Biosensor technology for food packaging applications

This article provides an introduction to the emerging technology of biosensors and its potential applications to food packaging. This biosensor technology may used to enhance the communication function of the intelligent packaging system (see the Intelligent Packaging article) to enhance the safety and quality of food products.

NEEDS OF FOOD QUALITY/SAFETY CONTROL

Food quality control is essential in the food industry; nowadays, an efficient quality assurance is becoming increasingly important. Consumers expect adequate quality of food product at a fair price, long shelf life, and high product safety, whereas food inspectors require safe manufacturing practices, adequate product labeling, and compliance with the U.S. Food and Drug Administration (FDA) regulations. Further more, food producers are increasingly demanding the efficient control methods, particularly through online or at-line quality sensors to satisfy consumers’ and regulatory requirements and to improve the feasibility of automated food processing and quality of sorting. Also, food producers are demanding a reduction in the production time (increase throughput) and the final product cost.

Novel sensing technologies using biomaterials or nanomaterials can be used to detect quality and safety attributes in packaged foods. These sensing technologies range from rapid non destructive and noncontact to highly specialized microsensing and nanobiosensing structures. Micro- and nano-based sensors that use a variety of transduction mechanisms to sense microbial and biochemical changes in food products are being explored. Extensive development of biosensors for food safety and quality control were stimulated by acquiring several new food safety and key quality concepts during the last decade, such as Hazard Analysis Critical Control Points (see the HACCP article), Total Quality Management (TQM), and ISO 9000 Certifications. The wave of terrorist acts and foodborne disease outbreaks have raised the importance of the food traceability and authentication (1, 2). There are specific safety problems (pathogenic micro-organisms, BSE, GMF, pollutants, etc.) that require intensive control, data logging, and data treatments, which can be controlled effectively only with the new generations of biodetection systems (3). All these tasks require rapid responce sensors for new integrated data analysis systems and are an indispensable part of the modern supply chain operation paradigm.

There are several possible sources of undesirable contaminations and/or changes in food products that can be combined in five groups by their localization and occurrence. Three of them are food manufacturing-related as follows: technology (processing and sequence of process operations), industrial hygiene (food safety management at the plant level and HACCP), and formulation (product development, interactions of food additives/ingredients with food matrix, and bioavailability). The sources of food raw materials and their quality are the issue of biosafety/biosecurity in the agricultural processing including postharvesting technologies and logistics. The fifth source of biohazards is the environment in the broadest sense, which includes pollution, climate changes, and anthropogenic environmental factors.

The major types of changes in foods are caused by the sources of undesirable contaminants. They can be instrumentally controlled; hence, they represent the primary targets for biosensors development and design. Indeed, the great challenge is to develop the real-time and online sensors and data systems suitable for surveying processes and products, controlling automated processes and the raw material stream, sensing the final products quality, typing the product labels with nutritional and health information, and much more. Today, the most important quality parameters and concepts in food production control are as follows:

• Sensory: appearance, flavor, taste, texture, and stability 
• Nutritional: including health implications, such as ‘‘high in fiber,’’ ‘‘low cholesterol,’’ and ‘‘GMF free’’ 
• Composition and labeling: additives lists as well as quality and ethical claims (e.g., ecological information) 
• Pollutants record: environmental pollutants, veterinary drugs, agricultural chemicals, BSE-prions, and mycotoxins 
• Detection of foreign bodies: such as stones, glass or metal fragments 
• Microbial safety: in particular Listeria, Salmonella, Campylobacter, Escherichia coli and Yersinia 
• Shelf life: microbial, sensory, chemical, sterility testing, and F0-values 
• Production hygiene: cleaning and decontamination; 
• HACCP: traceability and authentication 
• Process parameters control: machine settings, temperature, pressure, flow, aseptic conditions, and many others 
• Packaging: integrity, pinholes, gas permeability, and migration control 

BIOSENSORS: GENERAL FACTS

The potential susceptibility of the food supply chain to natural or intentional contamination could result in compromised safety and quality of foods. Nano-bio sensors and integrated microsystems could play a significant role of detecting deteriorative changes in food packaging. In modern food packaging development, appropriate sensing technologies are required to detect substances in parts per trillion for food safety, quality, and process control. The development of new sensing devices may be achieved by taking advantage of miniaturization of electronics and nano-bio materials. These novel sensing systems can be used to facilitate the online analysis of food stuffs.

Biosensor technology is a powerful alternative to conventional analytical techniques, harnessing the specificity and sensitivity of biological systems in small, low-cost devices. Despite the promising biosensors developed in research laboratories, there are not many reports of real applications in packaging. A sensor is the device that can detect a property or group of properties in a food product and respond to it by a signal, often an electric signal. This signal may provide direct information about the quality factor(s) measured or may have known relation to the quality factor. Usually, sensors are classified according to their mode of use (see Figure 1).

Biosensors usually are small, analytical bioelectronic devices that combine a transducer with a sensing biological component (biologically active substance). The transducer, which is in intimate contact with the biologically sensitive material, can the measure weight, electrical charge, potential, current, temperature, or optical activity of the substance. The biologically active species include enzymes, multienzyme systems, antibodies or antigens, receptors, populations of bacterial or eukaryotic cells, or whole slices of mammalian or plant tissue. Substances such as sugars, amino acids, alcohols, lipids, nucleotides, and so on can be specifically identified and their concentration measured by these sensors. A schematic functional representation of a biosensor and the detection principle is depicted in Figure 2. The biosensor consists of a biological sensing element integrated with a signal transducer; together, they produce a reagent-free sensing system specific for the target analyte. The biological component of a biosensor used for the molecular detection is made of highly specialized macromolecules or complex systems with the appropriate selectivity and sensitivity. Biosensors can be classified according to the biocomponents used for the detection.

