This competition and all ideas within it are archived
Check out the new way Marblar is realizing the promise of science.
The competition has now ended, but the conversation continues. Browse past entries, comment on your favourites, and join the discussion below. If you have any ideas for a new entry, please leave it in the discussion below.
Breath analysis
23 Sep '12
There's been some interest in using spectroscopic techniques to analyse breath. Molecules found in breath are markers for disease. The high stability and portability of your equipment is well suited to this field.

Ammonia is a sign of liver and kidney problems as described in Breath Ammonia Analysis: Clinical Application and Measurement by Hibbard and Killard.

Various markers are known for lung diseases as detailed in Biomarkers of some pulmonary diseases in exhaled breath by Sergei A. Kharitonov and Peter Barnes

The question that remains is that of sensitivity. Does MicroFTS have what it takes?
[1] Dweik, R. A., Amann, A., "Exhaled breath analysis: the new frontier in medical testing", Journal of Breath Research 2 (2008)

[2] Risby, T. H., Solga, S. F., "Current status of clinical breath analysis", Applied Physics B: Lasers and Optics 85, 421- 26 (2006)

[3] Risby, T., Tittel, F. K., "Current Status of Mid-Infrared Quantum and Interband Cascade Lasers for Clinical Breath Analysis", Optical Engineering 49, 000000-1 (2010)

[4] Phillips, M., "Breath tests in medicine", Scientific American 267, 74-79 (1992)
Gabriel Mecklenburg on Sep 23, 2012
Hey David,

That does indeed seem like a very relevant application. Hugh, it would be great to get your input on this. If you look at the Optical Engineering reference, the key requirement seems to be being able to detect certain organic compounds (such as nitric oxide) in breath at ppb accuracy.

Hugh, would this be feasible? If so, what kind of light source and detector would you need? The article discusses several high-quality light sources, but not detectors. Admittedly, this is not my area od expertise.

This contribution has won an assist marble
Mehmet Fidanboylu on Sep 23, 2012
Hello Dave,

This is actually a really cool idea. What about using it in a medical setting, like testing for H. pylori infection? I know the current gold-standard for testing this is a breath test now, but it's expensive, and you have to send it off to a lab for analysis that takes a week or two before getting your result.

If MicroFTS could pick up some metabolites that are specific to H. pylori (if they exist) then maybe we could have an instantaneous test that GP's could keep in their surgeries?

This contribution has won an assist marble
David Marshall on Sep 24, 2012
Thanks for input so far. Sorry for the formatting of the entry, I think submission from iOS gets a bit mangled. Is there a way to edit entries?
Gabriel Mecklenburg on Sep 24, 2012
Only if you're a moderator I'm afraid. Could you send it to me at and I'll change it to the correct formatting?
Hugh Mortimer (Inventor) on Sep 24, 2012
Hi David,

This is a great application area, and one that the microFTS could address give the correct breath sampling accessory. We've designed and tested a bench version of the instrument to measure SOx, NOx, ammonia, (NH3) and methane (CH4), in air and we are currently at the ppm level given our basic sampling setup. We would really like to develop a more sensitive version based on better detectors and sampling methods (gas cells etc.) that we believe would help. It would be really interesting to explore this idea further. What sensitivities and key gases do you think we would need to be able to measure for this to be a useful diagnostic tool?

David Marshall on Sep 26, 2012
Hi Hugh,

Thanks for sharing your invention with us, your comments and your question.

I've had a quick look at the literature to find some ppm values.

With regard to ammonia and liver diseases, I found the following from Shimamoto[1]:

"Breath ammonia levels were significantly higher in cirrhotic patients (0.745 ppm) than in controls (0.278 ppm), and higher in cirrhotic patients with hyperammonemia (0.997 ppm) than in those without (0.558 ppm)"

Looking at the case of acetone and diabetes, Deng et al. [2] state "acetone in diabetic breath was found to be higher than 1.71 ppmv, while its concentration in normal breath was lower than 0.76 ppmv".

With these values in mind, it looks like that could be stretching your current set up.

Looking back at Mem's suggestion, it's true that a urea breath test is very common test for H. pylori. Patients are given 13C labelled urea. If H. pylori is present, the urea is broken down by urease into CO2 and NH3 and the 13CO2 is detected. Could the MicroFTS detect the difference between 12CO2 and 13CO2?

Heading back to ammonia briefly, another test for H. pylori is to give the patient normal (12C) urea and look at ammonia levels. Kearney et al. [3] state that after a dose of 300 mg, levels of ammonia in breath are 0.04 ppm (+/- 0.09) for H. pylori negative people and 0.49 (+/- 0.24) for positive. Again, this could be a little low for the current set up.

This isn't an exhaustive list but just a few figures to get the ball rolling.

[1] Breath and blood ammonia in liver cirrhosis.Hepatogastroenterology. 2000 47(32):443-5

[2] 10.1016/j.jchromb.2004.08.013

[3] Breath ammonia measurement in Helicobacter pylori infection. Dig Dis Sci. 2002 47(11):2523-30
Gabriel Mecklenburg on Sep 27, 2012
Hi Dave,

I'll leave commenting on the specific ppm values to Hugh, but I can say that FTIR wouldn't be able to distinguish 13CO2. Isotope substitution only has a very minor effect on IR absorption related to bond vibrations (think of a slightly heavier weight connected to the same spring) which you would be very unlikely to be detected at these low concentration.

