The enhancement of light-matter coupling when light is confined to wavelength scale volumes is useful both for studying small sample volumes and increasing the overall sensing ability. At these length scales, nonradiative interactions are of key interest to which near-field optical techniques may reveal new phenomena facilitating next-generation material functionalities and applications. Efforts to develop novel chemical or biological sensors using metamaterials have yielded innovative ideas in the optical and terahertz frequency range whereby the spatially integrated response over a resonator structure is monitored via the re-radiated or leaked light. But although terahertz waves generally exhibit distinctive response in chemical molecules or biological tissue, there is little absorption for subwavelength size sample and therefore poor image contrast. Here, we introduce a method that spatially resolves the differential near-field phase response of the entire resonator as a spectral fingerprint. By simultaneously probing two metallic ring resonators, where one loaded with the sample of interest, the differential phase response is able to resolve the presence of guest molecules (e.g. methanol) as they are adsorbed or released within the pores of a prototypical porous coordination polymer.
© 2014 Optical Society of America
Terahertz (THz) time-domain spectroscopy [1,2] is well known to have great potential for chemical, biological, and medical sensing applications . In this frequency range, sensing of molecular systems relies on the inter- and intramolecular vibrations between weakly bound or high-molecular mass species . The structural and/or chemical properties of a specimen uniquely alter the shape of a THz wave during interaction. This means that the amplitude or phase of the electric field waveform can act as a spectral fingerprint of a chemical or physical change [1–4]. However, for many practical applications in chemical sensing or imaging of bio-systems, free-space spectroscopy still lacks the required sensitivity (e.g. for weakly absorbing materials). This is more critical when it comes to in situ measurements of samples that evolve dynamically. Such experiments are highly demanding in terms of acquisition speed versus sensitivity, and may additionally suffer from the intrinsically poor lateral spatial resolution  (i.e. 150 μm at a frequency of 1 THz considering the classical diffraction limit ~λ/2).
To enhance the sensing capability of electromagnetic waves in the optical and THz frequency ranges, various avenues arise from the recent advances in ultrafast plasmonics for extreme light concentration and manipulation [5–9]. From the longstanding question of the sub-wavelength interactions between metallic holes and incident electromagnetic waves , successful sensing applications using THz resonator structures [11–15] and waveguides [16,17] were demonstrated. In order to resolve these phenomena spatially in the THz range, intense efforts towards overcoming the diffraction limit have been made over the past decade ([18,19] and references therein), with the common approach being based on raster scanning (i.e. single pixel measurement). Such time consuming measurements make the resolution of weakly absorbing substances particularly challenging, and delays progress toward in-vivo near-field THz imaging applications. Here, we report a method that combines a recently proposed technique for single-shot THz near-field imaging  with a simple resonator device consisting of two metallic ring structures, of which one is loaded with a dynamically evolving porous coordination polymer (PCP) nanocrystals at room temperature undergoing molecular adsorption in its pores. The spectroscopic information is extracted by differentiation of the near-field dynamics simultaneously resolved between two identical and loosely coupled metallic structures in which one acts as a reference.
2. Experimental procedures
The basic operation of the system is illustrated in Fig. 1(a), combining intense THz pulses  with low Q-factor resonator structures. The laser used for near-field THz spectroscopy of PCPs material was a commercial Ti:sapphire regenerative amplifier (Spitfire/Spectra-Physics) seeded by a mode-locked oscillator (Mai Tai/Spectra-Physics). The amplifier produced 4 mJ pulses with a center frequency of 780 nm at a repetition rate of 1 kHz. The generation and detection pulse durations were set independently to 100 fs in order to optimize the performance of the THz microscope. In this microscope, the conventional electro-optic (EO) sampling technique is expanded to two-dimensional imaging using high power monocycle THz wave. The two-dimensional spatial distribution of the THz near field is obtained by setting a sample directly on the top surface of EO crystal. The subwavelength spectral and spatial THz components are resolved through the use of a 16-bit charge coupled device (CCD) camera from Hamamatsu (model C9100-13), in which an effective region of 220 x 480 pixels (i.e., after balanced and dynamic subtractions) is captured at a repetition rate of 14 Hz. The resulting imaged area is 170 x 350 μm2 in real space, which enables us to extract the full electric near-field of both ring structures simultaneously. Each 2-dimensional time-domain movie is composed of 310 frames, each consisting of 40 averaged images, which covers 8 ps of time evolution and takes 15 min of acquisition .
