Selective oxide sensors as non-invasive disease monitors

It has long been known that exhaled breath contains numerous chemical compounds that may signal metabolic changes in human and animal bodies, including diseases such as kidney failure and lung cancer. Hippocrates of Cos coined medical terms (such as fetor hepaticus) that relate the odor of exhaled breath to a disease or medical condition (see Figure 1). In the early 1980s, Linus Pauling advocated the concept of orthomolecular medicine—a theory that relates the concentration of human fluids to state of health—and started analyzing the chemical constituents of exhaled air using gas chromatography. Since then, it is known what it takes to ‘capture one's breath’: highly specific and sensitive gas detectors. Our work suggests that nanosensor devices made from metallic oxides are the most-promising, non-invasive, and personalized medical tools to this effect.

Figure 1. Timeline of breath analysis from antiquity to date. GC: Gas chromatography.

To date, what has limited the widespread use of breath analysis as a non-invasive methodology for medical diagnostics is the lack of appropriate sensor technology. Breath contains as many as 300 different gaseous species, and robust technologies would enable detection and discrimination of particular ‘disease-signaling’ gases from the complex odors of the breath. Currently, there are selective gas detectors that measure trace chemical concentrations in breath samples, but they are either too expensive,¹ bulky,² and/or require large samples of collected exhaled breath.³

Nanomedicine is a novel interdisciplinary field that comprises science, engineering, and medicine. It provides tools for the prevention and early detection of diseases, as well as effective therapeutic means for their treatment.⁴ Resistive chemosensor nanotechnology can provide novel and affordable breath analysis tools with the required gas selectivity. The principle of operation of gas-specific chemical sensors, based on nanostructured semiconducting oxides, involves selective interactions between targeted gas molecules and appropriate metal oxides.⁵ These interactions are determined by the arrangement of the oxide atoms on the surfaces exposed to the gas. This determines whether the gas-oxide interactions will be adsorption-based,^5,6 reaction-based,⁷ or due to ferroelectric poling.⁸ (These are the most prominent mechanisms observed to date.)

It is interesting to note that binary metal oxides are polymorphic materials. Depending on the pressure and temperature, compounds of the same composition may exist in different crystalline forms, each behaving as a distinct material with unique physical and chemical properties. Figure 2 shows the ε, γ, δ, β and α crystalline phases—polymorphs—of tungsten trioxide (WO₃) and their corresponding phase-transition temperatures. (One may refer to a metal oxide gas sensor by its crystalline structure rather than its composition.)

Figure 2. Phase transitions for bulk tungsten trioxide (WO₃) with the ε, δ, γ, βand α crystalline phases and their corresponding phase transition temperatures.

Using nanoscale processing, we stabilized polymorphs (that are not expected to exist at standard ambient conditions) at room temperature. Such is the case of the ε-phase of WO₃, a rare compound with ferroelectric nature that has shown specificity to polar molecules and which forms the basis of an acetone breath analyzer developed by our group for diabetes monitoring.⁹

To appreciate the importance of polymorph selection in gas sensing, we analyzed a nanosensor, based on the γ-phase of WO₃, that is specifically sensitive to nitrogen oxide (NO).¹⁰ Figure 3 shows that this monoclinic oxide phase is very selective of NO, while showing no response to reducing gases such as acetone, isoprene, methanol, and ethanol. This makes it a potential breath analyzer for asthma detection.

Figure 3. Sensitivity (Rg/Ra) of the γ-phase of tungsten trioxide (WO₃) to oxidizing gases such as nitrogen oxide (NO),¹⁰ specifically the response of WO₃to concentrations of 10ppm of NO, acetone and isoprene, and 50ppm of ethanol, methanol and carbon monoxide (CO).

Since oxide crystallography is the key to controlling gas affinity, single crystal nanowires are important to gas specificity. Their high aspect ratio configurations (nanoscale diameter and micro/macroscale length) offer high surface area for enhanced gas adsorption, as well as a high sensitivity to trace gas concentrations. Moreover, structurally well-defined and non-defective surfaces allow for a stable sensing response. We used electrospinning to produce one-dimensional oxide nanostructures as the next-generation nanosensing elements of inherent selectivity.^11,12

Non-invasive diagnostics, such as detection based on breath analysis, require affordable sensors that are highly specific to a given chemical. Our oxide nanosensors exhibit gas specificity and are based on a crystallo-chemical approach that matches classes of gases to specific crystalline phases of the WO₃ (rather than to specific oxide compositions). We have presented two different polymorphs of WO₃ with different gas specificities. The ε-phase of WO₃ was selective for a polar molecule such as acetone, while the γ-phase was selective for an oxidizing gas such as NO. Single crystal nanowires offer the promise of more sensitive and highly stable nanosensors. We have tested breath-analyzer prototypes in pre-clinical trials, and our next step is to try out this novel nanotechnology in human subjects.

This work has been primarily supported by the National Science Foundation and has been a collective effort of the author's research group. Key people carrying out the research described in this article were: Arun Prasad, Krithika Kalyanasundaram, and Lisheng Wang.