amplification of refractometric biosensor response through...

COMMUNICATION

1700023 (1 of 7) © 2017 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.advmattechnol.de

Amplification of Refractometric Biosensor Response through Biomineralization of Metal–Organic Framework Nanocrystals

Jingyi Luan, Rong Hu, Sirimuvva Tadepalli, Jeremiah J. Morrissey, Evan D. Kharasch, and Srikanth Singamaneni*

DOI: 10.1002/admt.201700023

considered to be highly promising for the development of simple, portable, sensitive, on-chip biodiagnostics for resource-limi ted settings such as at-home care, rural clinics, developing countries with low and moderate incomes and the battlefield.

While there has been tremendous pro-gress in the rational design of nanotrans-ducers with high sensitivity and the devel-opment of handheld read-out devices, the translation of these biosensors to resource-limited settings is hindered by the poor thermal, chemical, and environmental sta-bility of the natural antibodies, which are the most commonly employed biorecogni-tion elements. We have previously demon-strated that artificial antibodies, achieved through molecular imprinting on plas-monic nanostructures, are a viable alter-native to natural antibodies.[4,5] Although the artificial antibodies exhibit excellent temperature, chemical and environmental

stability, the sensitivity of plasmonic biosensors based on artifi-cial antibodies is generally inferior compared to those based on natural antibodies due to the relatively low binding affinity of artificial antibodies compared to natural antibodies.

Methods to overcome the limited sensitivity of the plasmonic biosensors based on artificial antibodies are extremely impor-tant to make this class of biosensors relevant to biodiagnostic applications in the real world. Amplification of the sensor response to a biomolecular binding event through an enzymatic reaction is a powerful technique that is employed in enzyme-linked immunosorbent assay (ELISA) for achieving high sen-sitivity and low detection limit. A number of amplification

Plasmonic biosensors based on the refractive index sensitivity of localized surface plasmon resonance (LSPR) are highly promising for on-chip and point-of-care diagnostics. In particular, plasmonic biosensors that rely on artificial antibodies are highly attractive for applications in resource-limited settings due to the excellent thermal, chemical, and environmental stability of these biorecognition elements. In this work, a universal LSPR response amplification strategy based on the biomineralization of a metal–organic framework (MOF) on the captured analyte proteins is demonstrated. The amplification relies on the differential ability of abiotic recognition elements and captured biomolecules to induce biomineralization of a MOF. The rapid amplification process (less than 10 min) demonstrated here results in nearly 100% higher sensitivity and three times lower limit of detection compared to the innate sensor. The amplification approach can be broadly applied to a wide variety of bioanalytes and can be rapidly implemented in real-world con-ditions without compromising the assay time or reusability of the plasmonic biochip.

J. Luan, Dr. R. Hu, S. Tadepalli, Prof. S. SingamaneniDepartment of Mechanical Engineering and Materials ScienceInstitute of Materials Science and EngineeringWashington University in St. LouisSt Louis, MO 63130, USAE-mail: [email protected]. J. J. Morrissey, Prof. E. D. KharaschDepartment of AnesthesiologyDivision of Clinical and Translational ResearchWashington University in St. LouisSt. Louis, MO 63110, USA

Prof. J. J. Morrissey, Prof. E. D. Kharasch, Prof. S. SingamaneniSiteman Cancer CenterSt. Louis, MO 63110, USAProf. E. D. KharaschDepartment of Biochemistry and Molecular BiophysicsWashington University in St. LouisSt. Louis, MO 63110, USAProf. E. D. KharaschThe Center for Clinical PharmacologySt. Louis College of Pharmacy and Washington University School of MedicineSt. Louis, MO 63110, USA

Artificial Antibodies

Localized surface plasmon resonance (LSPR) of metal nano-structures involves the collective oscillation of dielectrically confined conduction electrons, which results in a number of unique optical properties such as large absorption and scat-tering cross sections and large enhancement of electromagnetic field surrounding metal nanostructures.[1] LSPR wavelength of metal nanostructures is highly sensitive to composition, size, shape, dielectric constant of the surrounding medium, and proximity to other metal nanostructures (plasmon coupling).[2] The sensitivity of LSPR wavelength to the changes in the refrac-tive index of the surrounding medium lends itself to a powerful class of refractometric biosensors.[3] LSPR biosensors are