To avoid potential THz field-induced damage in the vicinity of a high Q-factor resonator [20,22], we employ a sensing device that exhibits lower field enhancement but covers a larger area (e.g., greater than the microscope linear spatial resolution ). The two metallic ring structures (as illustrated in Fig. 1(a)) were fabricated by a standard photolithography and lift-off technique of a 100-nm-thick gold layer on the high reflecting coating side of a 10-μm-thick x-cut LiNbO3 (LN) wafer mounted on a 500-μm-thick glass substrate (Tempax from Schott). A 2-nm-thick chromium layer was deposited before for improved gold adhesion. The THz spectral amplitude response located at the center position of one ring pattern is shown in the upper inset of Fig. 1(a) (blue line), exhibiting a two-fold amplitude enhancement compared to a reference pulse without the ring structure (black line, Fig. 1(a)) for frequencies ranging from 0.5 to 2.5 THz. The THz and probe polarization are oriented parallel between each other and follow the y direction, as referenced in Fig. 1(a). In addition, aligning the two metallic rings in the y direction minimize their coupling since the main radiated fields are propagating in x direction.
The flourishing area of PCPs, also called metal–organic frameworks (MOFs), has provided an extensive class of crystalline materials featuring high permanent porosity and diverse pore surface chemistries . Their modular framework construction (depicted schematically in Fig. 1(b)) allows the physical and chemical properties to be tuned according to the components selected at the time of synthesis. They are currently receiving significant attention for a variety of applications, ranging from high-density gas storage for automotive applications (e.g. hydrogen or methane) and industrially important gas separations (e.g. carbon capture), to chemical sensing, catalysis, and drug release [23,24]. An extensive suite of spectroscopic tools that facilitate the study of the host-guest interactions between host PCPs and targeted guests, the understanding of which is crucial for constructing custom-designed materials, has been developed in recent times . To date, far-field THz spectroscopy of PCPs has brought new insight into the dynamics of water guest molecules within the pores . Employing the near field of THz waves resolved in amplitude and in phase should, in principle, improve the sensing capability  and provide a complete understanding of the signal response anywhere inside the resonator structure, as oppose to the single pixel measurements [11,15].
Here, we chose Cu3(btc)2 (btc = 1,3,5-benzenetricarboxylate) as the host framework, due to its well-known physical and chemical properties, and high affinity toward a variety of guest molecules . The sorption kinetics of PCPs materials was altered using an environment-controlled system, in which the partial vapor pressure of methanol compounds in He carrier gas is adjusted by mass ñow controllers ranging from 0 to 80% at 298 K with a total mass ñow of 100 cm3 • min−1 . This system was connected to a small gas cell for near-field THz sensing via 2 meters long plastic tubing. The samples were prebaked at 80 °C under nitrogen purge environment to remove any water molecules that are easily bound to the PCPs framework in ambient air.
In Fig. 2, the visible images of the sensor with (i) and without Cu3(btc)2 (iv) are shown. The THz snapshots of Fig. 2 (ii and iii) (i.e. extracted from Media 1 and Media 2, respectively) at a time delay of 3.25 ps show a live observation of methanol adsorption in the Cu3(btc)2 sample. In (ii), the spatial phase shift (SPS) between the reference ring (upper ring) and the sample ring (lower ring) correspond to the Cu3(btc)2 framework signature (i.e. unloaded sample against no sample under a flow of helium (He), which is not adsorbed in Cu3(btc)2 at room temperature). Then, switching of the gas composition from pure He to one containing methanol vapor (80% concentration) into the He flow produced an additional SPS (iii) in the THz wave trapped in the lower ring (i.e. methanol loading), whereas the reference wave (i.e. in the upper ring) remains unaffected. The absence of SPS for an empty sensor measurement, under He (v) (i.e. extracted from Media 3) or methanol (vi), confirmed that the observed SPS came from the Cu3(btc)2 sample and not from the atmosphere in the gas cell. It is important to emphasize that the integrated amplitude in the lower rings of frames (ii) and (iii) are similar. Thus, only the differential phase of the THz pulse resolved spatially is usable to extract the spectroscopic information here, while such information would vanish in the far-field.