The biodetection principle can be schematically described as follows. A chemical, biological, or physical sensor produces a signal (e.g., voltage, absorbance rate, heat, or current) in response to a detectable event, such as binding between two molecules. In case of a biological or chemical sensor, this event typically involves a receptor (e.g., macrocyclic ligand, enzyme, or antibody) binding to a specific target molecule in a sample. Physical sensors, on the contrary, measure the inherent physical parameters of a sample, such as current or temperature, which can change because of reactions occurring in it. In any case, the signal is then transduced by passing it to a circuit where it is digitized. The obtained digital information can be stored in a memory, displayed on a monitor, or made accessible via digital communications port.

Because it is essential that the sensor’s response be detected, it is necessary that an appropriate transduction mode for electrochemical signals, optical signals using changes in the fluorescence or absorbance rate of a sample, or plasmon resonance be available. With most sensors, transduction is accomplished electrochemically or optically.

The transducer transforms the physicochemical variations occurring in the biosensing element as the result of a positive detection event into an electric signal, which is then amplified by an ad hoc designed electronic circuit and used for the control of external devices. The transducers can be electrochemical (amperometric, potentiometric, and conductometric/impedimetric), optical, piezoelectric, or calorimetric. Often, this classification is used to identify the type of a biosensor (see Figure 3).

The bio-specific elements of the biosensor and transducer can be coupled together in one of the four possible ways (4), that are schematically shown in Figure 4: membrane entrapment, physical adsorption, matrix entrapment/porous encapsulation, and covalent bonding.

In the membrane entrapment scheme, a semipermeable membrane separates the analyte and the bio-element, and the sensor is attached to the bioelement [collagen membranes, synthetic pre-activated membranes (5), and cellulose-acetate membranes]. The physical adsorption scheme depends on a combination of van der Waals forces, hydrophobic forces, hydrogen bonds, and ionic forces to attach the biomaterial to the sensor surface (6). The porous entrapment scheme is based on forming a porous encapsulation matrix around the biological material that helps in binding it to the sensor [nylon net (7), carbon paste (8), or graphite composites (9)]. In the case of covalent bonding, the sensor surface is treated as a reactive group to which the biologicalmaterial can bind (10). One of the bioselective elements most frequently used in biosensors is an enzyme. These large protein molecules act as catalysts in chemical reactions but remain themselves unchanged at the end of reaction.

Mechanical (Resonant) Biosensors.

In this type of biosensors, an acoustic wave transducer is coupled with an antibody (biosensitive element). When the analyte molecules (antigens) attach to the membrane (cantilever), the membrane mass changes, resulting in a subsequent change in the resonant frequency of the transducer (11). This frequency change is detected and measured (4).

Sensors Based on Electromagnetic Waves.

Electromagnetic sensors may be classified by the wavelength of the electromagnetic waves they use: visible (VIS) (400–700 nm), ultraviolet (10–400 nm), infrared [700– 30,000 nm: nearinfrared (NIR (12), FTIR (13), MRI (14)] waves, microwaves (15) (1–10 cm), radiofrequency (16) (1–10 m), and X-rays (17) (100pm1 nm). Each sensor class may be subdivided even more according to the molecular information that can be obtained through the interaction. For instance, infrared sensors may be subdivided into near-infrared (700–2500 nm), mid-infrared (2500–30,000 nm), far-infrared (up to 1,000,000 nm), and thermography (1–15 mm) sensors, which all extract different information from the molecules (sample) interacting with the waves. We may also classify these sensors according to their precise type of interaction: absorbance, transmittance, or reflectance of light.

Sensors based on interactions with electromagnetic radiation waves have been on the market for many years, in particular for laboratory purposes. Online examples of such sensors are also numerous: X-rays used for foreign body detection (18), visible light sensors for color recognition or machine vision inspections (19), near-infrared sensors for quality inspection and temperature measurements (20), or microwave sensors for the detection of water content (21).

Optical Detection Biosensors.

Strictly speaking, optical biosensors belong to the larger class of electromagnetic detectors, but because of their importance and broad use, they are usually considered as a separate group of biosensitive devices. The output signal measured in this type of biosensors is a light signal (22). These biosensors can be made based on optical diffraction or electrochemiluminescence (23).

Surface Plasmon Resonance (SPR).

SPR is another optical phenomenon used in new sensors, often in those that involve antibodies or enzymes. The optical range used is most often in the visible part of the spectrum, but it may also be in the NIR range. Traditionally, SPR devices detect minute changes in the refractive index of the sensing surface and its immediate vicinity. They may detect these changes by a diffraction grating, with a prism on a glass slide, or through an optical waveguide carrying a thin metal layer (gold). The metal layer carries a sensitizing layer (e.g., immobilized antibodies or other molecules binding the analyte specifically; this layer is in contact with the sample). Inside the device, a collective excitement of electrons in the metal film occurs and leads at a specific wavelength to a total absorption of light at a particular angle of incidence. This angle depends on the refractive indices on either side of the metal film. Specific molecules binding to the sensitizing layer change the refractive index, which changes an angle of total absorption; this angle is measured and correlated to the concentration of the analyte.

The SPR detection technique has been used by Hellnaes (24) for online and at-line detection of veterinary drug residues (hormones and antibiotics) in dairies and slaughterhouses. Clenbuterol and ethinyl-estradiol in bovine urine, sulfamethazine (SMT), and sulfadiazine (SDZ) in porcine bile, as well as SMT, SDZ, and enrofloxacin in milk have been successfully detected by the technique. The developed biosensor operates in real time and can simultaneously detect up to 8 different veterinary drugs with a throughput of up to 600 samples per day. The project participants have established a new company to produce and develop the sensor systems, and several new and elegant designs of SPR sensors are now under study. The SPR sensor principle has also been used by Patel (25). The sensor developed as a result of this research has been applied to the quantification of mycotoxins, Listeria, and markers for growth hormones [recombinant bovine somatotrophin (rBST)].

Most current research is focused on the NIR/VIS sensors, SPR sensors and nuclear magnetic resonance (NMR) sensors (pulsed and low resolution); some work has also been done on fluorescence sensors, MIR and Raman sensors, Fourier Transform NIR sensors, thermography- based sensors, and sensors that combine two or more sensor principles.

