Additional Applications

1. Nanoarray

We have created a technique for fabricating fiber optic-based nanoarrays for detection and analysis for DNA sequences. The fabrication of nanoarrays with two size formats is first discussed: one array has 700 nm features and another with 300 nm features. These arrays have an ultra high packing density in that they can contain 1 x 10 6 array elements/mm 2 (the 700 nm nanoarray) or 4.5 x 10 6 array elements/mm 2 (the 300 nm nanoarray). A straightforward etching protocol was used to create nanowells on the fiber bundle surface into which sensors were deposited. The surface characteristics of the etched arrays were examined with atomic force microscopy and scanning electron microscopy. Fluorescence microscopy was used to observe the arrays. The 300 nm array features and the 500 nm center-to-center distance approach the minimum feature sizes viewable using conventional light microscopy.

A random, multiplexed array composed of oligonucleotide-functionalized nanospheres was then fabricated on the proximal face of the etched fiber bundle. Arrays were used for parallel detection and analysis of fluorescently labeled DNA targets complementary to probe sequences on the nanosphere sensors. We have used these arrays to detect a variety of target sequences such as bacillus thuringiensis kurstaki and vaccina virus , biowarfare-related target sequences, as well as cystic fibrosis related target sequences.

J. M. Tam, L. Song, and D. R. Walt “Fabrication and Optical Characterization of Imaging Fiber-Based Nanoarrays” Talanta ( accepted) (2005)

2. Cross-reactive Enzymes

Many enzymes are considered selective; for example, L-glutamate oxidase reacts primarily with L-glutamate. Other enzymes are considered cross-reactive, such as, L-amino acid oxidase, which reacts with a range of L-amino acids at different rates. The incorporation of these non-specific enzymes into a sensor array is our approach to using the enzyme's inherent cross-reactive nature to create sensor arrays. The designed enzymatic assay combines esterases, esters, and a pH sensitive dye in the 96-well microtiter plate format to create a rapid and convenient fluorometric ester assay. The esterase hydrolyzes an ester causing a change in pH, which results in a unique fluorescence response. The rates of reaction were used to group the analytes through principal component analysis, and, so far nine esterases can cluster over twenty esters.

Enzyme Channeling/Compartmentalization

The goal of channeling is to create a hands free system for multiple enzyme catalyzed reactions. Channeling enzymes is a means to assembling multiple layers of enzymes in a manner, which dictates the order of reaction. Enzymes can be attached onto unreactive microspheres/nanospheres which can in turn be self-assembled. This forces an introduced substrate to react in a predetermined order to form a desired product. Using the known high affinity between avidin and biotin, enzymes are attached onto unreactive spheres via the biotin-avidin interaction. Figure 1 shows a schematic of the first two layers on one microsphere.

Figure 1 Self-assembled bienzyme structure.

Once assembled, substrate can be added to the multienzyme structures to begin a chain reaction seen below.

Enzyme Based Fiber Optic Sensor

The goal of this project is to create and deposit an array of sensing elements into the etched wells of an optical fiber bundle for the detection of specific compounds. Utilizing enzymes will be advantageous for two reasons. First, the target molecule will be degraded into products, which will elicit a chemical change to the sensor environment and allow for detection via a fluorescent dye. Second, because the enzyme is a catalyst, it can be reused many times. A specific class of enzymes called hydrolases degrade their substrate into products that include acid. This evolution of acid causes a pH change to occur at the sensor, which can be detected using a variety of pH sensitive dyes, fluorescein, for example. Placing the sensors in buffer of the starting pH will recharge the dye intensity.

Figure 2 A simple schematic showing the basics of sensor synthesis.

3. Explosives Detection

As an extension to the artificial nose technology which is being investigated in our laboratories, Department of Defense agency has supported research geared towards the detection of explosives vapor, specifically to be developed for landmine detection.  There is a tremendous need for a field-worthy explosives vapor detection system which can detect and discriminate between explosives compounds and the complex backgrounds within which they occur.  At present, a trained dog's nose is the best technology for field detection of land mines. However, dogs cannot cover substantial land area and they are limited by numerous factors including fatigue.  Research and progress made in this area are geared to ultimately be fused into a field-ready landmine detection system which could possible be housed in a soldier's or even a farmer's backpack.  In a recent article, Chemical and Engineering News (C & EN ; March 10, 1997; pp.14-22)   speaks of the global threat of landmines and along with artificial nose technology, the article highlights other methods currently being investigated by other research institutions for landmine detection.