Adv. Mater. Technol. 2017, 2, 1700023

www.advancedsciencenews.com

© 2017 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim1700023 (2 of 7)


strategies have been investigated in the context of plasmonic biosensors. The spectral shift of LSPR biosensor depends on the number of analyte molecules bound on the plasmonic nanotransducer, the size of the analyte species (which deter-mines the thickness of the adsorbate layer), and the refractive index difference between the analyte and the prior medium.[6] For the same number of molecules bound on the surface, a larger analyte molecule generally results in a thicker absorbate layer, thus inducing a larger spectral shift. This particular phe-nomenon has been exploited to amplify the LSPR shift. Fol-lowing the selective binding of the target analyte to the cap-ture antibodies immobilized on nanotransducers, exposure to a nonoverlapping primary antibody results in the binding of the primary antibody to the target analyte. The binding of the primary antibody to the analyte essentially increases the size of the adsorbate layer, thus increasing the LSPR shift.[7] Van Duyne and co-workers have demonstrated plasmon-coupling-based signal amplification using primary antibody conjugated to gold nanoparticles.[8] While these strategies are attractive, they require a specific nonoverlapping antibody and in some cases tedious labeling procedures that are expensive, time-consuming, and impractical in resource-limited settings. These considerations highlight the need for a rapid, inexpensive, and “universal” amplification strategy.

Owing to their large surface area, tunable porosity, organic functionality, and high thermal stability, metal–organic frame-works (MOFs), consisting of metal ions or clusters linked by organic ligands,[9] have received increased scientific and tech-nological interest.[10] In a recent study, Falcaro and co-workers demonstrated the encapsulation of a wide range of biomole-cules within MOFs by growing them in the presence of the biomolecules under mild and biocompatible conditions (e.g., aqueous solution at room temperature).[11] They have shown that a wide range of biomolecules (e.g., proteins and DNA) can efficiently localize MOF precursors, resulting in prenuclea-tion and rapid biomineralization of MOF crystals around the biomolecules. MOF growth exhibited an excellent spatial selec-tivity with crystal growth confined to the regions with immobi-lized biomolecules on solid substrates.[12]

Here, we demonstrate that biomineralization of MOF can be employed as a universal amplification strategy for LSPR biosen-sors based on abiotic recognition elements (e.g., artificial anti-bodies), where the captured target biomolecules serve as nucle-ation sites for the formation of MOF crystals on the nanotrans-ducer. The MOF crystals formed on the biomolecules increase the change in the refractive index and amplify the LSPR shift, effectively lowering the limit of detection of the biosensor. Owing to the widely available precursors and rapid growth, we have employed zeolitic imidazolate framework-8 (ZIF-8) as the MOF for LSPR amplification in this work. The MOF biomin-eralization-based amplification approach is universal and does not require any nonoverlapping antibodies or tedious labeling procedures. We demonstrate the generality of the amplification process using three different model analytes, namely, human serum albumin (HSA), lysozyme, and hemoglobin (Hb).

Prior to implementing the biomineralization-based ampli-fication, the preferential nucleation and growth of ZIF-8 on and around the biomolecules was investigated using atomic force microscopy (AFM). Owing to its atomically flat surface,