Electrochemical Biosensors.

Electrochemical biosensors are mainly used for the detection of hybridized DNA, DNA-binding drugs, glucose concentration, and so on. The underlying principle of these biosensors is that many chemical reactions produce or consume ions or electrons that in turn cause some changes in the electrical properties of the solution; these changes can be sensed out and measured (26). The electrochemical biosensor can be classified based on the measured electrical parameter as conductimetric, amperometric, or potentiometric (27).

Impedimetric/Conductometric Biosensors.

Many biological processes involve changes in the concentrations of ionic species. Such changes can be used by biosensors, which detect changes in electrical conductivity. The measured parameter is the electrical conductance/resistance of the solution. When electrochemical reactions produce ions or electrons, the overall conductivity/resistivity of the solution changes (28). This change is measured and calibrated to a proper scale. Conductance measurements have relatively low sensitivity. The electric field is generated using sinusoidal voltage, which helps in minimizing undesirable effects such as Faradaic process, double-layer charging, and concentration polarization (29).

Impedimetric biosensors use changes in the electrical conductivity in the frequency domain (impedance) of a biological system for sensing and detection (6, 30, 31). Impedance spectroscopy provides a powerful tool for investigating a variety of bioelectric processes for both electrical and nonelectrical applications. In impedance spectroscopy, the current flowing through a sample cell that contains a nano-scale patterned bio-interface and the voltage across this cell are measured as a function of frequency (32–34). The design of impedimetric sensors is similar to conductivity-based sensors (26, 35–39). Enzyme/ antibody immobilization on electrode surface makes these sensors highly selective and sensitive (40).

Amperometric Biosensors.

This highly sensitive biosensor can detect electro-active species present in biological test samples. Enzyme-catalyzed redox reactions can form the basis of a major class of biosensors if the flux of redox electrons can be determined (41). Normally, a constant voltage is applied between two electrodes, and the current, which is caused by the electrode reaction, is determined. The first and simplest biosensor was based on this principle. It was for the determination of glucose and made use of the Clark oxygen electrode. In the case of amperometric biosensors, the measured parameter is an electric current. Some of the most recent applications of amperometric biosensors include glucose sensor for meat freshness (43); glucose sensor for use in fermentation systems (43); rapid cell number monitor (44); monitor for herbicides in surface waters (45, 46); amperometric enzyme-linked immunosorbent assay (ELISA) method based on the self enzyme amplification system (36); amperometric and novel fluorescent DNA probes (47).

Potentiometric Biosensors.

In this type of sensors, the measured parameter is the oxidation/reduction potential of an electrochemical reaction (48). The simplest potentiometric technique is based on the concentration dependence of the potential, E, at reversible redox electrodes according to the Nernst equation (29) E ¼ E0 þ RT nF ln as; where Eo is the standard redox potential, R is the gas constant, T is the absolute temperature, F is the Faraday constant, n is the number of exchanged electrons of the substance S, and as is the activity of the substance S. Changes in ionic concentrations are easily determined by use of ion-selective electrodes (48). This forms the basis of potentiometric biosensors (30). Many biocatalyzed reactions involve charged species each of which will absorb or release hydrogen ions according to their pKa and the pH of the environment (49). This allows a relatively simple electronic transudction using the most common ion-selective electrode, which is the pH electrode (50).

Field Effect Transistors (FETs) and Ion-Selective Field Effect Transistors (ISFETs).

Potentiometric biosensors can be miniaturized by the use of FET. ISFETs are low-cost devices that are in mass production (51). A recent development from ion-selective electrodes is the production of ISFETs and their biosensor use as enzyme-linked field effect transistors (ENFETs). Enzyme membranes are coated on, the ion-selective gates of these electronic devices, the biosensor responding to the electrical potential change via the current output. Thus, these are potentiometric devices, although they directly produce changes in the electric current. Reference shows a diagrammatic crosssection through an npn hydrogen ion responsive ISFET with a biocatalytic membrane. The buildup of positive charge on this surface (the gate) repels the positive holes in the p-type silicon causing a depletion layer and allowing the current to flow. In Abdelmalek et al. (52), Langmuir- Blodgett films that contain butyrylcholinestrase (BuChE) are fabricated to realize an ISFET for the detection of organophosphorus pesticides in water.

Cell-Based Biosensors.

Cell-based biosensors have been implemented using micro-organisms, particularly for environmental monitoring of pollutants (53). Biosensors that incorporate mammalian cells have a distinct advantage of responding in a manner that can offer insight into the physiological effect of an analyte (54, 55). Several approaches for the transduction of cellular signals (56) are described in the literature: measures of cell metabolism, impedance (57), intracellular potentials, and extracellular potentials (58). Among these approaches, networks of excitable cells cultured on microelectrode arrays (53, 54, 59–61) are uniquely poised to provide rapid, functional classification of an analyte and ultimately constitute a potentially effective cell-based biosensor technology. Keese and Giaever (62) have designed a biosensor that can be used to monitor cell morphology in tissue culture environment. The sensing principle used is known as electric cell-substrate impedance sensing (ECIS). In this process, a small gold electrode is immersed into the tissue culture medium. After cells attach and spread over the electrodes, the electric impedance measured across the electrode chamber changes. These changes in impedance can be used for understanding cell behavior in the culture medium. The attachment and spreading of the cells are important factors for successful use of this biosensor. Unfortunately, some types of cells, e.g., cancerous cells, can grow and reproduce freely in a medium without being attached to any substrate/surface, that makes them impossible to detect with these sensors.

Cornell et al. (63) proposed biosensor mimics biological sensory functions and can be used with most types of receptor, including antibodies and nucleotides. The technique is flexible and even in its simplest form it is sensitive to pico-molar concentrations of proteins.

Lab-on-a-Chip Systems and DNA Detection Devices.