Figure 1 Prof. David R. Walt's Lab at Tufts University hopes to someday detect 2,4,6-trinitrotoluene (TNT) and other explosive-like vapors in the field using optical sensors and neural network analysis.

Main Objective

The main objective in our laboratory is to develop sensitive and semi-selective optical chemosensors for artificial nose 'sniffing' of TNT and TNT-like vapor (Figure 2) in the surface-crust above a buried landmine.  The artificial nose employs neural network analysis and has been shown to discriminate between many organic vapors using pattern recognition once a training set of analytes is used (see Dickinson et al  publications ).  An array of differentially-reactive sensors produces a fluorescence-temporal signature pattern when a specific analyte or analyte mixture is presented in pulsatile fashion.  This signature is mapped and is used to train a system on a specific set of analytes. Optical sensors which respond to 'sniffed' TNT and other nitroaromatic compounds can potentially be used in an array format for which neural networks can be trained for subsequent analysis.  The goal is to detect "a needle in a haystack" and with such low vapor concentrations, signal processing schemes may increase the likelihood of detecting part-per-billion (ppb) and sub-ppb TNT levels. The difficulty in detection methods include sensitivity, selectivity, sensor reversibility, speed of detection (real time?), dynamic range, sensor stability, and durability.  Using a non-traditional, analytical approach with fluorescent chemosensors to detect these ëneedlesí could ultimately lead to a new field detection method based on principles of the biological system (dog's nose).  Our cross-reactive sensor array approach may be very useful for eventually detecting landmines in real time.

Figure 2 The chemical structures for TNT and TNT-like compounds used in our laboratories.

Explosive Sensitive Sensors and Sensor Arrays

Once sensors are fabricated,* they are tested using a vapor delivery system which purges the headspace of sealed flasks containing solid explosives-like nitroaromatic compounds (NACís).  A saturated explosive vapor plug can be delivered to the sensor(s) in pulsatile form without raising temperatures.  Prior to delivery of analyte vapor onto a sensor, a diluent air stream can be used to lower vapor concentrations below saturation.  Preliminary work has shown that some sensors respond to compounds structurally similar to TNT, e.g. 2,4-dinitrotoluene (DNT), 2,6-DNT, 1,3-dinitrobenzene (1,3-DNB), and p-nitrotoluene (4-NT) at saturated vapor concentrations (Figure 3).  These preliminary results provide grounds for exploring the capacity of these and other new polymer/dye sensing combinations for detecting NAC's at low concentrations.  Concurrently, we are developing the enabling technology to incorporate these sensors into a small, portable system to detect land mines.

Figure 3 Average fluorescence change vs. time for a set of eighteen micrometer-sized sensors.  Four different analytes each produces a different signature when saturated vapor is delivered in a pulsatile fashion. Published in  Proceedings of SPIE:  Detection and Remediation Technologies for Mines and Minelike Targets III   13-17 April 1998, Vol. 3392, Part 1 of 2, p.426.

Also we are attempting to detect explosives and explosive-like materials using a high-density array of thousands of micrometer sized sensors.  This new array fabrication technology developed in our laboratory (see "Random Arrays ") allows thousands of micro-sensors to be simultaneously monitored upon a vapor 'puff.'  Using a sensor population of this magnitude benefits a field system where false-positives might affect performance.

*For sensor materials and fabrication see artificial nose .

Collaborative Efforts

            (a)  University South Carolina (USC), Prof. Michael L. Myrick
            (b)  Lawrence Livermore National Laboratories (LLNL), Dr. Fred Milanovich & Steve Brown
            (c)  Massachusetts Institute of Technology (MIT), Prof. Timothy M. Swager

Collaborative efforts with MIT and their sensory material (TNT-quenching polymers as seen in Journal of the American Chemical Society , (1998) 120: 5321-5322) has led to a broader sensor design incorporating more selectivity for nitroaromatic compounds.
    Collaborators from Lawrence Livermore National Laboratories and University of South Carolina have worked with Tufts to construct a first generation portable detection system for landmines which mimics our current bench-top imaging and detection system.  Take a look at these photos: of the completed portable system (Figure 4), the fiber-bundle sensor end (Figure 5), and one outside at Tufts (Figure 6). Special thanks to Steve Brown, Fred Milanovich, Brian Bodtker, Jim Richards (all of LLNL) and Michael Myrick of USC for system design, development, and software and hardware interface.

