Article pubs.acs.org/ac
Fast and Highly Sensitive Fiber-Enhanced Raman Spectroscopic Monitoring of Molecular H2 and CH4 for Point-of-Care Diagnosis of Malabsorption Disorders in Exhaled Human Breath Stefan Hanf,†,‡ Timea Bögözi,† Robert Keiner,† Torsten Frosch,*,†,§ and Jürgen Popp†,§,∇ †
Leibniz Institute of Photonic Technology, Jena 07745, Germany Max Planck Institute for Biogeochemistry, Jena 07745, Germany § Institute for Physical Chemistry, Friedrich Schiller University, Jena 07745, Germany ∇ Abbe Center of Photonics, Friedrich Schiller University, Jena 07745, Germany ‡
S Supporting Information *
ABSTRACT: Breath gas analysis is a novel powerful technique for noninvasive, early-stage diagnosis of metabolic disorders or diseases. Molecular hydrogen and methane are biomarkers for colonic fermentation, because of malabsorption of oligosaccharides (e.g., lactose or fructose) and for small intestinal bacterial overgrowth. Recently, the presence of these gases in exhaled breath was also correlated with obesity. Here, we report on the highly selective and sensitive detection of molecular hydrogen and methane within a complex gas mixture (consisting of H2, CH4, N2, O2, and CO2) by means of fiber-enhanced Raman spectroscopy (FERS). An elaborate FERS setup with a microstructured hollow core photonic crystal fiber (HCPCF) provided a highly improved analytical sensitivity. The simultaneous monitoring of H2 with all other gases was achieved by a combination of rotational (H2) and vibrational (other gases) Raman spectroscopy within the limited spectral transmission range of the HCPCF. The HCPCF was combined with an adjustable image-plane aperture pinhole, in order to separate the H2 rotational Raman bands from the silica background signal and improve the sensitivity down to a limit of detection (LOD) of 4.7 ppm (for only 26 fmol H2). The ability to monitor the levels of H2 and CH4 in a positive hydrogen breath test (HBT) was demonstrated. The FERS sensor possesses a high dynamic range (∼5 orders of magnitude) with a fast response time of few seconds and provides great potential for miniaturization. We foresee that this technique will pave the way for fast, noninvasive, and painless point-of-care diagnosis of metabolic diseases in exhaled human breath.
T
lactose intolerance.7 Consequently H2 and CH4 can serve as combined biomarkers for colonic fermentation of undigested or unabsorbed oligosaccharides. Hence, innovative multigas sensors for H2 and CH4 with sensitivities in the range of 10− 100 ppm are required for the monitoring of HBTs (for example, the lactose breath test).8−10 Conventionally available hydrogen sensors can be classified as (electro-)chemical,11,12 electronic,13−15 mechanical,16,17 acoustic,18−20 and (chemo-)optical21−24 sensors. Although detection limits in the low ppm region can be achieved, most of these sensor types require an oxygen atmosphere for proper functionality.25 Palladium and platinum sensors susceptibly suffer from mechanical damage upon hydrogen exposure and hydrogen absorption into the bulk material.26 Electrochemical sensors are affected by aging and sensor surface poisoning and show cross-sensitivities.27 Acoustic sensors lack selectivity and
he analysis of biomarkers in exhaled breath is an emerging research field for early-stage diagnosis of metabolic disorders or diseases.1 Human breath is a matrix of major gas components (e.g., N2, O2, and CO2) and also contains various minor gaseous substances that can be indicators for disorders or diseases.1 Hydrogen (H2) in concentrations of several tens of ppm within breath is only generated by anaerobic metabolism of colon bacteria Escherichia coli during metabolic disorders, including lactose intolerance,2 fructose malabsorption,3 and bacterial overgrowth syndrome.4,5 An increase in H2 concentration by 20 ppm above basal within 90 min after oligosaccharide (lactose) ingestion is considered to be a positive hydrogen breath tests (HBT), 5 because of a malabsorption problem regarding the tested oligosaccharide. An increase of more than 12 ppm of H2 over basal within 90 min is also an indicator for bacterial overgrowth in the small intestine.6 For some patients, also methane (CH4) can be detected, which is produced by another large bowel bacterium (Methanobrevibacter smithii) from four molecules of H2. A cutoff value of 20 ppm of CH4 above basal is an indicator for © 2014 American Chemical Society
Received: September 14, 2014 Accepted: December 14, 2014 Published: December 29, 2014 982
DOI: 10.1021/ac503450y Anal. Chem. 2015, 87, 982−988
Article
Analytical Chemistry face a temperature-dependent response time.28 Further alternative detection methods such as gas chromatography (GC) in combination with a thermal conductivity detector (TCD) are bulky and expensive.29 Optical spectroscopic detection techniques promise high selectivity, because of spectral discrimination. However, homonuclear diatomic gas molecules, such as H2, N2, and O2, have no permanent dipole moment and cannot be detected using direct infrared vibrational absorption methods.30 Raman spectroscopy is an emerging technique31−36 that is based on the inherent molecular vibrations of the molecules.37−43 Hence, Raman gas spectroscopy would be ideally suited for rapid and simultaneous monitoring of various gases (H2, N2, and O2, etc.) without all the above-mentioned problems of saturation, aging, poisoning, transducers, and cross-sensitivities. However, the application of Raman spectroscopy for gas sensing has been limited so far, because of the weak inelastic scattering process and the low molecule density in gases. Recently, innovative enhancement techniques based on high finesse optical cavities44−46 and microstructured hollow core fibers47−53 pushed the sensitivity limits for Raman gas sensing down to low-ppm levels. Gas sensing in innovative optical hollow core fibers enables enhanced light−analyte interaction on long path lengths and at very small sample demand (nanolitres volume per centimeter fiber length). Thus, fiber-enhanced Raman spectroscopy (FERS) of molecular hydrogen within a multigas mixture is introduced in this contribution as a promising new methodology for noninvasive, painless breath gas analysis.