Significant advances have been made in the development of micro-scale technologies for biomedical and drug discovery applications. The first generation of microfluidicsbased analytical devices [Lab-on-a-Chip (64)] have been designed and are already functional. Microfluidic devices offer unique advantages in sample handling (65–67), reagent mixing (68–70), separation (71–73), and detection (74). These devices include, but are not limited to are devices for cell sampling (75), cell trapping and cell sorting devices (59, 76–79), flow cytometers (67, 80, 81), devices for cell treatment: cell lysis, poration/gene transfection, and cell fusion devices (82).

Biosensors used for DNA detection are used to identify small concentrations of DNA (of micro-organisms such as viruses or bacteria) in a large sample. The detection relies on comparing sample DNA with a DNA of known microorganism (probe DNA) (83). Because the sample solution may contain only a small number of micro-organism molecules, multiple copies of the sample DNA need to be created for proper analysis (84). This is achieved with an aid of the polymerase chain reaction (PCR). PCR starts by splitting the sample’s double-helix DNA into two parts by heating it. If the reagents contain proper growth enzymes, then each of these strands would grow the complementary missing part and form the double-helix structure again. This happens after the temperature is lowered. Thus, in one heating/cooling cycle, the amount of sample DNA is doubled (85). In general, PCR is power consuming, so previously it was not possible to fabricate portable biodetectors that can perform PCR. But, using newly developed MEMS devices, such biodectors (also known as lab-on-achip systems) have been created. In these MEMS-based devices the amount of reagent used is scaled down (86).

DNA-Based Sensors/Assays.

The general principle of DNA probe assay is similar to the immunoassay described in. Indeed, even the applications of DNA probes and monoclonal antibody immunoassay frequently overlap, thus establishing a ‘‘competition’’ between the two possible approaches.

One of the most important applications for DNA probes is the testing for virus infections (87). For probes of infectious disease, it is assumed that all strains can still contain a common DNA sequence region and thus be identified by a single probe. Recognized by the cell as a foreign body, viruses will induce an antigenic reaction causing antibody generation so they can also be detected in an immunoassay (88).

Another type of biosensor developed by the Naval Research Laboratory (85) uses magnetic field instead of optics or fluorescence. This sensor equipped with magnetic sensors and microbeads (89) can detect the presence and concentration of bioagents. The magnetic sensor (group of sensors) is coated with single-stranded DNA probes specific for a given bioagent or sample DNA. Once a single strand of DNA probe and a single strand of sample DNA find each other, they form a double-stranded (doublehelix) structure, which in turn binds a single magnetic microbead. When a magnetic bead is present on a sensor surface, its resistance decreases, which can be detected and measured.

BIOSENSORS AND FOOD PACKAGING

Integrated Sensor-Packaging Systems

The basic function of sensors-packaging system is to monitor the package environment, product state/status, and perform data exchange with the external databases, providing information for decision making. Intelligent packaging is the integrated system of a complex structure with discrete/distributed sensor and data carrier elements. Technical realization of such a packaging system includes integration of sensitive elements into the packaging materials and/or labels, and integration of sensors’ data into the intelligent packaging information flow, i.e., data layer. Such integration allows to perform online and ‘‘on-shelf’’ control of the internal and/or external package environments. Biosensors can be created by chemical modification and sensibilizing of packaging labels, embedding microdevices (‘‘smart chips’’), or depositing microand nano-sensitive elements into the label or the package itself. In the future, this system can be used to trigger the controllable release of bioactive components and targeted delivery of protective agents into the food product. One of the most important elements of the Internet Protocol (IP) concept is integration of sensors and detectors into the package itself. Examples of monitoring activities include visual observations/control and measurements of temperature, time, pH, and moisture levels. In the future, this system can be used to trigger the controllable release of bioactive components and the targeted delivery of protective agents into the food product.

Based on the type of the sensor and their location, it is possible to recognize three major types of sensor-packaging systems (see Figure 5):

• Off-package — remote sensing devices without requirements of direct contact with the package materials of packaging content, i.e., various spectroscopic devices (Raman, IR, NIR, and radiofrequensy). 
• On-package — truly integrated devises that located on the packaging itself. RFID devices can be related to this category. Sensing packaging materials, which are sensors with distributed parameters, also belong to this type of sensors. 
• In-package — sensing device is placed inside of the package. In some cases sensor can be part of the product formulation. 

Integrated Sensor-Packaging Systems May Enhance Food Safety

Existing food distribution network includes producers, logistic operators, and processors. Timely detection of unsafe foods entering this network is the main issue that food safety system should address. The HACCP system, which has become the industrial standard for food safety management is designed to identify health hazards and to establish strategies and procedures to prevent, eliminate, or reduce their occurrence. Each stage of the food products flow is individually controlled, but there is no integrated food safety system that combines material and informational flows into one continuous food safety management process. There is a need to enhance the existing HACCP procedures by integration of material and information flows and to ensure early detection of deliberate food contamination at any point along the production pathway. In general the intelligent packaging framework is perfectly adapted to be used in the HACCP workflow and uses the principle of information system cyclic interactions with the environment (Figure 6).

The intelligent packaging framework enhances the existing HACCP system by incorporating its virtual data layer into the food safety management structure. Intelligent packaging can be integrated into the HACCP system as a data carrier, which delivers proper information about product history, establishes critical limits, and acts as a self-reporting data tag that creates natural link with an external knowledge base. Hence, the Critical Control Points that implement elements of an intelligent packaging will receive new functionality. Combining data of product properties and proper storage regimes with the data generated by the packaging sensors, the IP system can direct the distribution/transportation network on proper operation sequence of product handling. In-time correction actions, continuous monitoring of product environment/ quality, and record keeping are the main functionalities of the HACCP that can be realized within the intelligent packaging conceptual framework. Established critical limits can be stored in product labels. If progressive scan of sensors’ data indicates the difference between established critical limits and current food product properties, then corrective actions will be performed. Intelligent packaging will be able to participate in these actions providing timely and sufficient information necessary for decision making.