a freshly cleaved highly ordered pyrolytic graphite (HOPG) was employed as a substrate for the AFM investigation. HSA (molecular weight = 66.5 kDa) is one of the most abundant pro-teins in human blood, and has a heart-shaped spheroid struc-ture and dimensions of ≈9.5 × 5 × 5 nm.[13] HSA was adsorbed onto the HOPG surface from a dilute solution (200 ng mL−1). AFM images revealed uniform and sparse distribution of HSA on HOPG surface (Figure 1A) with an average height of ≈1.9 ± 0.7 nm (Figure 1C). The height of HSA globules adsorbed on HOPG was found to be significantly lower com-pared to the reported dimensions of HSA (9.5 × 5 × 5 nm) possibly due to the spreading of proteins upon adsorption on the surface. Subsequently, HSA immobilized HOPG substrate was immersed into MOF precursor solution, followed by thor-ough rinsing and drying under a stream of dry nitrogen. AFM images revealed globules with a significantly higher thickness (≈16.9 ± 4.5 nm) compared to HSA (1.9 ± 0.7 nm) (Figure 1B and the height histogram shown as Figure 1D). Conversely, the density and distribution of the globules were found to be similar to that of the HSA before exposure to MOF precursor, indicating the preferential growth of MOF around the sparsely adsorbed HSA (Figure 1A,B).

To further confirm the biomineralization of MOF, we have employed HSA-coated Au nanorods (AuNR). The HSA-coated AuNRs adsorbed on a glass substrate were exposed to MOF precursor solution. LSPR wavelength of HSA-coated AuNR exhibi ted a red shift of 15.1 nm after exposure to MOF precursor solution for 10 min indicating the formation of MOF crystals on the AuNR. As a negative control, pristine AuNR did not exhibit a red shift after being exposed to MOF precursor solution of identical concentration, which further confirms the preferential growth of MOF around the HSA biomolecule (Figure 1E). We have employed Raman spectroscopy and powder X-ray diffrac-tion (XRD) to confirm the chemical composition of the MOF formed around HSA. We observed strong Raman bands at 1146 and 1458 cm−1 corresponding to C5N stretching and methyl bending, respectively, which confirm the formation of ZIF-8 nanocrystals (Figure 1F, and Figure S1, Supporting Informa-tion).[14] XRD and Fourier transform infrared (FTIR) spectro-scopy further confirmed the formation of ZIF-8 nanocrystals on HSA (Figure 1G, and Figures S2 and S3, Supporting Informa-tion). The peak positions in the XRD pattern are in agreement with the expected structure of ZIF-8 except for the absence of (011) and (112) plane, implying the possible orientation of ZIF-8 formed on immobilized HSA.[11] The XRD pattern also shows a strong peak at 10.88°, indicating the partial orientation of the crystals in (001) direction. It has been previously reported that the surface properties can significantly influence the nucle-ation and crystal growth of ZIF-8.[15]

Now we turn our attention to the MOF biomineralization-based amplification of LSPR biosensor response. In this study, we have employed AuNR as the nanotransducers due to the large tunability and high refractive index sensitivity of the lon-gitudinal LSPR wavelength, and the electromagnetic hot spots at the edges. The AuNR, synthesized through a seed-mediated approach,[16] exhibited a narrow size distribution with a length of 49.5 ± 3.8 nm and a diameter of 12.7 ± 0.8 nm (Figure 2A, inset of Figure 2B). The vis–NIR extinction spectrum of the as-synthesized aqueous AuNR suspension is characterized





by two peaks: one at higher wavelength and the other one at lower wavelength, corresponding to the longitudinal and transverse surface plasmons, respectively (Figure 2B). The longitudinal plasmon band, which exhibits a higher refrac-tive index sensitivity compared to the transverse band, was employed for monitoring the molecular imprinting process and implementing the artificial antibody-based biosensor. Various steps involved in the fabrication and implemen tation of the biosensor are illustrated in Figure 2C. The AuNRs, immo-bilized on a clean glass substrate, are modified with p-ami-nothiophenol (pATP) and glutaraldehyde (GA) (step 1), which serve as linkers to immobilize HSA (model target protein) on the surface of AuNR by forming reversible imine bonds (step 2). Two monomers, namely, (3-aminopropyl) trimethox-ysilane (APTMS) and trimethoxysilane (TMPS), are then copo-lymerized around the immobilized template protein (step 3). The methoxy group of two functional monomers undergoes rapid hydrolysis and subsequent condensation to form an amorphous polymer network, leaving the functional groups (NH3+, OH, CH3) interacting with the template protein through electrostatic, hydrogen bonding, and hydrophobic interactions. Subsequently, the template protein is released from the polymer matrix and AuNR surface by a mixture of oxalic acid and sodium dodecyl sulfate (SDS) solution (step 4), which breaks the imine bond and overcomes the noncovalent

