Goals
The goal of this research is to carry out the initial studies which will lead to the development of a small, easily fieldable instrument to measure organic vapors in air from sources, in industrial atmospheres, or in the ambient air. The SAMP sensors, in theory, offer high sensitivity and selectivity, small size, and it is simple to change their selectivity profile by employing different dyes. This makes them ideal for assembling into an array. There are well developed techniques for 'training' an array of differently sensitive sensors by exposing the system to various known sample concentrations, so that unknown samples can be characterized. An 'ozone precursor index' is an example of a possible output of this sensor array system.
The sensitivity of the sensor's response varies with the overlap of the sorbed gases' IR spectra with that of the photosensitizing dye. Initial versions of the device have been made with rhodamine dyes, which contain aromatic rings in their structures. Sensors coated with rhodamine are sensitive to such compounds as benzene, toluene, xylenes and other substituted benzenes (BTEX group). In addition, the molecular vibrations of olefinic compounds may also resonate with rhodamine, and will certainly do so with other suitable dyes. Initial work will focus on the BTEX compounds and possibly olefins, as these seem to be the most promising and are also of great environmental interest, as reactive ozone precursors. Eventually, the development of a small, rugged, portable device which can give an overall index of a particular air quality parameter is the goal of this research.
Background
Policy making and source identification for airborne organic vapors such as ozone precursors depend heavily on analytical techniques. However, analysis of air samples for organic vapors is a exacting and expensive process. Methods currently used include trapping organics on packed traps, desorbing these either thermally or with a solvent, and gas chromatographic analysis, often with mass spectrometric detection. However, methods using arrays of simple sensors, each sensitive to a somewhat different range of compounds, and each giving a different output signal for the particular mix of compounds to which it is exposed, holds promise for a simpler approach. The general technique is currently in commercial use for what are called 'electronic noses.' In this project, we are developing a novel class of semiconductor based, solid state sensors, which show high sensitivity for organic vapors and varied selectivity. These can be made with a variety of surface coatings for use in a sensing array.
Solid state semiconductor sensors, seem to offer a nearly ideal method for monitoring. They are chemically selective, reversible, rapid in response, low in noise, sensitive, not easily poisoned, small in size and low in cost. Selectivity, however, is a problem if these sensors are to be used for monitoring volatile organic compounds (VOC) in complex environmental mixtures.
Commonly used solid state sensors are based on adsorption-induced changes in the characteristics of the surface, such as the dark conductivity (sd), photoconductivity (sp), or surface potential (Y). These characteristics are governed by the surface electronic states (SES), which are determined primarily by the structure of terminal chemisorbed molecular groups. Adsorption of atoms or molecules on the semiconductor surface changes the distribution of the existing SES and creates new SES. With van der Waals adsorption, the energy parameters of the existing SES change to only an insignificant extent. On disordered surfaces, the energy spectrum of SES is quasi-continuous, and, while the sorption of molecules may be fairly selective, this selectivity in sorption is not reflected in the sensor response. The variations in sd, sp, and Y which occur due to absorption are believed to be due to a change in the occupation of the SES. This purely charge-controlled approach cannot be relied upon to solve the selectivity problem. In this project the traditional semiconductor sensor techniques will be combined with the use of fine vibronic effects caused by adsorbed dye molecules interacting with the sorbed analyte, to produce novel semiconductor sensors capable of monitoring vapors with much higher selectivity.
Dye molecules cause an increase in the photoconductivity of a semiconductor upon which they are adsorbed. When an analyte molecule is adsorbed with the dye, energy can be transferred within the dye molecule from the excited electronic state, into vibrational modes. If the vibrational states of the dye and the sorbed molecules are in resonance, some energy will be transferred to the analyte, rather than to the semiconductor, changing the amount of energy available for the production of photosensitized conductivity. This changed conductivity is measured and related to the amount of sorbed analyte.
Our (VZ) experiments with semiconductor surface films made of Ge, ZnO, or CdS have shown that dyes coated on the surface can be photoexcited and cause an increase in conductivity. These energy transfers cause a large increase in the photosensitized conductivity of the semiconductor film. The photoexcited dyes will also undergo resonant vibrational energy transfers with co-adsorbed analyte molecules which have corresponding vibrational modes. The enhanced photoconductivity effect is reduced when energy is drained into co-adsorbed analyte molecules. The selectivity therefore depends on the overlap of the rich vibrational spectra between the dye and the analyte molecule. The Semiconductor-based Adsorption-Modified Photosensitization (SAMP) sensors, which will be developed in this project have already been shown to be sensitive to ppm level gases, and to be quite selective. It has been shown, for instance, that the sensor can distinguish between H2O and D2O, and between naphthalene and its deuterated analog. Current chemiresistor sensors cannot distinguish between such similar compounds.