Prior to simulating our result, we have measured the THz refractive index of the Cu3(btc)2 material by a THz time-domain spectroscopy (THz-TDS) system based on photoconductive antenna emission and detection. For that measurement, the Cu3(btc)2 material was prepared in a ~0.17 mm-thick pellet form and placed inside our gas cell under a He purge. The reference and signal THz pulses are presented in Fig. 3(a) with their corresponding Fourier transforms shown in the inset. The measured unloaded Cu3(btc)2 refractive index, = 1.20 ± 0.05 retrieved from the data taken in Fig. 3(a), is presented in Fig. 3(b) and was employed for simulations of the activated sample prior to exposure to methanol. The comprehensive analysis of our results was done using a 3-dimensional model based on a finite-difference time-domain (FDTD) method from the Lumerical software package. The simulations were performed using a linearly polarized THz excitation of Gaussian shape ranging from 0.1 to 2.5 THz in bandwidth. A 0.4 x 0.4 x 1 mm3 sized substrate and a 0.4 x 0.4 mm2 by 3-μm-thick high reflective coating material were set using the optical constants of SiO2. The thicker substrate was selected to extend the simulation range to 15 ps without disturbance of a second echo. A 0.4 x 0.4 by 10-μm-thick lithium niobate crystal with anisotropic THz refractive indexes nordinary = 6.46 and nextraordinary = 5.11 was sandwiched between the two glass layers.
In order to simulate the experiment, we have added and varied the complex refractive index of methanol (nMeOH) of a fractional part of the sample volume while keeping the remaining volume at a refractive index of n = 1.2. This scenario is depicted in Fig. 3(c). Figure 3(d) shows the near-field mapping of the electric field at a fixed time delay of 3.25 ps for three conditions: an empty sensor (i), a 180 μm-thick sample with constant refractive index of 1.2 placed on top of the lower ring (ii), and the same sample with 60% of its volume (which corresponds to the total pore volume) modified with nMeOH (iii). Note that the selected simulated time frames displayed in Fig. 3(d) are chosen to match the experimental image frames presented in Fig. 2.
In the near-field range, electromagnetic waves can have physical properties that are considerably different from their free-propagating counterparts . Particularly at the metal –crystal interface, the incoming beam can couple to various modes, such as surface plasmon polaritons  that might interfere inside the resonator at specific frequencies and localizations . Generally for single pixel measurements, these effects are averaged out in the response. With full 2-dimensional information, successful demonstrations were achieved for samples that exhibit strong intermolecular vibrational modes, with  and without a resonator . Here, PCP and methanol are free of spectral features in THz range rendering amplitude signal changes more difficult to monitor. To extract accurate spectroscopic information of these weakly absorbing materials, a phase differentiation method was adopted.
The first step in the analysis is to Fourier transform the time-domain two-dimensional near-field maps. By working in the frequency domain, we gain access to the amplitude and phase information as a function of THz frequency and resolved spatially (as described in ). Secondly, we normalized the unloaded phase maps with the sensor phase map responses (i.e. without sample), and the methanol loaded phase maps with the unloaded ones. This operation adjusts for the small initial phase deviation existing between the upper and lower rings. Consecutively, we differentiate the phase information between the upper and lower rings at each THz frequency. In Fig. 4(a) and (b), we show an example of the experimental and simulation phase maps at 0.72 and 0.73 THz, respectively, including the differentiated zones marked with the dotted circles. Each corresponding pixels inside the upper and lower rings were differentiate together and the averaged differentiated values (Δθ) were plot as a function of THz frequency. This technique insures good repeatability of extracting spectroscopic information from weakly absorbing medium in the nanogram scale using the phase of the THz waves. The signal-to-noise ratio (i.e. about >100 at any pixel in the phase data) also allows mapping of the sample signature at a micron scale anywhere inside the ring structure (i.e. limited by the linear spatial resolution of the system, which is about 10 μm here). At higher frequencies, we observed standing waves inside the rings and the appearance of nodes at resonant frequencies, which contains no useful information. The pixels at which these nodes appeared were omitted in the calculation of the Δθ values. Finally, we have tested the phase differentiation between the reference with and without sample and measured a constant phase (Δθ) of 0 ± 0.1 radian for single pixel measurement. This observation indicates the absence of cross talk between the reference and sample when the data are taken inside the ring structures.