Figure 4 The complete buggy detection system.  Laptop controlled hardware components include a CCD camera, a complicated optical block with two LED excitation sources, an air-flow pump system, and sensor housing with built in fan for vapor generation.

Figure 5 The seven-sensor portion of the portable system: a six-around-one fiber bundle shown with green LED excitation.  The outer diameter of the fiber-bundle is smaller than that of a dime.

Figure 6  Our system outside at Tufts University prior to field deployment Sept.1998. (L to R):  Steve Brown of LLNL, Keith Albert of Tufts, and M. L. Myrick of USC.

4. Materials Chemistry

Preparation of Core-Shell Composite Materials-Colliodal Assembly

Colloidal assembly is a process by which particles ranging in size from nanometers to micrometers are organized into structures by mixing two or more particle types. Assembly is controlled by either specific or non-specific interactions between particles. Examples include chemical bonding, biological interactions, electrostatic interactions, capillary action and physical adsorption. The assembly process is performed such that smaller particles assemble around larger ones. We are able to assemble polymer nanoparticles (50-200 nm diameter) onto silica particles (3-5 m m diameter) using specific chemical interactions (i.e. amine-aldehyde) or biospecific interactions (avidin-biotin).

Figure 1 Self-assembled nanospheres

Figure 2  200 nm diameter PS nanospheres assembled on 5 m m Silica Microspheres

Annealing the assembled composites at temperatures above the glass transition (T g ) of the polymer nanospheres allows polymer to flow and uniformly coat the microsphere surfaces. Polystyrene and poly(methyl methacrylate) nanospheres were used to produce such materials. These materials may be applied to the development of novel sensing materials for use in the artificial nose or in the development of novel stationary phases for chromatography applications.

Figure 3 100 nm PS assembled on 3.1 uM silica (left). Ps metling at 170 - 180° C

Core-Shell Composite Materials-Surface Confined Living Radical Polymerization

We have made core-shell composites consisting of an inorganic core (silica) and polymer shell (poly (benzyl methacrylate), polystyrene or poly (methylmethacrylate) etc.) by surface immobilization of an initiator for living radical polymerization, followed by polymerization of various selected monomers in the presence of these “macro-initiators”. Core-shell materials prepared with this technique have also been used as sensors in the ‘artificial nose'.

Figure 4 Coating of Silica Beads with Poly(benzyl methacrylate) by LRP

Production of Hollow Polymeric Microspheres by Surface Confined Living Radical Polymerization on Silica Templates

We are able to synthesize uniform hollow polymer microspheres by coating silica microsphere templates with poly(benzyl methacrylate) using surface initiated controlled/living radical polymerization and subsequently removing the core by chemical etching. Shell thickness was controlled by varying the polymerization time. Hollow microspheres have many applications including: encapsulation of drugs for drug delivery or as pigments in the development of new coatings and paints.

Figure 5 Schematic representation of the production of hollow polymeric microspheres.

Figure 6 Hollow MBzMA microspheres (left). Hollow PMMA microspheres (right).

5. Near-Field Arrays

Arrays of near-field tips have been fabricated from high-density fiber optic bundles. Bundles containing approximately 30,000 individual fibers were chemically etched to form an array of silicon tips. To create apices for each tip, a metal layer was subsequently sputtered over the entire surface to prevent light escaping from the sides. An insulating polymer layer was then electrochemically deposited followed by heat curing to expose gold at each tip. The gold was then chemically removed.

6. Lectin Screening Array

This project employs lectins (carbohydrate binding proteins) as the recognition element. Different carbohydrate structures, synthesized in the laboratory of Prof. Peter Seeberger (MIT), were coupled to beads with bovine serum albumin as a linker. The beads are internally encoded with a fluorescent dye for positional registration. A different dye concentration encodes for each carbohydrate-sensor. The modified beads are loaded into the etched wells of a fiber optic imaging bundle. Two different arrays were tested: on the first array, two carbohydrates were screened for their affinity to Concanavalin A; on another array, five different carbohydrate structures were screened for binding to cyanovirin, an anti-HIV agent. The results demonstrated both a very high selectivity and sensitivity of the sensors with low nonspecific adsorption.