Figure 1. Schematic sketch of the experimental setup for fiberenhanced Raman spectroscopy (FERS). A recently introduced setup52 was further improved and consists of the following components: mass flow controllers (MFC1, MFC2), laser, telescope (T), edge filter (F), microscope objective (OL), fiber adapter assemblies (A1, A2), hollow core photonic crystal fiber (PCF), power meter (PM), pinhole (P), edge filter (E), aspheric lenses (L1, L2, L3), spectrometer (SPEC), and CCD detector.
rotational and vibrational Raman spectroscopy within the spectrally limited transmission window of the HCPCF (see Figure S1 in the Supporting Information), in order to analyze H2 (rotational bands) simultaneously with other gases (vibrational bands). Thus, an additional filter was implemented in the FERS setup in order to suppress the background signal of the fiber (as discussed later). This modal filter consists of a pinhole (P) in the focal plane of two achromatic lenses (L1 and L2). Different sizes of pinholes were used (10, 15, 25, 30, and 50 μm) for optimization of the signal-to-noise ratio (SNR). The filtered Raman signal was then focused with an achromatic lens (L3) on the slit of the spectrometer (SPEC) from Roper Scientific (Model Acton 2556), which was equipped with liquid-nitrogen-cooled CCD . A thermoelectrically cooled CCD can be used for point-of-care (POC) applications. A grating with 600 lines/mm (resolution of 3.5 cm−1/pixel at 557 nm), a slit size of 105 μm, and exposure times of 1−100 s were used for all experiments, if not stated otherwise. For the quantification of H2, the rotational peak S0(1) was integrated at 587 cm−1. A mixture of different concentrations of various gases was realized with help of mass flow controllers (MFC_1 and MFC_2) with three flow rate ranges (3−10, 93−281, and 861− 2600 sccm) and a forming gas from Linde AG (consisting of 5% H2 and 95% N2). A minimal concentration of 10 ppm hydrogen (with a relative concentration error of 3.9%) defined the lower limit and 5 vol % defined the upper limit of the H2 concentrations in the performed experiments. The dead volume of the entire setup consisting of MFC, HCPCF, and all tubing was flushed with pure nitrogen or argon between every mixing step. Every gas mixture passed a sinter filter with 0.5 μm pore size in order to ensure homogeneous mixing before filling into the fiber. The routine for quantitative measurements of unknown gas concentrations within a multigas mixture was performed as previously described.52 The attenuation of the HCPCF was determined by transmission measurements for two different fiber lengths (0.3 m, 2 m) with a white light source and an optical spectrum analyzer (OSA) in the wavelength range of 500−950 nm (see Figure S1 in the Supporting Information). The optical transmission characteristic of the fiber in the bandgap range was proven to be unaffected for the applied slight bending of
■
MATERIALS AND METHODS Raman spectroscopy is an extremely versatile technique with great selectivity, but conventionally suffers from low sensitivity. The Stokes Raman intensity (IStokes) depends on the laser intensity (I0) and the number (N) of molecules that contribute to the signal as well as the frequencies of the laser and scattered light and the molecular polarizability (ωL, ωS, and α), respectively: IStokes ≈ N × I0 × (ωL − ωS)4 × |α|2
(1)
Innovative hollow core photonic crystal fibers (HCPCFs) can tremendously enhance the interaction of the guided light and gas molecules52 (factors N and I0) and provide a highly miniaturized sample container. Thus, a recently introduced setup for fiber-enhanced Raman spectroscopy (FERS)52 was further improved (Figure 1) in order to achieve the necessary enhancement factors of 3 orders of magnitude and high signalto-noise ratios, which are implicitly essential for high accuracy analysis of low-concentration gas measurements of H2 and CH4 for performing the hydrogen breath tests. The FERS setup consists of two compact switchable lasers. These diode-pumped solid-state lasers (DPSS) have a small footprint size and the following parameters: λL1 = 607 nm, IL1 = 60 mW (Lasos Lasertechnik), and λL2 = 660 nm, IL2 = 1 W (Laser Quantum). The excitation laser light passed a telescope (T), was reflected by a long pass edge filter (F) into a microscope objective Nikon SLWD 20×/0.35 (OL) and focused into the HCPCF (NKT). Both ends of the fiber were fixed in specially constructed fiber connectors (A).36 The backscattered Raman signal from one fiber end (A1) was collected by the same objective and passes the edge filter (F). One of the major innovations of this paper is the combined 983
DOI: 10.1021/ac503450y Anal. Chem. 2015, 87, 982−988
Article
Analytical Chemistry
Figure 2. (A) Fiber-enhanced ro-vibrational Raman bands of (A) the Q-branch of H2 and (B) the S-branch of H2.