In Fig. 4(c), we present the experimental phase differentiation (Δθ) plots for different gas exposure conditions as a function of THz frequency (i.e. extracted from the two-dimensional phase maps corresponding to these conditions). The black curve indicates the Δθ obtained for the empty sample under helium. The full blue circle curve shows the additional Δθ if we change the environment to methanol gas (i.e. when compare to the previous condition). The half blue circles show the Δθ curve 5 minutes after reinjection of helium gas instead of methanol gas. Waiting 12 more hours almost returns Δθ to a zero phase shift, when compared to the initial condition (see the empty blue circle).
In Fig. 4(d), we show the comparison of Δθ between experiment and simulation assuming a 180 μm-thick Cu3(btc)2 sample, respectively. The black curve was obtained for an activated sample (i.e. just after heating under He purge in experiment and for = 1.2 in simulation). The agreement between the black curve in Figs. 4(c) and (d) as a function of frequency confirms the choice in our simulation for a refractive index of 1.2 for the empty Cu3(btc)2 framework, as seen in Fig. 3(b). This signature is altered after introducing methanol vapor to the sample at a relative pressure (P/P0) of 0.8, as shown by the blue full circles of Fig. 4(c). Our method assumed a simple volume ratio weighting to correlate our simulations with the mixture of indices of refraction when methanol is loaded or unloaded. In Fig. 4(d), we have tried to replicate this condition by simulating Δθ for 60%, 50%, 30%, and 10% volume ratios of nMeOH when compared to the remaining volume of = 1.2 (see Fig. 4(d) the blue and red curves). Notice that we selected the complex refractive index of liquid methanol following the result from infrared spectroscopy of PCP under methanol gas exposure that clearly shows the broad peak -OH stretching vibration, typical for interconnected (hydrogen bonded) MeOH .
From a structural point of view of the Cu3(btc)2 framework, the maximum loading volume can be estimated from its solvent-accessible volume (ca. 65%), which translates to a methanol loading density of approximately 0.6 g per gram of the Cu3(btc)2 framework . According to the sensor size, the inspected volume of PCPs material for this experiment was approximately 0.9 μm3, which correspond to 300 ng of Framework. The maximum loaded of methanol was 180 ng (140 nl assuming a liquid phase inside PCPs material). Noticeably, when adding methanol, the best agreement between experiment and simulation is obtained for a volume fraction of 60% of nMeOH (see the full blue circles of Fig. 4(c) and 4(d)), also corresponding to the real available loaded volume-size as described above. This correlation was found for nMeOH using the single Debye model with parameter values derived from Ref . (i.e. were = 2.05, ε0 = 35.5 and τD = 48.3). The experimental agreement with this model strongly suggests the presence of a phase similar in density to liquid methanol within the framework pores, as also observed in  using infrared spectroscopy.
The subsequent data presented in Fig. 4(c) are the Δθ responses after 5 minutes and after 12 hours of purging with He gas, upon exposure to methanol. Interestingly, 5 minutes after starting the He purging that follows the methanol loading (blue half circles in Fig. 4(c)), the observed SPS indicates that the desorption process of Cu3(btc)2 is not completed and the material still contains adsorbed methanol. This observation is consistent with the prolonged purging time required for desorption of methanol from Cu3(btc)2 due to a so-called metastable state that emerges during desorption as reported previously . Here, the desorption rate during a He purge is temporarily slowed or even stopped due to the anchoring of stable, hydrogen-bonded methanol clusters to the exposed Cu(II) cation sites on the pore surface, and desorption slowly resumes once these clusters are disrupted. Indeed, purging for several hours under He (Fig. 4(c), open circles) results in a smaller SPS compared to the precedent measurement (Fig. 4(c), half-filled circles), although the methanol is not completely evacuated even after this time. As such, full desorption of methanol molecules is expected to require heating of the sample or the application of a high vacuum . Notice that we validated the sample integrity by X-ray diffraction technique before and after 8 hours of THz irradiation.