interactions, respectively. The removal of the template pro-tein leaves cavities in the siloxane copolymer that are com-plementary in size, shape, and chemical functionality to the template protein. These cavities serve as artificial antibodies, preferentially capturing template (like) species in the analyte solution (step 5). Following the capture of the target proteins, the biochip is exposed to MOF precursor solution resulting in the biomineralization of MOF around the captured biomol-ecules (step 6).[11] To evaluate the HSA binding and the MOF formation (steps 4–6), we performed AFM imaging of the plasmonic biochip at each stage. After the template removal, the HSA-imprinted AuNR exhibited a smooth surface and the AuNR were found to be uniformly distributed on the sub-strate with no signs of aggregation (Figure 2D). Exposure to HSA solution resulted in the capture of the biomolecules (as evidenced by the LSPR shift discussed below), however, no significant change in the morphology of the AuNR was noted (Figure 2E). In stark contrast, after the exposure of the HSA captured biochip to MOF precursors, the AFM images revealed small globular features, presumably MOF nanocrys-tals, around the AuNR (Figure 2F). Additionally, AFM investi-gation also revealed the preferential growth of the small glob-ules at the two ends of AuNR (inset AFM image in Figure 2F showing an individual AuNR with MOF nanocrystals). In fact, we have demonstrated that during the imprinting process,


Figure 1. AFM height images of HSA absorbed on HOPG substrate A) before (height scale: 15 nm) and B) after incubation in ZIF-8 precursors (height scale: 30 nm) (insets show the schematic illustration of MOF growth around immobilized HSA). C) Height histogram of HSA adsorbed on HOPG obtained from AFM image in Figure 1 A. D) Height histogram of globules formed after incubation in ZIF-8 precursors obtained from AFM image in Figure 1 B. E) LSPR shift versus MOF growth time on HSA modified AuNR and bare AuNR, respectively. F) Raman spectra of HSA before and after exposure to ZIF-8 precursors. G) XRD pattern of HSA coated with ZIF-8 nanocrystals.





Figure 2. A) TEM image of AuNRs employed as nanotransducers for artificial antibody-based LSPR biosensor. B) Vis–NIR extinction spectrum of aqueous suspension of AuNR (inset shows the histogram of the length of AuNR obtained from TEM image, revealing the average length of AuNR ≈49.5 nm). C) Schematic illustration showing the MOF-amplified plasmonic biosensor based on artificial antibodies. Schematic illustration depicts the various steps involved in fabrication of MIP-based plasmonic biosensor, target protein capture, followed by the mineralization of MOF around the captured protein to enhance the LSPR signal. Representative AFM images of AuNR after D) template protein removal, E) recapture of the target protein, and F) MOF biomineralization (insets show individual AuNR, Z scale is 40 nm for all the images).




the molecular linkers (pATP and GA) tend to bind at the ends of AuNR due to the relatively low density of surfactant at the ends compared to the side walls.[4,5] This phenomenon results in site-specific formation of artificial antibody and subsequent capture of the antigen. The captured antigen present at the ends of AuNR in turn serves as the nucleation point for the MOF nanocrystals.