the fiber, which is prerequisite for reliable quantification. The spectral bandgap, with attenuation values smaller than 5.5 dB/ m, ranges from 607 nm to 740 nm (Figure S1 in the Supporting Information). Thus, Raman spectra can be acquired with low attenuation up to ∼3200 cm−1 for excitation λL1 = 607 nm. This spectral range still covers the vibrational stretching mode of CH454 at 2917 cm−1 but not the vibrational modes of H2 at 4156 cm−1 anymore. Excitation with λL2 = 660 nm provides light guidance with very low attenuations of 2.9 dB/m in the middle of the bandgap and thus results to strong Raman signals up to 1640 cm−1, including the vibrational stretching mode of O2 at 1556 cm−1. For both excitation wavelengths, an extremely low attenuation of the Raman signal of the pure rotational bands of H2 (ranging from 350 cm−1 to 1050 cm−1) is ensured by the bandgap characteristics. An excellent incoupling efficiency of more than 90% of the laser light into the fiber was provided by the FERS setup (Figure 1), as calculated by Beer−Lambert law with the help of the attenuation curve (see Figure S1 in the Supporting Information) and the measured light transmission power.
■
The vibrational Raman transition of hydrogen is dominated by the separated Q-branch lines (∼4156 cm−1; see Figure 2A). Although this molecular vibration occurs in a very high wavenumber region with relatively high attenuation losses of the Raman signal, the hydrogen Q-branch is still detectable, because of a generally high enhancement provided by the FERS setup. The individual bands of the Q-branch are determined by the thermal population of rotational ground-state energy levels with an intensity relation of 3:1 between odd and even J-values, because of different spin states.30 Since hydrogen is allocated in two nuclear spin states (I = 1 and I = 0), Q1(0) and Q1(2) are related to para-H2, Q1(1) and Q1(3) to ortho-H2, respectively. The peak of Q1(1) at 4156 cm−1 shows the highest intensity (Figure 2A). In the rotational Raman spectrum of hydrogen (Figure 2B) at lower wavenumbers, para-H2 is represented by the states S0(0) and S0(2), whereas S0(1) and S0(3) belong to ortho-H2. The most intensive rotational band (at room temperature) S0(1) is located at 587 cm−1. The advantages and disadvantages of the HCPCF for highly sensitive H2 analysis must be carefully examined and balanced against each other. The bandgap characteristic of the fiber provides low attenuation for the rotational bands; however, Figure 2A shows that Raman measurements at the high wavenumber region (∼4150 cm−1) are effectively backgroundfree and no other molecular ro-vibrational (or rotational) bands will emerge. By thoroughly examining the signal-to-noise ratio (SNR) of the strongest rotational S0(1) and ro-vibrational Q1(1) bands of hydrogen, it was found that the SNR of S0(1) is 644 times higher than the SNR of Q1(1) (see Table 1). Thus, for the following trace gas experiments, the area of the peak of S0(1) at 587 cm−1 was integrated for the quantification of H2. In order to quantify the enhancement of the FERS setup, the Raman peak at 587 cm−1 was first analyzed with fiber and second under identical conditions with carefully removed fiber.