An advanced food safety management system based on intelligent packaging will be able to correctly (a) identify potential hazards; (b) identify hazards which must be controlled; (c) conduct a biohazard analysis; (d) recommend controls, critical limits, and procedures for monitoring and verification; and (e) recommend appropriate corrective actions when a deviation occurs. This integrated food safety system combines data from multiple sensors (from different packages and/or products) and relates information about food environment and process conditions. Intelligent packaging is a unique entity that creates a dynamic link among the food product, environment, safety management system, and business operations. Its universality is determined by integration of information and data flows in one system.

APPLICATIONS OF BIOSENSORS IN FOOD SCIENCE AND FOOD PACKAGING

The need for fast, online, and accurate sensing, e.g., in situ analysis of pollutants in crops and soils, detection and identification of infectious diseases in crops and livestock, online measurements of important food processing parameters (90), monitoring animal fertility, and screening therapeutic drugs in veterinary testing are well described in another work (91).

Sensors for Pathogens Detection

The broad spectrum of foodborne infections keeps changing dramatically over time, as well-known pathogens have been controlled or eliminated, and the new ones have emerged. The burden of foodborne diseases remains substantial: One in four Americans is estimated to have a significant foodborne illness each year. Most of these illnesses are not caused by known pathogens, so more of them remain to be discovered. Among the known foodborne pathogens, the recently identified predominate, suggesting that as more and more is learned about pathogens, they would come under control. In addition to the emergence or recognition of new pathogens, other trends include global pandemics of some foodborne pathogens, the emergence of antimicrobial resistance, the identification of pathogens that are highly opportunistic and affect only the most high-risk subpopulations, and the increasing identification of large and dispersed outbreaks. New pathogens can emerge because of changing ecology or technology that connects potential pathogen with the food chain. They also can emerge by transferring the mobile virulence factors, often through bacteriophage (92).

Over the past decade, many improvements have been observed in both conventional and modern methods of pathogenic bacteria detection in foods (26). Modification and automation of conventional methods in food microbiology involve sample preparation, plating techniques, counting, and identification test kits. Adenosine triphosphate (ATP) bioluminescence techniques are increasingly used for measuring the efficacy of surfaces and utensils cleaning. Cell counting methods, including flow cytometry, and the direct epifluorescent filter technique are suitable for rapid detection of contaminating micro-organisms, especially in fluids. Automated systems based on impedance spectroscopy can screen high numbers of samples and make total bacterial counts within 1 day. Immunoassays in various formats make a rapid detection of as many pathogens as possible. Recently, there have been important developments in the nucleic acid-based assays and their application for the detection and subtyping of foodborne pathogens. The sensitivity of these methods has been significantly increased by employing the polymerase chain reaction and other amplification techniques. Alternative and rapid methods must meet several requirements concerning accuracy, validation, speed, automation, sample matrix, and so on. Both conventional and rapid methods are used in the frame of biohazard analysis critical control point programs. Additional improvements especially in immunoassays and genetic methods can be expected, including applications of biosensors and DNA chip technology (93).

In recent work by Bokken et al. (94), a surface plasmon resonance biosensor was used to detect Salmonella pathogen through antibodies reacting with Salmonella group A, B, D, and E (Kauffmann-White typing). In the assay designed, anti-Salmonella antibodies immobilized onto the biosensor surface were allowed to bind injected bacteria, followed by a pulse with soluble anti-Salmonella immunoglobulins to intensify the signal. No significant interference was found for mixtures of 30 non-Salmonella serovars at 109 CFU mL1. A total of 53 Salmonella serovars were successfully detected at 107 CFU mL1, except those from groups C, G, L, and P, as expected.

Another sensor technology recently developed uses a micro-electrophoretic system (mFFE) that separates and concentrates the analyte in question by several electrophoretic methods: preparative zone, interval zone, isotachophoresis, or isoelectric focusing. The mFFE system can be designed as a plain glass substrate 1.5mm thick, and a cross-linked polydimethyl-siloxane (PDMS) top layer with micromachined sample channels. The central separation chamber (1240.15mm) is connected to 34 inlet channels for sample injection, and 36 outlet channels for sample collection. The detector unit can be based on several principles. In the case of Listeria, the detector unit may be a well-known ATP luminescence detector. For other analytes, the SPR detection system may be used with immobilized bio-specific layer, e.g., antibodies (95).

New ion-channel biosensor based on supported bilayer lipid membrane for direct and fast detection of Campylobacter species has been reported (96). The sensing element was composed of a stainless-steel working electrode, covered with an artificial bilayer lipid membrane (BLM). Antibodies to bacteria embedded into the BLM are used as channel-forming proteins. The biosensor has a strong signal amplification effect, which is defined as the total number of ions transported across the BLM. The biosensor has demonstrated a good sensitivity and selectivity to Campylobacter species.

A novel assay system for the detection of E. coli O157:H7 has been recently developed. The detection is based on the immunomagnetic separation of the target pathogen from a sample and absorbance measurements of p-nitrophenol at 400nm from p-nitrophenyl phosphate hydrolysis by alkaline phosphatase (EC 3.1.3.1) on the ‘‘sandwich’’ structure complexes (antibodies coated onto micromagnetic beads – E. coli O157:H7-antibodies conjugated with the enzyme) formed on the microbead surface (97). The selectivity of the system has been examined, and no interference from other pathogens including Salmonella typhimurium, Campylobacter jejuni, and Listeria monocytogenes was observed. The sensor’s working range is from 3.2102 to 3.2104 CFU/mL, with the relative standard deviation of 2.5–9.9%. The total detection time is less than 2 h.

An improved antibody-coated sensor system based on quartz crystal microbalance analysis of Salmonella spp. has been developed using thiolated antibody immobilization onto the gold electrode of the piezoelectric quartz crystal surface (98). The best results in sensitivity and stability were obtained with the thin layer of a thiolcleavable, hetero-bifunctional cross-linker. The long bridge of this reagent can function as a spacer, facilitating antibody-Salmonella interaction on the gold electrode. The sensor’s response was detected for the microbial suspension concentrations that ranged from 106 to 1.8108 cfu/mL.