We monitored the LSPR wavelength shift corresponding to each step in the imprinting process, protein capture, and MOF amplification. Cumulative red shift of ≈11 nm was observed after the first three steps of the imprinting process (i.e., formation of cross-linker layer (pATP+GA), protein (HSA) immobilization, polymerization) followed by a blue shift of ≈4.6 nm after the extraction of the template protein (steps 1–4 in Figure 3A, and the corresponding vis–NIR spectra are shown in Figure S4, Supporting Information). Following the molecular imprinting process using HSA as template protein, the plasmonic biochip was exposed to a relatively high concen-tration of HSA (25 µg mL−1), resulting in the binding of HSA to artificial antibodies on AuNR and a consequent LSPR red shift of 3.9 nm (step 5 in Figure 3A). Subsequently, the biochip with the captured analyte was exposed to ZIF-8 precursor solu-tion for 10 min resulting in the growth of MOF nanocrystals on the AuNR. The MOF crystal growth around HSA resulted in an additional red shift of 7.9 nm (step 6 in Figure 3A). The LSPR wavelength shifts after the capture of HSA and after the MOF-based amplification are also represented as steps 4–5 and steps 5–6 in Figure 3B, respectively. The MOF growth resulted in a nearly 200% enhancement (7.9 nm) in the LSPR shift com-pared to the shift from the target protein binding (3.9 nm). It is known that ZIF-8 dissociates under acidic conditions because of the loss of the coordination between the zinc ions and imida-zole.[11,17] As such, one should expect a blue shift in the LSPR wavelength after exposure to acidic buffer. After step 6, the plasmonic chip was exposed to buffer solution at pH 5, which resulted in a blue shift of ≈5.9 nm. Finally, we demonstrate the regrowth (exposure to ZIF-8 precursor solution) and redis-solution (exposure to acidic buffer) of ZIF-8 on the plasmonic biochip as depicted in steps 8 and 9, respectively. In the past, we have demonstrated that one of the significant advantages of the plasmonic biosensors based on artificial antibodies is

their reusability after desorbing the bound target protein using a mixture of sodium dodecyl sulfate (SDS) and oxalic acid.[4] Considering the fact that the MOF could be dissociated under acidic conditions, the amplification process introduced here does not compromise the reusability of this class of biosensors. These results demonstrate the robust and repeatable amplifica-tion of the LSPR shift in a plasmonic biosensor with a simple and rapid MOF growth process.

Next, we attempted to quantify the improvement in sen-sitivity and limit of detection of the artificial antibody-based plasmonic biosensor using MOF-based amplification. The plasmonic biosensor with artificial antibodies specific to HSA was exposed to phosphate buffer (at pH 8) spiked with various concentrations of HSA. The LSPR shift exhibited a monotonic increase with increase in the concentration of HSA. The LSPR shift at the highest concentration tested here (25 µg mL−1) was found to be ≈3.9 nm (Figure 4A). Following exposure to the analyte solution at each concentration, the LSPR shift was amplified by exposing the plasmonic biochip to ZIF-8 precursor solution. The LSPR wavelength exhibited a much larger shift corresponding to each concentration after ampli-fication process (Figure 4A). Figure 4B showed the innate and amplified response of the plasmonic biosensor to low concentrations of HSA. Over this small concentration range, where the plasmonic biosensor exhibited a linear response, the innate sensitivity was calculated to be 0.09 nm nm−1. The sen-sitivity after amplification process (0.16 nm nm−1) was found to be nearly twice compared to the innate sensitivity. Further-more, defining the minimum detectable LSPR shift as 0.5 nm (3σ noise level), the limit of detection of the pristine plasmonic biosensor was found to be ≈400 ng mL−1 whereas the same after amplification process was found to be nearly three times lower (130 ng mL−1).

For efficient deployment and widespread utility in point-of-care and resource-limited settings, the amplification process would need to be simple and rapid. In most previous cases, the amplification process to enhance the LSPR response signifi-cantly prolongs the testing period. We investigated the optimal time for MOF growth-based amplification reaction. The LSPR shift upon exposure to ZIF-8 precursors showed a rapid increase within the first 60 s and essentially plateaued within


Figure 3. A) LSPR shift of AuNR after each step along the fabrication of MIP, target protein capture, MOF growth, MOF dissolution, and subsequent regrowth and dissolution of MOF. B) Extinction spectra exhibiting a red shift in the LSPR wavelength of MIP-AuNR after the capture of HSA and a significant enhancement of the shift after biomineralization.