RESULTS AND DISCUSSION
In order to develop fiber-enhanced Raman spectroscopy (FERS) as a novel miniaturized point-of-care (POC) technique for easy and rapid performance of the hydrogen breath test (HBT), the spectral signatures of H2 and CH4 within a multigas mixture were thoroughly examined; then, methodological improvements were developed and finally proven on a simulated positive HBT. Combination of Rotational and Vibrational Raman Spectroscopy for Highly Sensitive FERS Analysis of H2 within a Spectrally Limited Bandgap Range. The key idea for highly sensitive FERS analysis of H2, simultaneously with CH4 and other gases, is the combination of rotational and vibrational Raman spectroscopy. Thus, the great advantages of the hollow core photonic crystal fiber (HCPCF) for confinement of light and gas molecules and the extremely low attenuation of the guided light can also be exploited for H2 sensing within the spectrally limited bandgap range of the fiber (see Figure S1 in the Supporting Information). This combination of rotational and vibrational Raman spectroscopy for sensing H2 simultaneously with other gases within a limited spectral range is also of essential importance for further miniaturization with compact and robust spectrometers, which are conventionally equipped with fixed dispersion gratings and provide only a limited spectral detection range.
Table 1. Comparison of Signal-to-Noise Ratios (SNR) for the Strongest Rotational and Vibrational Modes and Relative Raman Cross Sections Related to N2a hydrogen band
relative Raman cross section at 607 (660) nm
fiber transmission for band (%)
signal-tonoise ratio, SNR
S0(1) (rotational) Q1(1) (rovibrational)
4.7 (4.9)
75.0
8371
3.4 (3.2)
0.2
13
a
984
Data taken from ref 55. DOI: 10.1021/ac503450y Anal. Chem. 2015, 87, 982−988
Article
Analytical Chemistry
Figure 3. (A) Fiber-enhanced rotational Raman spectrum of the S-branch of H2 and the background silica Raman signal of the fiber. The reduction of the silica background is shown for several diameters (10 μm (1), 15 μm (2), 25 μm (3), no pinhole) of the filtering pinhole. (B) Electron microscopic picture of the fiber end-face. The circular areas marked by “1”, “2”, “3”, and “4” equal the collection areas with pinholes sizes of 10, 15, 25, and 50 μm in the focal plane, respectively (see Figure 1).
(in comparison to a SNR of 1162 without pinhole filtering), as illustrated in Figure 3 and summarized in Table 2.
From both measurements, a signal enhancement of 5249 was derived for FERS, in contrast to conventional measurements with no fiber. One drawback of the application of the HCPCF is the background Raman signal of the silica material of the fiber itself. The silica Raman signal consist mainly of the W1 band at 200− 500 cm−1 resulting from bending motions of the six-member rings, the defect bands D1 (500 cm−1) and D2 (650 cm−1) related to breathing modes of four- and three-membered rings, unresolved TO-LO phonon pairs W3, and also TO and LO phonons at 1060 and 1190 cm−1.56 This silica Raman background signal restricts the limit-of-detection (LOD) for sensing of H2 by means of the rotational Raman bands in the lower wavenumber region from 300 cm−1 to 1000 cm−1. In order to minimize this background, a separation can be achieved by spatial filtering due to the different geometric distribution of the silica Raman signal (originating from the cladding) and the H2 Raman signal (from the fiber core). First, the spectrometer was used as a spatial filter; then, an additional pinhole (P) was inserted in the setup (recall Figure 1). This pinhole is operating as a mode filter in the focal plane, to eliminate the higher modes of the silica noise from the Gaussian-like shaped H2 Raman signal. The microscope objective (OL, f =10 mm) and the achromatic lens (L1, f = 35 mm) magnify the fiber core diameter of 6 μm (and the mode field diameter of 4.4 μm) to an effective diameter of 21 μm (and 15 μm, respectively) in the focal plane (Figure 1). Different pinhole sizes were tested (Figure 3) in order to suppress the higher-order silica Raman modes (mainly distributed in the outer core region and inner cladding region) from the H2 Raman fundamental mode (and partly the higherorder modes of the H2 Raman signal). The use of pinhole sizes larger than 25 μm in diameter did not provide any benefit by mode-filtering. The pinhole sizes of 10 and 15 μm represented effective collection diameters of 2.9 and 4.3 μm on the fiber end face (Figure 3B) and provided an improved SNR of the rotational band S0(1) of H2 by a factor of 6 and 3, respectively
Table 2. Comparison of Different Pinholes Sizes, Effective Collection Diameters, H2 Raman Counts, Background Noise, and Resulting SNR of the Rotational Band S0(1) of H2 pinhole size (μm)