Method of Approach and Facilities
Photo-induced singlet-singlet S0ÆS1 transitions in adsorbed dye molecules are known to initiate electronic transitions in solids. Combined electrophysical measurements of the surface charge and spectral measurements of molecular luminescence have been used to study the fine vibronic effects in the semiconductor-dielectric-dye structures. The semiconductor electronic subsystem has been found to be sensitive to any changes in the adsorbed phase due to strong competition of two energy transfer channels: 1) inside the molecular phase (channel M) and 2) into semiconductor substrate (channel S). A strong fluorescence quenching occurs and the DT emptying rate decreases, when molecules with overlapping luminescence and absorption bands are present on the surface. When there is resonance between electron transitions in molecular phase, the energy of an excited electron from a donor molecule is transferred to vibrational modes of this molecule and then to adjacent acceptor molecules, if their vibrational spectra overlap.
In Dr. Zaitsev's laboratories, initial experiments have shown that the devices, photosensitized with rhodamine B or rhodamine 6G, were sensitive to naphthalene, for instance, but not to deuterated naphthalene, whose vibrational absorption bands are shifted. This demonstrates that the decrease in photoconductivity with exposure to the vapor is a function of a very specific interaction between the sorbed vapor and the dye, rather than a gross function of surface sorption. Therefore, since the response depends on a specific interaction, there is promise that the sensitivity to different compounds can be controlled and varied by the use of different dyes. The sensors used this project consist of a thin semiconductor film on a dielectric substrate. Micro-quantities of organic dye molecules are adsorbed on the film. The film has metal contacts to measure conductivity. The light used to photoexcite the dye molecules must be limited to the wavelength band in which the dye absorbs, to avoid excess direct stimulation of photoconductivity in the semiconductor. Broadband interference filters are being tested now, but other kinds of filters may be needed.
The sensors studied so far have been made in the MSU laboratories, but the basic semiconductor structures are essentially the same as commercial solid state devices used in automatic lighting. Some commercially available, solid state devices are now being tested. If these can be dye coated and used in this project, it will reduce the cost and the time required for commercialization of the finished sensor.
The experiments under real atmospheric conditions are being performed at New Jersey Institute of Technology. After toluene, the response to other saturated, unsaturated hydrocarbons and aromatic hydrocarbon vapors will be studied. Interferences from water, carbon dioxide and other gases present at high concentration in the atmosphere will be examined. These are expected to be minimal, since the photosensitizing dye molecule will be chosen so that its IR spectra does not overlap significantly with the interferents. Once the sensor technology has been developed and a sufficient number of different sensors have been tested in single gases and mixtures of gases, initial studies on the sensor array will be undertaken.
This research is a collaborative project between New Jersey Institute of Technology (NJIT) and Moscow State University (MSU). The co-principal investigators, Dr. Zaitsev at MSU and Dr. Kebbekus at NJIT have different areas of expertise which cover those needed for such an interdisciplinary project. Dr. Zaitsev, is producing the devices, and studying the selectivity variations brought about by use of different dyes in the surface layers. The laboratory at MSU is well equipped to do the proposed sensor development. The laboratory has the equipment necessary for the initial production and testing of the sensors. This includes high vacuum systems for in vacuo experiments; a piezoresonance balance apparatus which is capable of measuring the amount of sorbed dye at levels as low as 1012 molecules per cm2; electronic equipment for measuring the surface charge and conductivity; and a 1 kW high pressure xenon lamp with monochromator to illuminate the samples.
Dr. Kebbekus has experience in the field of gas analysis, especially the determination of organics at the ppb level in ambient atmospheres. She has developed the system to test the devices. The laboratory at NJIT is equipped with necessary apparatus for gas analysis and calibration of such systems.
Summary of Progress and Accomplishments
Work carried out during the first few months of the currently funded project has shown the feasibility of using the sensor at atmospheric pressures in air.
The solid-state sensors are being prepared by Dr. Zaitsev at Moscow State University. Thin films of CdS were deposited onto 2 mm thick glass substrate from a water solution of Cd salt and thiourea. The specific surface area of the film, is 10 to 20 m2/kg. The thickness of the films was chosen to be 0.5 mm.