A label-free immunosensor for the detection of pathogenic bacteria using screen-printed gold electrodes (SPGEs) and a potassium hexacyanoferrate (II) redox probe has been reported by Susmel et al. (99). Gold electrodes were produced using screen printing, and the gold surfaces were modified by a thiol-based self-assembled monolayer (SAM) to facilitate antibody immobilization. In the presence of analyte, a change in the apparent diffusion coefficient of the redox probe was observed, that can be attributed to impedance of the diffusion of redox electrons to the electrode surface caused by the formation of the antibody–bacteria immunocomplexes. No change in the diffusion coefficient was observed when a nonspecific antibody [mouse immunoglobuline (IgG)] was immobilized and antigen added. The system has been demonstrated to work with Listeria monocytogenes and Bacillus cereus.

Sensors to Monitor Food Packaging and Shelf Life

In recent work by Yano et al. (100), a cell-based biosensor has been used to control meat freshness. Samples of fresh meat stored at 51C were periodically removed from storage and washed with water for periods of up to 2 weeks. The water was then charged into a flow-injection analysis (FIA) system combined to the microbial sensor using yeast (Trichosporon cutaneum) as a sensitive element. This sensor has been specifically developed in this work for monitoring the freshness of meat. Relationships among the sensor signals obtained by the FIA system, the amounts of polyamines and amino acids produced from the meat, and the number of bacteria that had been multiplying in the meat during the aging process were investigated. The sensor response has been found to correspond to the increase in amino acid levels and viable counts in the meat during the first stage of aging. This is because amino acids produced initially by enzymes in the meat serve as a nutrition source for septic bacteria, and as a result, the amount of bacterial cells increases with increasing level of amino acids.

Foreign Body Detection.

The presence of foreign bodies in processed food is of major concern to the producers. Mechanical separation techniques based on size and weight of different components have been used for many years to help finding foreign bodies in powdered and flowing products. Optical inspection techniques were able to extend the range of detectable foreign objects in free-flowing materials with regard to their shape and color. Metal detectors enabled metallic particles inside the product to be found. With recent achievements in sensor technologies advanced foreign body detection systems are becoming available (101).

The working principle and design of an ultrasonic transducer system with auto-alignment mechanism was first described by Zhao et al. (102). The proposed system has been used for detecting foreign bodies in beverage containers. Variations in reflection amplitude were analyzed as a function of the ultrasound beam incident angle to the beverage container surface. It has been concluded that a quadratic relationship exists between the strength of the reflected signal and the incident angle. Furthermore, a calculation for effective angular increment for searching the normal to a curved surface was introduced. Experiments conducted using the sensor prototype have demonstrated that foreign bodies are detectable in containers of various juices. This sensor design is also applicable to the non destructive inspection of canned food products for the foreign bodies presence.

Biosensors for Food Quality/Additives Control

Existing food packaging and packaging equipment may include microprocessors that are activated by electronic or biological sensors. Recent advances in electronic vision and computer technology have opened the research horizons for greater accuracy in process control, product sorting, and operation. The development of new sensors and instruments in this area is focused on measuring/ evaluating the product internal and external quality and flavor (91).

The aim of food additives control and measurement is to develop, extend, and enhance the instrumental methods to improve consumer-perceived macroscopic quality factors. Several types of electronic sensors for quality assessment, grading, and sorting of food products have been investigated and described in literature.

A near infrared sensing technique can rapidly determine the sugar content of intact peaches. This technology has been extended to many other commodities, including testing avocados for oil content, and kiwifruits for starch and sugar content. The NMR method, for example, can be used for nondestructive detection and evaluation of internal product quality factors, such as the existence of bruises, dry regions or worm damage, stage of maturity, oil content, sugar content, tissue breakdown, and the presence of voids, seeds, and pits.

The machine vision for postharvest product sorting and grading is being investigated for several commodities. Recent research has included the development of a highspeed prune defect sorter, color and defect detector for fresh-market stone fruits, raisin grading, and flower grading machines. In this technology, electronic cameras are used for monitoring the product in various packing-line handling situations. Quality features are computed from digitized images, and the control system allows for product grading and sorting. The NIR/VIS region has been used in several different sensors. Thus, Crochon (103) has presented the design of a glove-shaped apparatus equipped with various miniaturized sensors providing information on fruit quality parameters, i.e., sugar content, maturity, mechanical properties (firmness and stiffness), and internal color. The sugar content and internal color were measured by a miniaturized spectrometer (NIR/VIS) coupled with optical fibers. A sound sensor evaluated the mechanical properties, and the size was measured by a potentiometer placed at the hand aperture. These sensors were coupled to a microcomputer that delivered processed information about the fruit overall quality grade, based on previously established variety and quality classes. The weight of the glove prototype was 400 g, and the electronic devices were held in a rucksack weighing 1000 g. The glove may be used before harvest to control the growth and to estimate the harvest date, at harvest to select fruits with specific qualities, or after harvest to control and measure the quality of the crop.

In Baendemaeker (104), chlorophyll fluorescence and reflectance in the NIR/VIS spectrum has been used for the mechanical quality factors assessment of green beans, broccoli, and carrots. Biosensors have been used for evaluation the effects of pasteurization on the vegetables quality by measuring the remaining enzymatic activity. Using MIR spectroscopy, as well as Raman scattering for online quality assessment in bakeries, breweries, dairies, and fruit farms has been reported (105).

Another method working in the NIR/VIS range, called time-resolved diffuse reflectance spectroscopy (TDRS) has been used to measure the internal quality of fruits and vegetables (106). The group has developed statistical models for the analysis of relationships between the TDRS signals and the firmness, sugar, and acid content of kiwifruit, tomato, apple, peach, nectarine, and melon. They have also developed the classification models to sort apples, peaches, kiwifruits and tomatoes into quality classes. Using a pulsed laser diode (70–200 ps/pulse), the single measurement time was about 100 ms. The absorption coefficient was related to the tissue constituents, whereas the scattering coefficient was related to the firmness and fiber content. A real-time sensors in the NIR/VIS range can be also used to measure product quality traits, such as maturity, flavor, or internal diseases and defects in potatoes, apples, and peaches (107).