effective diameter on fiber cross section (μm)
Raman signal for S0(1) (counts/s)
Raman noise (counts/s)
signal-tonoise ratio, SNR
10 15 25
2.9 4.3 7.1
47122 108759 100995
7.0 33.4 64.9
6741 3259 1556
The extremely important silica noise reduction was an essential prerequisite to achieve better SNR values for highly sensitive detection of H2 by FERS and enabled a linear calibration of the Raman intensity of the S0(1) peak of H2 at 587 cm−1 in dependency from the H2 concentration (see Figure S2 in the Supporting Information) in the range from 10 ppm up to 5 vol % (50 000 ppm), in accordance to eq 1. Although the maximum hydrogen concentration was fixed to 5 vol %, the sensor range can be extended to 100 vol %, based on the linear dependency of the Raman intensity (eq 1). A minimal silica Raman background noise of 2.1 counts/s was measured for 0 ppm of H2 in N2. The concentrations of 10 and 25 ppm of H2 resulted in value of 13.4 and 28.1 counts/s for the integration of the rotational Raman peak S0(1). Thus, the limit of detection (LOD) for H2 with this FERS setup is 4.7 ppm (SNR = 3). Highly Selective and Sensitive Detection of H2 and CH4 within a Complex Gas Mixture for Early Stage Diagnosis of Metabolic Disorders or Diseases. With the help of these methodological improvements, it was now possible to monitor the concentrations of H2 and CH4 selectively within a complex gas mixture such as human breath (e.g., CO2, O2, and N2) with just one single Raman measurement. The simultaneous FERS monitoring of such 985
DOI: 10.1021/ac503450y Anal. Chem. 2015, 87, 982−988
Article
Analytical Chemistry complex multigas mixture was acquired for the first time (Figure 4), and it was nicely shown that all the individual gases
Table 3. Assignment of the Individual Peaks in the Raman Spectrum of the Multicomponent Gas Mixture (see Figure 5) and relative Raman Scattering Cross Sections at λexc = 607 nm Raman transition
Raman shifta (cm−1)
relative Raman scattering cross section at 607 nm
reference
H2
S0(1) S0(2) S0(3) S0(4) Q1(1)
587r 814r 1034r 1243r 4156v
4.66 0.65 0.42 0.01 3.43
57 57 57 57 55
N2O
2ν2 ν1 ν3
1168v 1284v 2224v
0.08 2.34 0.51
58 55 55
12
CO2
hot band 2ν2 ν1 hot band
1270v 1285v 1388v 1410v
0.08 0.95 1.48 0.14
59, 60 55 55 59, 60
13
CO2
ν1*
1370v
1.48
55
O2
1556v
1.26
55
N2
2331v
1.00
55
1535v 2585v 2917v 3021v
0.10 0.01 5.78 0.18
54 61 55 61
gaseous component
Figure 4. Fiber-enhanced Raman spectrum of the rotational and rovibrational bands of a complex gas mixture: (A) rotational Raman bands of H2 and ro-vibrational Raman bands of CO2, N2O, O2, CH4, and N2 in the wavenumber region 550−1800 cm−1; and (B) rovibrational Raman bands of N2O, N2, CH4, and H2 in the wavenumber region 2200−4300 cm−1. A spectrometer grating with 1800 lines/mm was used for these measurements.
CH4
a
(H2, CO2, N2O, O2, CH4, and N2) can be selectively quantified due to their sharp and separated Raman peaks. The thorough interpretation of this complex rotational/ro-vibrational Raman spectrum is prerequisite for further breath gas experiments. Besides the spectral signatures of H2 (compare Figure 2) one can also observe the typical Fermi diad of the major natural isotope 12CO2 (ν̃− at 1285 cm−1 and ν̃+ at 1388 cm−1) as well as the hot bands of 12CO2 (ν̃−(1) at 1270 cm−1 and ν̃+(1) at 1410 cm−1) and the isotope 13CO2 (ν̃+ at 1370 cm−1). The spectrally resolved Q-branch of 14N2 at 2331 cm−1 is accompanied by the ro-vibrational transitions of the O- and S-branches, which show the typical (2J + 1) degeneracy, in combination with the Boltzmann distribution (approximately defined as exp(−J2)) for these transitions, and an intensity alternation of 2:1, caused by the statistical weight of the nuclear spin states of even and odd J. Similarly, the component 16O2 shows a ro-vibrational fine structure around the Q-branch transitions at 1556 cm−1 with no intensity variation (odd J, because of a total spin number of I = 0). Molecular methane shows several vibrational transitions, namely, the CH symmetric stretching band ν1 (at 2971 cm−1) including a ro-vibrational fine structure, the asymmetric stretching mode ν3 (at 3021 cm−1), the bending mode ν2 (at 1535 cm−1), and the bending mode overtone 2ν4 (at 2585 cm−1). All assignments are summarized in Table 3. With help of these important improvements, it was now possible to use the novel FERS methodology for the analysis of biomarkers in exhaled breath and thus for early stage diagnosis of metabolic disorders or diseases. As explained in the beginning of this report, H2 is only observed within breath in concentrations up to several tens of ppm, because of the
ν2 2ν4 ν1 ν3
Superscript legend: r, rotational mode; v, vibrational mode.
