroscopy system was used to acquire spectral signal from sample data and identify benzoyl peroxide based on Raman spectral peak position. The optical devices consisted of Raman spectrometer and CCD camera, 785 nm laser module, optical fiber, prober, and a translation stage to develop a real-time, nondestructive detection system. Pure flour, pure benzoyl peroxide and different concentrations of benzoyl peroxide mixed with flour were prepared as three sets samples to measure the Raman spectrum. These samples were placed in the same type of petri dish to maintain a fixed distance between the Raman CCD and petri dish during spectral collection. The mixed samples were worked by pretreatment of homogenization and collected multiple sets of data of each mixture. The exposure time of this experiment was set at 0.5s. The Savitzky Golay (S-G) algorithm and polynomial curve-fitting method was applied to remove the fluorescence background from the Raman spectrum. The Raman spectral peaks at 619 cm^-1, 848 cm^-1, 890 cm^-1, 1001 cm^-1, 1234 cm^-1, 1603cm^-1, 1777cm^-1 were identified as the Raman fingerprint of benzoyl peroxide. Based on the relationship between the Raman intensity of the most prominent peak at around 1001 cm^-1 and log values of benzoyl peroxide concentrations, the chemical concentration prediction model was developed. This research demonstrated that Raman detection system could effectively and rapidly identify benzoyl peroxide adulteration in wheat flour. The experimental result is promising and the system with further modification can be applicable for more products in near future.

With the continuous development of living standards and the relative change of dietary structure, consumers’ rising and persistent demand for better quality of meat is emphasized. Colour, pH value, and cooking loss are important quality attributes when evaluating meat. To realize nondestructive detection of multi-parameter of meat quality simultaneously is popular in production and processing of meat and meat products. The objectives of this research were to compare the effectiveness of two bands for rapid nondestructive and simultaneous detection of pork quality attributes. Reflectance spectra of 60 chilled pork samples were collected from a dual-band visible/near-infrared spectroscopy system which covered 350-1100 nm and 1000-2600 nm. Then colour, pH value and cooking loss were determined by standard methods as reference values. Standard normal variables transform (SNVT) was employed to eliminate the spectral noise. A spectrum connection method was put forward for effective integration of the dual-band spectrum to make full use of the whole efficient information. Partial least squares regression (PLSR) and Principal component analysis (PCA) were applied to establish prediction models using based on single-band spectrum and dual-band spectrum, respectively. The experimental results showed that the PLSR model based on dual-band spectral information was superior to the models based on single band spectral information with lower root means quare error (RMSE) and higher accuracy. The PLSR model based on dual-band (use the overlapping part of first band) yielded the best prediction result with correlation coefficient of validation (Rv) of 0.9469, 0.9495, 0.9180, 0.9054 and 0.8789 for L*, a*, b*, pH value and cooking loss, respectively. This mainly because dual-band spectrum can provide sufficient and comprehensive information which reflected the quality attributes. Data fusion from dual-band spectrum could significantly improve pork quality parameters prediction performance. The research also indicated that multi-band spectral information fusion has potential to comprehensively evaluate other quality and safety attributes of pork.

Raman spectroscopy is a useful, rapid, and non-destructive method for both qualitative and quantitative evaluation of chemical composition. However it is important to measure the depth of penetration of the laser light to ensure that chemical particles at the very bottom of a sample volume is detected by Raman system. The aim of this study was to investigate the penetration depth of a 785nm laser (maximum power output 400mw) into three different food powders, namely dry milk powder, corn starch, and wheat flour. The food powders were layered in 5 depths between 1 and 5 mm overtop a Petri dish packed with melamine. Melamine was used as the subsurface reference material for measurement because melamine exhibits known and identifiable Raman spectral peaks. Analysis of the sample spectra for characteristics of melamine and characteristics of milk, starch and flour allowed determination of the effective penetration depth of the laser light in the samples. Three laser intensities (100, 200 and 300mw) were used to study the effect of laser intensity to depth of penetration. It was observed that 785nm laser source was able to easily penetrate through every point in all three food samples types at 1mm depth. However, the number of points that the laser could penetrate decreased with increasing depth of the food powder. ANOVA test was carried out to study the significant effect of laser intensity to depth of penetration. It was observed that laser intensity significantly influences the depth of penetration. The outcome of this study will be used in our next phase of study to detect different chemical contaminants in food powders and develop quantitative analysis models for detection of chemical contaminants.

The wireless phage-based magnetoelastic (ME) biosensor has proven to be promising for real-time detection of pathogenic bacteria on fresh produces. The ME biosensor consists of a freestanding ME resonator as the signal transducer and filamentous phage as the biomolecular-recognition element, which can specifically bind to a pathogen of interest. Due to the Joule magnetostriction effect, the biosensors can be placed into mechanical resonance when subjected to a time-varying magnetic field alternating at the sensor’s resonant frequency. Upon the attachment of the target pathogen, the mass of the biosensor increases, thereby decreasing its resonant frequency. This paper presents an investigation of blocking reagents immobilization for detecting Salmonella Typhimurium on fresh food surfaces. Three different blocking reagents (BSA, SuperBlock blocking buffer, and blocker BLOTTO) were used and compared. The optical microscope was used for bacterial cells binding observation. Student t-test was used to statistically analysis the experiment results. The results shows that SuperBlock blocking buffer and blocker BLOTTO have much better blocking performance than usually used BSA.

The European directive 2008/98/CE establishes a legal framework for the treatment of waste within the Community. It aims at protecting the environment and human health through the prevention of the harmful effects of waste generation and waste management. In order to better protect the environment, the Member States should adopt measures for the treatment of their waste according to a hierarchy as outlined: prevention, preparing for reuse, recycling, energy recovery, disposal. In this context, the European project LIFE12 ENV/IT/000356 “RESAFE” is addressed to produce and utilize a new class of fertilizers characterized by reduced salinity in order to substitute chemical and mineral fertilizers through a technological route based on Urban Organic Waste (UOW), Farm Organic Residues (FOR), Bio-Chars (BC) and Vegetable Active Principles (VAP) processing. Following this approach, it will be possible for farmers and urban waste managers to reduce costs and to obtain environmental and economic incomes. Furthermore, environmental impacts will be also reduced contributing to decrease the greenhouse emissions from landfills and from the production of mineral fertilizers. In this paper, specific innovative sensing architectures, based on Hyper-Spectral Imaging (HSI) devices working in the near infrared (NIR) range, and related detection architectures, is presented and discussed in order to define and apply smart detection engines to follow the transformations of the complex material, resulting from UOW, FOR, BC and VAP based recipes during the different stages of the fertilizer production process. Results show as the fertilizer production process can be monitored adopting the NIR-HSI approach.