The conductivity of the films in the darkness is about 10-9 to 10-7 ohm-1cm-1. The films show photo-induced conductivity of 10-2 to 102 ohm-1cm-1 when illuminated with ultraviolet or visible light. The photoelectric threshold is shifted towards higher energies for the films in comparison with single crystals of CdS. This is due to the structure of the film. Electron microscopy shows that the film consists of the small grains of CdS. The grain size varies due to the deposition conditions. The initial films showed slow (103 to 106 s) kinetics of photoconductivity relaxation after switching off the light. To stabilize the films and to eliminate the slow photoconductivity kinetics we have treated the films at the temperature of 773 K under the atmospheric conditions for 30 minutes. The long photoconductivity decay disappeared after this treatment. The grain size in the treated films is about 100 A, measured by electron microscopy. The same result is given by the number of "blue" shift of the film photoelectric threshold, which is due to the quantum dimensional effect in the CdS grains.
An organic dye rhodamine G (RhG) is deposited on the CdS films from an ethanol solution over a period of 15 minutes. The surface concentration of the RhG is determined by the solution molar concentration, which was chosen to be 10-3 M/l to provide a surface concentration of about 2x1013 molecules/cm2. This surface concentration was estimated by the microbalance method and gives the best results for the spectral sensitization of the CdS film with the dye molecules. When the film with deposited RhG molecules is illuminated with light, it shows a new spectrum of photoconductivity. It consists of an "old" CdS photoconductivity band and a new additional wide peak of photo conductance with a maximum at about 530 nm. This new peak lies in the region where the initial film was not light sensitive and it is due to the light absorption by the RhG molecules and energy transfer from the molecules to the CdS film. To trigger the photosensitization, the specimens are illuminated with monochromatic light. In these experiments, a xenon lamp and monochromator are being used to obtain the proper wavelength.
The atmospheric pressure gas generation system was assembled. It consists of a 250 ml glass chamber, which is flushed gas (air or nitrogen). The toluene is added to the diluent stream by diffusion through a short capillary tube. With a flow of 13 ml/min, the setup gives a calculated toluene concentration of approximately 4.4 ppm. The concentration can be varied by changing the diameter or length of the capillary tube, and the diluent flow. A gas chromatograph will be attached to the exit of the system to monitor the concentration. Since the chamber acts as an exponential dilution flask, the concentration rises from zero (pure diluent) to 4.4 ppm, over a period of about 75 min. If there is efficient mixing within the chamber, the concentration rise follows an exponential curve, approaching 4.4 ppm asymptotically. At this point, we cannot follow this change experimentally, and so are not sure if the mixing is complete. The apparatus was assembled under considerable time pressure, in order to obtain some results before this proposal was due, and many compromises were made to expedite obtaining the first readings. Several improvements to the system are now being made.
The conductivity is measured using a current source and a sensitive analog electrometer. Two wide bandwidth interference filters centered around 450 and 500 nm respectively and a halogen lamp have provided the light source. The photoresistance was constant in either air or nitrogen. When readings are taken as the gas flowing into the flask is changed from pure air to 4.4 ppm toluene in air, a decrease in photoresistance is shown as the toluene concentration increases. The sensor appears to respond nearly linearly as the concentration increases. However, when the calculated concentration begins to level off, at around 4 ppm, the sensor continues to show a changing signal. There are two possible explanations for this. If the mixing in the flask is not efficient, the concentration may still actually be changing at this point in the run, and the sensor may be giving correct information. Or, the resistance may be decreasing because its response is slower than the concentration change. However, in either case, there is a measurable resistance change at the first point taken, in a mixture containing about 1.3 ppm toluene, which indicates that the sensor is working. The gas dilution system is being refined and upgraded and new sensors have been prepared and shipped to NJIT.
Status of Original Plan
Original revised scope of work and current state of completion
Future Plans
Handling of QA/QC
The primary purpose is to develop a new class of sensors, and apply them to measurements in air. To validate their response, the results obtained from the sensors will be compared with current methods, primarily gas chromatography, using gravimetrically prepared or purchased gas standards. The experimental measurements made using the new system will be compared with gas chromatographic methods, with which the researchers at NJIT have had extensive experience. The usual blanks and synthetic samples will be used to validate the method, define the detection limits and assess the overall stability of the system, and its response to both analytes and possible interferences. The data will be collected in the laboratory at NJIT, retained in the laboratory notebooks of the graduate student working on the project, and will eventually be published in the peer reviewed literature. Copies of laboratory data from the group at MSU will also be filed at NJIT to make them more readily available. Since this is a novel application of these devices, patents are also possible.
Key Personnel
Research Assistant:
Ashish Agrawal
[Source and Control Projects] [Transport and Transformation Projects] [Monitoring and Source Attribution Project]