Simon Bayly on Oct 10, 2012
The challenge is not just sensitivity, it is also selectivity. MicroFTS needs to be able to detect ppb analyte levels not just in air, but n breath, which is a fairly complex mixture of stuff

There's a timely article on this in the Wall Street Journal:

There is already an Oxford spin-out working on breath analysis for diabetes detection
Oxford Medical Diagnostics are using Cavity Enhanced Absorption spectroscopy and Plasma Emission spectroscopy.
Andrius Jonas Dagilis on Oct 12, 2012
Hello all,

Going along very much the same lines, this technology might be useful in Tuberculosis diagnosis. Currently, diagnosing an active (that is, pathogenic) form of tuberculosis is usually done by culturing the bacteria. This requires a large amount of sputum, if it is produced, or more invasive samples when it is not. Furthermore, the resulting cultures need to be properly identified, and this is a fairly complex process. In effect, full diagnosis is a very long and arduous process.

There are alternative methods of identification, however. One that had caught my eye is identification of the bacilli through Raman spectroscopy ( While I understand that your machine is not a raman spectrometer, and my knowledge of chemical instruments is a (tad) simplified, as far as I can tell, all you need to do is strap a laser to your machine to make it a raman spectrometer.

The main reason why your machine may be far better suited for this than many others is the far lower price. Regions with the highest Tuberculosis incidence tend to not be able to afford the faster, more accurate TB diagnosis tools. Thus, your machine, coupled with a hopefully cheap laser, might be able to fill a niche with high demand.

I think this application is fairly straightforward, but as another (last) defense, I'll cite the authors from the study posted: "Important features of these methods are the relative ease by which measurements can be performed, the limited amount of sample handling involved, the small amounts of biomass required and the high degree of reproducibility. Fourier transform infrared spectroscopy proved to be a convenient approach for classifying NTM at the species level."

Andrius J. Dagilis
This contribution has won an assist marble
Gabriel Mecklenburg on Oct 12, 2012
It appears I misspoke in my last comment - goes to show how reliable our assumptions actually are. Turns out that the peaks for the two isotopes are actually 100cm-1 apart [1] - and can be analysed using FTIR. If the microFTS could handle the relevant concentrations, the H pyllori test could be a great application!

Regarding Andrius' comment: Hugh, how easy would it be to re-jig the microFTS into a Raman instrument? From my own lab experience I know Raman systems as pretty huge (since they usually require powerful lasers), but apparently there are hand-held ones out there already [2].

Hugh Mortimer (Inventor) on Oct 22, 2012
This is a very exciting feed, so thanks to everyone contributing. I think that both Davids and Simon are making very valid comments. The gas mixtures associated with breath is especially complex considering the increased water vapour associated with exhalation. There are methods that can address the interference between consecutive features, from physically filtering the gas sample through to digitally post processing the data. But what I have taken from Davids post is that to be able to measure upshot is that to be a realistic monitoring device for Cirrhotic the microFTS would need to be able to measure NH3 to a level of 0.05ppm, for diabetes the instrument would need to be able to measure acetone to 0.1ppm and H. pylori the instrument would need to be able to measure NH3 to 0.01ppm. Do you think this is a fair summary?
Simon Bayly on Oct 22, 2012
For diabetes/acetone 0.1ppm may be slight overkill. According to Oxford Medical Diagnostics 'below' 1ppm accuracy is required. Acetone is around 0.5 ppm in healthy people, but for untreated/undiagnosed diabetics the level can rise to 5 ppm.
This contribution has won an assist marble
David Marshall on Oct 22, 2012
Hi Hugh, just to clarify, are you talking about minimum detectable levels or precision of measurement required?
Hugh Mortimer (Inventor) on Oct 29, 2012
Hi David,

I think that knowing the minimum detectable level should lead to a required precision of the measurement technique. It would be great to know both the detection limit and precision required to make a meaningful measurement. I'm assuming that one would want to be able to perform a quantitative analysis of the composition of the breath to an accuracy that could determine the level of ammonia above the expected.
Hwang Jaeyun on Nov 02, 2012
you would need a database to compare with. I'm sure the lay people would not know the proper composition of normal breath, plus their diet might also influence results.
Andrea Mica on Nov 05, 2012
TB doesnt just happen in poor countries. Bovine TB could do with rapid diagnosis and of course there is the re-rise of TB in many countries. There's an Oxford spin-out that knows everything about TB.
This contribution has won an assist marble
Hugh Mortimer (Inventor) on Nov 05, 2012
Hi Hwang, The reference database is actually one of the most important parts of determining accurate gas concentrations from the spectra measured. The quantitative analysis of the infrared spectra (chemometrics) uses principle component analysis (or something equivalent) to fit known sample spectra measured spectra. These libraries can be bought, but the results are often better if the reference spectra are taken with the spectrometer that you are using. This is something that we need to look into further, especially in the case where sensitivity and selectivity are critical

David Marshall on Nov 12, 2012
Sorry for the delay, but here's a table of concentrations.

Ammonia and cirrhosis [1]
Cirrhotic - 0.745 ppm
Cirrhotic with hyperammonemia - 0.997 ppm
Control - 0.278 ppm

Acetone and diabetes [2]
Diabetic - higher than 1.71 ppmv
Normal - lower than 0.76 ppmv

Ammonia and H. pylori [3]
After 300mg urea dose
Positive 0.49 ppm (+/- 0.24)
Negative - 0.04 ppm (+/- 0.09)

[1] Breath and blood ammonia in liver cirrhosis.Hepatogastroenterology. 2000 47(32):443-5
[2] Determination of acetone in human breath by gas chromatography–mass spectrometry and solid-phase microextraction with on-fiber derivatization 10.1016/j.jchromb.2004.08.013
[3] Breath ammonia measurement in Helicobacter pylori infection. Dig Dis Sci. 2002 47(11):2523-30
David Marshall on Nov 14, 2012
I missed Hwang's entry! That is great point about building a database. If we're doing 1000 genomes, let's do 1000 breaths :)