10 min (Figure 4C). In order to verify whether the saturation in the LSPR shift was due to the depletion of precursors, we immersed the MOF-amplified plasmonic biochip (after 10 min amplification reaction) to fresh precursor solution for an extra 30 min and no significant red shift was observed (Figure S5, Supporting Information). As a negative control, pristine plas-monic biochip (i.e., plasmonic nanostructures with artificial antibodies) was exposed to ZIF-8 precursor solution. The pris-tine biochip did not exhibit significant LSPR shift indicating the absence of MOF crystal growth (Figure 4C). These results indicate the artificial antibodies themselves do not possess suf-ficient affinity to the ZIF-8 precursors to induce prenucleation and growth of ZIF-8 nanocrystals.[18] It is important to note that the absence of MOF growth on artificial antibodies is in stark contrast to natural antibodies, which induces biominer-alization of ZIF-8. In fact, ZIF-8 has been employed as a pro-tective coating to render thermal stability to natural antibodies immobilized on plasmonic nanostructures.[19] Overall, we introduced an ultrafast process to enhance the LSPR response, which outperforms most of the existing LSPR enhancement methods.

Finally, we demonstrate the generalizability of the amplifi-cation process for the detection of various proteins. We have employed lysozyme and hemoglobin as two other model bio-markers. The LSPR shift corresponding to each step along the

imprinting process is shown in Figure S6 (Supporting Infor-mation). The biochips imprinted with lysozyme and hemo-globin were exposed to 20 µg mL−1 of the corresponding pro-tein solutions to capture the target biomolecules, followed by the immersion in ZIF-8 precursors. Both sensors exhibited enhancement in the LSPR shift, although with a small differ-ence in the amplification efficiency, compared to the corre-sponding innate responses (Figure 4D). The difference in the amplification efficiency is probably due to the difference in amino acid composition of the proteins, resulting in a differ-ence in the affinity to ZIF-8 precursors and ability to induce biomineralization of ZIF-8 crystals.

To summarize, we introduced a biomineralization-based signal amplification of refractometric biosensors based on abi-otic recognition elements. To the best of our knowledge, this is the first demonstration of biomolecule-free and label-free LSPR enhancement of a plasmonic biosensor. At the heart of the amplification process lies the abiotic nature of the biorecog-nition elements employed in this class of sensors. The differen-tial affinity of MOF precursors to the artificial antibodies and the captured protein biomarkers and their differential ability to induce biomineralization have been exploited to realize the amplification of LSPR response. The highly specific and ultrafast growth of MOF crystals enhance the LSPR response of the plasmonic biosensors by ≈200% and significantly lower

Figure 4. A) LSPR shift versus concentration of the target protein biomarker (HSA) before and after MOF amplification. B) Concentration-dependent LSPR shift of innate and amplified biosensor at low target protein concentration. C) MOF growth reaction time-dependent LSPR shift on molecularly imprinted AuNR before and after target protein capture. D) MOF amplification of MIP-based plasmonic for three different model protein analytes showing the generality of the signal amplification strategy (results are the mean ± SD, n ≥ 3).





the limit of detection without significantly increasing the assay time. We also demonstrate that this ultrafast and label-free method introduced here could be broadly applied to various protein biomarkers, albeit with small variations in the enhance-ment efficiency.

Supporting InformationSupporting Information is available from the Wiley Online Library or from the author.

AcknowledgementsThe authors acknowledge support from National Science Foundation (CBET1254399 and CBET1512043) and National Institutes of Health (R21DK100759 and R01 CA141521). The authors thank Nano Research Facility (NRF) at Washington University for providing access to electron microscopy facilities.

Conflict of InterestThe authors declare no conflict of interest.

Keywordsartificial antibodies, localized surface plasmon resonance (LSPR), metal–organic framework, molecular imprinting, plasmonic biosensors

Received: January 31, 2017Revised: March 22, 2017

Published online: May 18, 2017

[1] J. N. Anker, W. P. Hall, O. Lyandres, N. C. Shah, J. Zhao, R. P. Van Duyne, Nat. Mater. 2008, 7, 442.

[2] K. L. Kelly, E. Coronado, L. L. Zhao, G. C. Schatz, J. Phys. Chem. B 2003, 107, 668.