The differential method proposed in this paper enables to probe few hundred nanograms of weakly absorbing material, which introduces a significant reduction in required sample size for detection, sensing and imaging purposes in the THz range. Using our method provides a direct way of dynamically resolving the changes in the near-field electric field distribution that correspond to the loading and unloading of methanol guest molecules in PCP sample. According to our simulations, this technique allows quantifying and qualifying the loading of the guest molecules inside the PCP pores. The good agreement between experiment and simulation confirm the amount of guest molecules upon methanol gas exposure and strongly suggest a phase similar to liquid methanol inside the pores. This information is generally obtained through complementary techniques such as masses measurement and infrared spectroscopy. We envision this method as a new and powerful tool to probe the THz response of sub-wavelength and weakly absorbing materials, including studies of living cells.
We thank P. C. M. Planken for valuable discussions. The iCeMS is supported by World Premier International Research Center Initiative (WPI), MEXT, Japan. We wish to acknowledge financial support from MEXT/JSPS KAKENHI grants number 23244065. F. B. wishes to acknowledge NSERC and Le Fonds Québécois de la Recherche sur la Nature et les Technologies (FQRNT) contract number 138131 and MEXT/JSPS KAKENHI grants number 00571358. K. S. thanks the Japan Society for the Promotion of Science Postdoctoral Fellowship for Foreign Researchers. M. T. is grateful to the DAAD for a postdoctoral fellowship. This work was supported in part by Kyoto University Nano Technology Hub in “Nanotechnology Platform Project” sponsored by the Ministry of Education, Culture, Sports, Science and Technology (MEXT), Japan.
References and links
1. M. C. Nuss and J. Orenstein, “Terahertz Time-Domain Spectroscopy,” in Millimeter and Submillimeter Wave Spectroscopy of Solids, G. Grüner, ed., (Springer, Berlin, 1998).
2. D. Mittleman, Sensing with Terahertz Radiation (Springer-Verlag, 2003).
3. M. Tonouchi, “Cutting-edge terahertz technology,” Nat. Photonics 1(2), 97–105 (2007). [CrossRef]
4. M. Walther, B. M. Fischer, A. Ortner, A. Bitzer, A. Thoman, and H. Helm, “Chemical sensing and imaging with pulsed terahertz radiation,” Anal. Bioanal. Chem. 397(3), 1009–1017 (2010). [CrossRef] [PubMed]
6. J. A. Schuller, E. S. Barnard, W. Cai, Y. C. Jun, J. S. White, and M. L. Brongersma, “Plasmonics for extreme light concentration and manipulation,” Nat. Mater. 9(3), 193–204 (2010). [CrossRef] [PubMed]
7. L. Novotny and H. van Hulst, “Antennas for light,” Nat. Photonics 5(2), 83–90 (2011). [CrossRef]
8. N. Liu, T. Weiss, M. Mesch, L. Langguth, U. Eigenthaler, M. Hirscher, C. Sönnichsen, and H. Giessen, “Planar metamaterial analogue of electromagnetically induced transparency for plasmonic sensing,” Nano Lett. 10(4), 1103–1107 (2010). [CrossRef] [PubMed]
9. D. Brinks, M. Castro-Lopez, R. Hildner, and N. F. van Hulst, “Plasmonic antennas as design elements for coherent ultrafast nanophotonics,” Proc. Natl. Acad. Sci. U.S.A. 110(46), 18386–18390 (2013). [CrossRef] [PubMed]
10. T. W. Ebbesen, H. J. Lezec, H. F. Ghaemi, T. Thio, and P. A. Wolff, “Extraordinary optical transmission through sub-wavelength hole arrays,” Nature 391(6668), 667–669 (1998). [CrossRef]