anaerobic metabolism of colon bacteria Escherichia coli during malabsorption disorders, and an increase of 20 ppm above basal within 90 min after oligosaccharide (lactose) ingestion is considered to be a positive HBT. Thus, a precise concentration series of three sequential measurements of 5, 10, 20, and 50 ppm of H2 was performed (Figure 5A). It is nicely shown that the FERS technique can be exploited for monitoring a rise by 20 ppm of H2 in the HBT and an increase of more than 12 ppm of H2 as an indicator for bacterial overgrowth in the small intestine (Figure 5A). Monitoring of the CH4 level can also be used as indicator for lactose intolerance (due to the conversion of H2 into CH4 by the large bowel bacteria Methanobrevibacter smithii). Therefore, the ability of FERS for precise monitoring of concentrations of 5, 10, 20, and 50 ppm of CH4 was also proven (Figure 5B). A cutoff value of 20 ppm of CH4 above basal can be monitored as indicator for lactose intolerance (Figure 5B). The monitoring of H2 and CH4 (Figure 5) as combined biomarkers for colonic fermentation of undigested oligosaccharides can improve the correct interpretation and help to find the right diagnosis.
■
CONCLUSION AND OUTLOOK A novel methodology for early-stage, noninvasive, point-of-care (POC) breath gas diagnosis of metabolic diseases was introduced based on combined rotational and ro-vibrational fiber enhanced Raman spectroscopy. The analytical capabilities were thoroughly analyzed and a spatial mode filter was included in the focal plane, such that fast monitoring of molecular 986
DOI: 10.1021/ac503450y Anal. Chem. 2015, 87, 982−988
Article
Analytical Chemistry
Figure 5. Simulation of fiber-enhanced Raman spectroscopic monitoring of typical concentrations in a hydrogen breath test (HBT): (A) stepwise monitoring of the Raman signal of H2 at concentrations of 5, 10, 20, and 50 ppm and (B) stepwise monitoring of the Raman signal of CH4 at concentrations of 5, 10, 20, and 50 ppm. The values of the Raman intensities were corrected for the background values for 0 ppm of H2 (15 counts) and 0 ppm of CH4 (6 counts), respectively.
■
ACKNOWLEDGMENTS Funding of the research project by the Free State of Thuringia (Germany) and the European Union (EFRE) is highly acknowledged (FKZ: 2012 FGR 0013). S.H. was supported by the International Max-Planck Research School “Global Biogeochemical Cycles”. T.B. is grateful for support from the “Studienstiftung des Deutschen Volkes”. The authors thank Anka Schwuchow for her help in obtaining the fiber attenuation characteristics.
hydrogen can be achieved down to a LOD of 4.7 ppm. The simultaneous monitoring of H2 and CH4 within a mixture of the major breath components N2, O2, 12CO2, and 13CO2 was demonstrated. The new FERS sensor provides an excellent linearity and a dynamic range of 5 orders of magnitude. The combined rotational and vibrational Raman spectroscopy enables the exploitation of the great advantages of innovative microstructured hollow core photonic crystal fibers (HCPCF) for enhanced analytical sensitivity and low sample demand as well as the use of miniaturized spectrometers (with limited spectral range) for the design of affordable and robust POC devices. These unique abilities of the novel multigas sensor were proven with the precise monitoring of H2 and CH4 in the concentration range of 5−50 ppm in order to demonstrate the potential for the diagnosis of malabsorption disorders, including lactose intolerance, fructose malabsorption, and bacterial overgrowth syndrome. This novel FERS hydrogen sensor provides a manifold of advantages compared to other types of hydrogen sensors: (i) in comparison to electrical sensors, the FERS sensor could be operated in explosive environments, because it does not generate sparks and has no electrical contacts; (ii) the FERS sensor shows no cross-sensitivity for other gases, such as carbohydrates or CO; and (iii) the FERS sensor can be used for remote sensing. Moreover, the abilities of the new FERS sensor for simultaneous monitoring of H2 and CH4 will find various additional applications in the context of novel hydrogen-based energy systems (e.g., “power to gas”)62,63 and biogas production.64
■
■
ASSOCIATED CONTENT
S Supporting Information *
This material is available free of charge via the Internet at http://pubs.acs.org.