[3] a) H. Chen, X. Kou, Z. Yang, W. Ni, J. Wang, Langmuir 2008, 24, 5233; b) K. M. Mayer, J. H. Hafner, Chem. Rev. 2011, 111, 3828; c) B. Sepúlveda, P. C. Angelomé, L. M. Lechuga, L. M. Liz-Marzán, Nano Today 2009, 4, 244.

[4] A. Abbas, L. Tian, J. J. Morrissey, E. D. Kharasch, S. Singamaneni, Adv. Funct. Mater. 2013, 23, 1789.

[5] L. Tian, K.-K. Liu, J. J. Morrissey, N. Gandra, E. D. Kharasch, S. Singamaneni, J. Mater. Chem. B 2014, 2, 167.

[6] L. Guo, J. A. Jackman, H.-H. Yang, P. Chen, N.-J. Cho, D.-H. Kim, Nano Today 2015, 10, 213.

[7] A. J. Haes, L. Chang, W. L. Klein, R. P. Van Duyne, J. Am. Chem. Soc. 2005, 127, 2264.

[8] W. P. Hall, S. N. Ngatia, R. P. Van Duyne, J. Phys. Chem. C 2011, 115, 1410.

[9] O. M. Yaghi, M. O’Keeffe, N. W. Ockwig, H. K. Chae, M. Eddaoudi, J. Kim, Nature 2003, 423, 705.

[10] a) H. Furukawa, K. E. Cordova, M. O’Keeffe, O. M. Yaghi, Science 2013, 341, 974; b) M. L. Foo, R. Matsuda, S. Kitagawa, Chem. Mater. 2014, 26, 310; c) G. Lu, O. K. Farha, W. N. Zhang, F. W. Huo, J. T. Hupp, Adv. Mater. 2012, 24, 3970.

[11] K. Liang, R. Ricco, C. M. Doherty, M. J. Styles, S. Bell, N. Kirby, S. Mudie, D. Haylock, A. J. Hill, C. J. Doonan, Nat. Commun. 2015, 6, 7240.

[12] K. Liang, C. Carbonell, M. J. Styles, R. Ricco, J. Cui, J. J. Richardson, D. Maspoch, F. Caruso, P. Falcaro, Adv. Mater. 2015, 27, 7293.

[13] M. Dąbkowska, Z. Adamczyk, M. Kujda, Colloids Surf., B 2013, 101, 442.

[14] G. Kumari, K. Jayaramulu, T. K. Maji, C. Narayana, J. Phys. Chem. A 2013, 117, 11006.

[15] S. Li, W. Shi, G. Lu, S. Li, S. C. J. Loo, F. Huo, Adv. Mater. 2012, 24, 5954.

[16] K.-S. Lee, M. A. El-Sayed, J. Phys. Chem. B 2005, 109, 20331.[17] J. Zhuang, C.-H. Kuo, L.-Y. Chou, D.-Y. Liu, E. Weerapana,

C.-K. Tsung, ACS Nano 2014, 8, 2812.[18] E. T. Hwang, R. Tatavarty, J. Chung, M. B. Gu, ACS Appl. Mater.

Interfaces 2013, 5, 532.[19] C. Wang, S. Tadepalli, J. Luan, K. K. Liu, J. J. Morrissey,

E. D. Kharasch, R. R. Naik, S. Singamaneni, Adv. Mater. 2016, 29, 1604433.

本文献由“学霸图书馆-文献云下载”收集自网络，仅供学习交流使用。

学霸图书馆（www.xuebalib.com）是一个“整合众多图书馆数据库资源，

提供一站式文献检索和下载服务”的24 小时在线不限IP

图书馆。

图书馆致力于便利、促进学习与科研，提供最强文献下载服务。

图书馆导航：

图书馆首页文献云下载图书馆入口外文数据库大全疑难文献辅助工具

http://www.xuebalib.com/cloud/http://www.xuebalib.com/http://www.xuebalib.com/cloud/http://www.xuebalib.com/http://www.xuebalib.com/vip.htmlhttp://www.xuebalib.com/db.phphttp://www.xuebalib.com/zixun/2014-08-15/44.htmlhttp://www.xuebalib.com/

amplification of refractometric biosensor response through...

Documents