Among the optical sensor systems developed and demonstrated in industrial environments are machine or artificial vision sensors. The system for olives sorting, using a traditional vision camera and three CCD color sensors for the shape, size, and color evaluation has been described (108). The new algorithm allowed the olives to be sorted into four classes with the speed of 132 olives/sec, and 6 images/sec.

The molecular imprinted polymers (MIP) technology is the new technology used for the development of biosensor substrates (109). The polymers are produced by imprinting the recognition sites of predetermined specificity into cross-linked synthetic polymers. The polymer is consequently able to selectively rebind the imprinted molecule (110). These sensor materials are called ‘‘artificial antibodies’’ (111). The MIP technology has particular strengths for small molecular analytes up to about 400 Dalton; it may be used to bind and detect many chemicals polluting food products, e.g., pesticides and veterinary drugs in meat and dairy products.

This technology has been successfully employed to develop and optimize plug-in detection cartridge supporting the molecularly imprinted polymer assay (112) for detection of different b-lactam antibiotics in milk. The sensor consists of a microfabricated column accommodating an optical detection window. Molecular imprinted polymers in the form of beads were used as packing materials and recognition elements; analyte binding was detected by the fluorescence. The same MIP technology has been used in several other studies, the overall objective of which was to develop novel and robust MIP-based technology that can be used in sensors for real-time measurements of food product contaminants (113–115). The results of the study indicate that MIP can be used to prepare both selective and general recognition matrices for either individual analytes or groups of compounds , with very good detection reproducibility and stability (115). SPR based sensor shows similar results for dairy product quality applications (116). The MIP developed for clenbuterol has been successfully applied in preparing a novel sensor comprising MIP as the selective element and amperometric detector as the transducer (109). The responses from several sensors were determined to have a variability of 10%. The feasibility for an oxacillin MIPbased sensor was also demonstrated.

At-line immunological sensors using amperometric detection of the resulting antibody-antigen complexes were described (117). The target quality factor assessed in this project was the presence of toxic chlorophenolic fungicides and their chloroanisole breakdown products in potable water, wine, and fruit juices. The electrochemical immunosensor uses monoclonal antibody preparations. The investigations of the effects of liquid food matrices on electrochemical transduction processes indicated that horseradish peroxidase is a suitable label for interrogation of the analyte–antibody immune complex, using amperometry and in-house fabricated screen-printed electrodes. The detection of hormonal substances for growth promotion, which is also based on immuno-sensors, has been recently reported by Guilbault (118). The sensor has to be used prior to slaughtering, and it can detect and measure testosterone, methyltestosterone, 19-nortestosterone, stanozolol, and trenbolone levels in biological fluids (blood). Analysis time achieved was about 30 min, compared with 24–36 h for tests used in laboratories today.

Biosensors for Sensory Evaluation of Food Products

‘‘Electronic noses’’ (119, 120) and ‘‘electronic tongues’’ (121) are the common names of devices responding to the flavor/odor (volatiles) or taste (solubles) of a product using an array of simple and nonspecific sensors and the pattern-recognition software system (122). Historically, the sensors used were advanced mass spectrometers or gas/liquid chromatographs that produce unique fingerprint of the analyte. Nowadays, these sensors have been substituted by arrays of simple electric and/or frequency sensors, or sensors measuring changes in voltage or frequency as a response to the food contact.

Electronic noses and tongues are used in food production and quality control of different products, typically for laboratory tests or at-line control but may be further developed for in-line operation in the future. Testing times are often in the range of a few minutes, and the largest drawback of these devices is the lack of sensor stability. Examples of claimed successful applications include (123) the following:

• Discrimination between single volatile compounds 
• Tracking of aroma evolution of ice-stored fish or meat 
• Tracking of the evolution of cheese aroma during aging 
• Classification of wines 
• Determination of boar odor (androsterone) in pork fat 
• Classification of peaches and other fruits 
• Differentiation of spices by the area 
• General raw materials control 
• Testing of coffee, soft drinks, and whiskey 
• Control of beer quality and faults 

Essentially, each odor or taste leaves a characteristic pattern or fingerprint on the sensor array, and an artificial neural network is trained to distinguish and recognize these patterns (see Pattern recognition is gained by building a library of flavors from known flavor mixtures given to the network. Thus, e-noses and tongues are the devices intended to simulate human sensory response to a specific flavor, sourness, sweetness, saltiness, bitterness, and so on (123, 124).

The potentiometric chemical sensors such as ion selective sensors are most often used in the electronic noses. Considerable interest exists in the development of cheap, portable electronic noses to detect, online or at-line, the odor quality of many foods. For instance, olive oil producers would tremendously benefit from the possibility of detecting oil quality and shelf-life, and classifying the oils by their quality (e.g., extra virgin olive oil). This was the objective of a project in course of which scientists from olive producing countries have developed electronic noses especially for the olive production plants and tested them with great success (125).

In Dutta et al. (126), different tea samples were used to evaluate the applicability of electronic noses for sensory studies. A metal oxide sensor-based electronic nose has been used to analyze tea samples with different qualities, namely, drier month, drier month again overfired, well fermented normal fired in oven, well fermented overfired in oven, and under fermented normal fired in oven.

Electronic tongues are also widely used to assess the taste quality of various products. An electronic tongue based on voltammetry measurements, and a multichannel lipid membrane taste sensor based on potentiometry were compared using two aqueous solutions: detergent and tea (127). The electronic tongue consists of four electrodes made of different metals, a reference electrode, and a counter electrode. The measurement principle is based on pulse voltammetry technique in which an electric current is measured during the amplitude change of the applied potential. The taste sensor consists of eight different lipid/polymer membranes. The voltage difference between the electrodes and an Ag/AgCl reference electrode is measured when the current is close to zero. The multichannel electrochemical (potntiometric) sensors have demonstrated better sensitivity, faster dynamic response, but lower reproducibility of the results.