11. P. H. Bolivar, M. Nagel, F. Richter, M. Brucherseifer, H. Kurz, A. Bosserhoff, and R. Büttner, “Label-free THz sensing of genetic sequences: towards ‘THz biochips’,” Philos. T. Roy. Soc. A 362(1815), 323–335 (2004). [CrossRef]
12. N. I. Landy, C. M. Bingham, T. Tyler, N. Jokerst, D. R. Smith, and W. J. Padilla, “Design, theory, and measurement of a polarization-insensitive absorber for terahertz imaging,” Phys. Rev. B 79(12), 125104 (2009). [CrossRef]
13. J. R. Knab, A. J. L. Adam, R. Chakkittakandy, and P. C. M. Planken, “Terahertz near-field microspectroscopy,” Appl. Phys. Lett. 97(3), 031115 (2010). [CrossRef]
14. H. Tao, W. J. Padilla, X. Zhang, and R. D. Averitt, “Recent Progress in Electromagnetic Metamaterial Devices for Terahertz Applications,” IEEE J. Sel. Top. Quantum Electron. 17(1), 92–101 (2011). [CrossRef]
15. J. R. Knab, A. J. L. Adam, E. Shaner, H. J. A. J. Starmans, and P. C. M. Planken, “Terahertz near-field spectroscopy of filled subwavelength sized apertures in thin metal films,” Opt. Express 21(1), 1101–1112 (2013). [CrossRef] [PubMed]
18. A. J. L. Adam, “Review of Near-Field Terahertz Measurement Methods and Their Applications,” Int. J. Infrared Millim. Waves 32(8-9), 976–1019 (2011). [CrossRef]
19. P. C. M. Planken, A. J. L. Adam, and D. Kim, “Terahertz Near-Field Imaging,” Spr. Ser. Opt. Sci. 171, 389–413 (2013). [CrossRef]
20. F. Blanchard, A. Doi, T. Tanaka, and K. Tanaka, “Real-Time, Subwavelength Terahertz Imaging,” Annu. Rev. Mater. Res. 43(1), 237–259 (2013). [CrossRef]
21. H. Hirori, A. Doi, F. Blanchard, and K. Tanaka, “Single-cycle terahertz pulses with amplitudes exceeding 1 MV/cm generated by optical rectification in LiNbO3,” Appl. Phys. Lett. 98(9), 091106 (2011). [CrossRef]
22. M. Liu, H. Y. Hwang, H. Tao, A. C. Strikwerda, K. Fan, G. R. Keiser, A. J. Sternbach, K. G. West, S. Kittiwatanakul, J. Lu, S. A. Wolf, F. G. Omenetto, X. Zhang, K. A. Nelson, and R. D. Averitt, “Terahertz-field-induced insulator-to-metal transition in vanadium dioxide metamaterial,” Nature 487(7407), 345–348 (2012). [CrossRef] [PubMed]
24. H. Uehara, S. Diring, S. Furukawa, Z. Kalay, M. Tsotsalas, M. Nakahama, K. Hirai, M. Kondo, O. Sakata, and S. Kitagawa, “Porous Coordination Polymer Hybrid Device with Quartz Oscillator: Effect of Crystal Size on Sorption Kinetics,” J. Am. Chem. Soc. 133(31), 11932–11935 (2011). [CrossRef] [PubMed]
25. M. Tsotsalas, P. Hejcik, K. Sumida, Z. Kalay, S. Furukawa, and S. Kitagawa, “Impact of Molecular Clustering inside Nanopores on Desorption Processes,” J. Am. Chem. Soc. 135(12), 4608–4611 (2013). [CrossRef] [PubMed]
26. K. Schröck, F. Schröder, M. Heyden, R. A. Fischer, and M. Havenith, “Characterization of interfacial water in MOF-5 (Zn4(O)(BDC)3)-a combined spectroscopic and theoretical study,” Phys. Chem. Chem. Phys. 10(32), 4732–4739 (2008). [CrossRef] [PubMed]
27. L. Novotny and B. Hecht, Principles of Nano-Optics Second edition, (Cambridge University Press, 2012).
28. C. R. Williams, S. R. Andrews, S. A. Maier, A. I. Fernández-Domínguez, L. Martín-Moreno, and F. J. García-Vidal, “Highly confined guiding of terahertz surface plasmon polaritons on structured metal surfaces,” Nat. Photonics 2(3), 175–179 (2008). [CrossRef]
30. C. D. Abeyrathne, M. N. Halgamuge, P. M. Farrell, and E. Skafidas, “An ab-initio Computational method to determine dielectric properties of biological materials,” Nature Sci. Rep. 3, 1796 (2013). [CrossRef] [PubMed]