■
REFERENCES
(1) Buszewski, B.; Kęsy, M.; Ligor, T.; Amann, A. Biomed. Chromatogr. 2007, 21, 553−566. (2) Calloway, D. H.; Murphy, E. L.; Bauer, D. Am. J. Dig. Dis. 1969, 14, 811−815. (3) Gibson, P. R.; Newnham, E.; Barrett, J. S.; Shepherd, S. J.; Muir, J. G. Aliment. Pharmacol. Ther. 2007, 25, 349−363. (4) Eisenmann, A.; Amann, A.; Said, M.; Datta, B.; Ledochowski, M. J. Breath Res. 2008, 2, 046002. (5) Ghoshal, U. C. J. Neurogastroenterol. Motil. 2011, 17, 312−317. (6) Yu, D.; Cheeseman, F.; Vanner, S. Gut 2011, 60, 334−340. (7) de Lacy Costello, B. P.; Ledochowski, M.; Ratcliffe, N. M. J. Breath Res. 2013, 7, 024001. (8) Ghoshal, U. C.; G, U.; Das, K.; Misra, A. Indian J. Gastroenterol. 2006, 25, 6−10. (9) Simren, M.; Stotzer, P. O. Gut 2006, 55, 297−303. (10) Ghoshal, U. C.; Srivastava, D. World J. Gastroenterol. 2014, 20, 2482−2491. (11) Stetter, J. R.; Li, J. Chem. Rev. 2008, 108, 352−366. (12) Martin, L. Solid State Ionics 2004, 175, 527−530. (13) Aroutiounian, V. Int. J. Hydrogen Energy 2007, 32, 1145−1158. (14) Potje-Kamloth, K. Chem. Rev. 2008, 108, 367−399. (15) Boon-Brett, L.; Black, G.; Moretto, P.; Bousek, J. Int. J. Hydrogen Energy 2010, 35, 7652−7663. (16) Iannuzzi, D.; Slaman, M.; Rector, J.; Schreuders, H.; Deladi, S.; Elwenspoek, M. Sens. Actuators, B 2007, 121, 706−708. (17) Chou, Y.-I.; Chiang, H.-C.; Wang, C.-C. Sens. Actuators, B 2008, 129, 72−78. (18) Drafts, B. IEEE Trans. Microwave Theory Tech. 2001, 49, 795− 802. (19) Wan, J. K. S.; Ioffe, M. S.; Depew, M. C. Sens. Actuators, B 1996, 32, 233−237. (20) D’Amico, A.; Palma, A.; Verona, E. Sens. Actuators 1982, 3, 31− 39. (21) Butler, M. A. Appl. Phys. Lett. 1984, 45, 1007. (22) Ando, M. TrAC, Trends Anal. Chem. 2006, 25, 937−948. (23) Silva, S. F.; Coelho, L.; Frazao, O.; Santos, J. L.; Malcata, F. X. IEEE Sens. J. 2012, 12, 93−102.
AUTHOR INFORMATION
Corresponding Author
*E-mail addresses:
[email protected], torsten.frosch@ gmx.de. Notes
The authors declare no competing financial interest. 987
DOI: 10.1021/ac503450y Anal. Chem. 2015, 87, 982−988
Article
Analytical Chemistry (24) Villatoro, J.; Luna-Moreno, D.; Monzón-Hernández, D. Sens. Actuators, B 2005, 110, 23−27. (25) Boonbrett, L.; Bousek, J.; Castello, P.; Salyk, O.; Harskamp, F.; Aldea, L.; Tinaut, F. Int. J. Hydrogen Energy 2008, 33, 7648−7657. (26) Hübert, T.; Boon-Brett, L.; Black, G.; Banach, U. Sens. Actuators, B 2011, 157, 329−352. (27) Boonbrett, L.; Bousek, J.; Moretto, P. Int. J. Hydrogen Energy 2009, 34, 562−571. (28) Boonbrett, L.; Bousek, J.; Black, G.; Moretto, P.; Castello, P.; Hübert, T.; Banach, U. Int. J. Hydrogen Energy 2010, 35, 373−384. (29) Ottens, A. K.; Harrison, W. W.; Griffin, T. P.; Helms, W. R. J. Am. Soc. Mass Spectrom. 