In study performed by Legin et al. (128) the electronic tongue based on a sensor array comprising 23 potentiometric cross-sensitive chemical detectors, and pattern recognition and multivariate calibration data processing tools has been applied to the analysis of Italian red wines.

Biosensors and Biosecurity

Food industry is one of the major potential targets for bioterrorism. The most damage can be attained through (a) final product contamination using either chemical or biological agents with an intent to kill or cause illness among consumers; (b) disruption of food distribution systems; and (c) damaging the food producing cycle by introducing devastating crop pathogens or exotic animal diseases such as foot-and-mouth disease, which could severely impact the food system.

Efforts to develop recognizing preparedness and response strategies for protecting the nation’s food supply pose substantial challenges for many reasons, including the following (129–131):

• The food system encompasses many different industries. 
• A great variety of biological and chemical agents could potentially contaminate the food supply, and the possible scenarios for deliberate contamination are essentially limitless. 
• The public health system is complex, and responsibilities for foodborne diseases prevention and control may overlap, or much worse, fall in the ‘‘gray area’’ between authorities of different agencies. 

To achieve an adequate food supply chain and agricultural security, the improvement is needed in the activities on bioterrorism prevention, detection, and response. In addition, appropriate areas for applied research must be identified as follows:

• Recognition of a foodborne bioterrorism attack. This may be delayed because of background levels of foodborne diseases and potential wide distribution of the contaminated product or ingredient. 
• Rapid diagnostic methods for identifying food-contaminating agents. They are not yet consistently available, and coordinated laboratory systems for pathogens detection are not fully operational. 
• Rapid trace-back procedures for potentially contaminated products. 

Biosensors and HACCP

Timely detection of unsafe foods is the main issue that the food-safety system should address, providing guidance for the design and integration of such system into the existing food safety management structures, i.e., HACCP. The preventive detection of the biohazard can be accomplished by direct measurements with the biosensors, or indirect detection by the process/environment monitoring and control. Such detection is based on the data from physical and chemical sensors, which are reliable and allow scaledown (meaning the possibility of easy integration into the existing information carriers). The HACCP system for food-safety management is designed to identify health hazards and to establish strategies to prevent, eliminate, or reduce their occurrence. An important purpose of corrective actions is to prevent potentially hazardous foods from reaching consumers. Where there is a deviation from the established critical limits, corrective actions are necessary. Therefore, corrective actions should include the following elements: (a) determine the disposition of noncompliant product, (b) determine and correct the cause of noncompliance, and (c) record the corrective actions that have been taken.

Currently, the use of HACCP is voluntary, but it is widely used in the food processing industry as a successful component of comprehensive food safety program. HACCP is a food safety management system in which food safety is addressed through the analysis and control of biological, chemical, and physical hazards from raw material production, procurement, and handling, to manufacturing, distribution, and consumption of the final product. The terms ‘‘HACCP’’ and ‘‘food safety’’ are used interchangeably in the food industry, implying that HACCP may be the only approach to achieving food safety. HACCP is designed for use in all segments of the food industry from growing, harvesting, processing, distributing, and merchandising, to preparing food for consumption (132).

However, there is a need for enhancement and integration of existing HACCP system into the total quality management system and food safety/biosecurity management on higher levels. This system currently includes the mechanisms to decrease the potential for contamination of or damage to the food supply from farm to table (i.e., prevention activities); systems to ensure early detection of deliberate food contamination at any point along the production pathway, including surveillance, rapid laboratory diagnostic and communication systems; and systems to ensure a rapid and thorough response if a bacterial contaminant is detected, including protection of workers and consumers (i.e., emergency response, control, traceback, and mitigation activities).

The ultimate goal is the integration of sensors and sensor networks into the food-safety management structure. Such integration will allow to perform online and ‘‘on-shelf’’ control of the internal and/or external food product quality and package environment.

The integrated sensor information system combines data from multiple sensors (from different packages and/ or products) and the information about environmental and process conditions to achieve highly specific information that cannot be obtained by using a single, independent microbiological assay. The emergence of new information carriers and advanced processing methods will make the food-safety management system increasingly dependable. A successful biohazard detection system should be able to (a) identify potential hazards; (b) identify hazards, which must be currently controlled; (c) conduct hazard analysis; (d) recommend control factors, critical limits, and procedures for hazard monitoring and verification; and (e) recommend appropriate corrective actions if a deviation occurs.

Based on a comprehensive model for multisensor data processing, developed by the U.S. Joint Directors of Laboratories (JDL) Data Fusion Group on Department of Defense (DoD) request (133), the integrated concept of multiple sensors data processing has been developed for the existing HACCP system of food safety monitoring and bio-hazard prevention. This model is specifically adapted to the HACCP workflow and uses the principle of information system cyclic interaction with the environment. The four major steps, including observation/ detection, hazard recognition, decision making and corrective actions strictly correspond to the seven HACCP principles. Integration of such system does not require the redesigning of existing manufacturing and control processes.

The new integrated sensors can monitor the HACCP control points with corresponding material packaging flow on a continuous basis or with predetermined monitoring frequency. Statistically designed data collection or sampling systems lend themselves to this purpose. Issues that need to be addressed when considering implementation of an integrated food safety monitoring system include where the system would be established; how it would be funded; how the data would be generated, analyzed, summarized, and disseminated; and how ‘‘snap surveys’’ could be used as a part of the system.

Microbiological tests are rarely effective for food safety monitoring because of their time-consuming properties and problems with assured detection of contaminants. Physical and chemical measurements are preferred because they are rapid and usually more effective for the control of microbiological hazards. For example, the safety of pasteurized milk is based on the measurements of heating time and temperature rather than on testing the processed milk for the absence of surviving pathogens.