2002, 13, 1120−1128. (30) Haken, H.; Wolf, H. C. Molecular Physics and Elements of Quantum Chemistry: Introduction to Experiments and Theory, 2nd ed.; Springer: Berlin, 2004. (31) Frosch, T.; Koncarevic, S.; Becker, K.; Popp, J. Analyst (Cambridge, U.K.) 2009, 134, 1126−1132. (32) Frosch, T.; Koncarevic, S.; Zedler, L.; Schmitt, M.; Schenzel, K.; Becker, K.; Popp, J. J. Phys. Chem. B 2007, 111, 11047−11056. (33) Frosch, T.; Meyer, T.; Schmitt, M.; Popp, J. Anal. Chem. 2007, 79, 6159−6166. (34) Frosch, T.; Popp, J. J. Biomed. Opt. 2010, 15, 041516. (35) Frosch, T.; Tarcea, N.; Schmitt, M.; Thiele, H.; Langenhorst, F.; Popp, J. Anal. Chem. 2007, 79, 1101−1108. (36) Frosch, T.; Yan, D.; Popp, J. Anal. Chem. 2013, 85, 6264−6271. (37) Frosch, T.; Küstner, B.; Schlücker, S.; Szeghalmi, A.; Schmitt, M.; Kiefer, W.; Popp, J. J. Raman Spectrosc. 2004, 35, 819−821. (38) Frosch, T.; Popp, J. J. Mol. Struct. 2009, 924−926, 301−308. (39) Frosch, T.; Schmitt, M.; Bringmann, G.; Kiefer, W.; Popp, J. J. Phys. Chem. B 2007, 111, 1815−1822. (40) Frosch, T.; Schmitt, M.; Noll, T.; Bringmann, G.; Schenzel, K.; Popp, J. Anal. Chem. 2007, 79, 986−993. (41) Frosch, T.; Schmitt, M.; Popp, J. Anal. Bioanal. Chem. 2007, 387, 1749−1757. (42) Frosch, T.; Schmitt, M.; Popp, J. J. Phys. Chem. B 2007, 111, 4171−4177. (43) Frosch, T.; Schmitt, M.; Schenzel, K.; Faber, J. H.; Bringmann, G.; Kiefer, W.; Popp, J. Biopolymers 2006, 82, 295−300. (44) Frosch, T.; Keiner, R.; Michalzik, B.; Fischer, B.; Popp, J. Anal. Chem. 2013, 85, 1295−1299. (45) Keiner, R.; Frosch, T.; Hanf, S.; Rusznyak, A.; Akob, D. M.; Kusel, K.; Popp, J. Anal. Chem. 2013, 85, 8708−8714. (46) Keiner, R.; Frosch, T.; Massad, T.; Trumbore, S.; Popp, J. Analyst 2014, 139, 3879−3884. (47) Russell, P. Science 2003, 299, 358−362. (48) Buric, M. P.; Chen, K. P.; Falk, J.; Woodruff, S. D. Appl. Opt. 2008, 47, 4255−4261. (49) Buric, M. P.; Chen, K. P.; Falk, J.; Woodruff, S. D. Appl. Opt. 2009, 48, 4424−4429. (50) Hartung, A.; Kobelke, J.; Schwuchow, A.; Wondraczek, K.; Bierlich, J.; Popp, J.; Frosch, T.; Schmidt, M. A. Opt. Express 2014, 22, 19131. (51) Benabid, F.; Couny, F.; Knight, J. C.; Birks, T. A.; Russell, P. S. J. Nature 2005, 434, 488−491. (52) Hanf, S.; Keiner, R.; Yan, D.; Popp, J.; Frosch, T. Anal. Chem. 2014, 86, 5278−5285. (53) Benabid, F.; Knight, J. C.; Antonopoulos, G.; Russell, P. S. J. Science 2002, 298, 399−402. (54) Schrötter, H. W.; Klöckner, H. W. In Raman Spectroscopy of Gases and Liquids; Weber, A., Ed.; Springer: Berlin, Heidelberg, Germany, 1979; pp 123−166. (55) Fenner, W. R.; Hyatt, H. A.; Kellam, J. M.; Porto, S. P. S. J. Opt. Soc. Am. 1973, 63, 73−77. (56) Dove, P. M.; Han, N.; Wallace, A. F.; De Yoreo, J. J. Proc. Natl. Acad. Sci. U.S.A. 2008, 105, 9903−9908. (57) Flora, F.; Giudicotti, L. Appl. Opt. 1987, 26, 4001−4008. (58) Anderson, A.; Sun, T. S. Chem. Phys. Lett. 1971, 8, 537−542. (59) Eichmann, S. C.; Weschta, M.; Kiefer, J.; Seeger, T.; Leipertz, A. Rev. Sci. Instrum. 2010, 81, 125104−125107.
(60) Eggers, D. F.; Crawford, B. L. J. Chem. Phys. 1951, 19, 1554. (61) Crawford, M. F.; Welsh, H. L.; Harrold, J. H. Can. J. Phys. 1952, 30, 81−98. (62) Conte, M. J. Power Sources 2001, 100, 171−187. (63) Schiermeier, Q. Nature 2013, 496, 156−158. (64) Lu, F.; Ji, J.; Shao, L.; He, P. Biotechnol. Biofuels 2013, 6, 92.
988
DOI: 10.1021/ac503450y Anal. Chem. 2015, 87, 982−988
