Self-Organized Frameworks on Textiles (SOFT): Conductive Fabrics for Simultaneous Sensing, Capture, and Filtration of Gases

Article abstract of DOI:10.1021/jacs.7b08840

Wearable electronics have the potential to advance personalized health care, alleviate disability, enhance communication, and improve homeland security. Development of multifunctional electronic textiles (e-textiles) with the capacity to interact with the local environment is a promising strategy for achieving electronic transduction of physical and chemical information. This paper describes a simple and rapid approach for fabricating multifunctional e-textiles by integrating conductive two-dimensional (2D) metal-organic frameworks (MOFs) into fabrics through direct solution-phase self-assembly from simple molecular building blocks. These e-textiles display reliable conductivity, enhanced porosity, flexibility, and stability to washing. The functional utility of these integrated systems is demonstrated in the context of chemiresistive gas sensing, uptake, and filtration. The self-organized frameworks on textiles (SOFT)-devices detect and differentiate important gaseous analytes (NO, H₂S, and H₂O) at ppm levels and maintain their chemiresistive function in the presence of humidity (5000 ppm, 18% RH). With sub-ppm theoretical limits of detection (LOD for NO = 0.16 ppm and for H₂S = 0.23 ppm), these constitute the best textile-supported H₂S and NO detectors reported and the best MOF-based chemiresistive sensors for these analytes. In addition to sensing, these devices are capable of capturing and filtering analytes.

Full text of DOI:10.1021/jacs.7b08840

Article
pubs.acs.org/JACS
Cite This: J. Am. Chem. Soc. XXXX, XXX, XXX-XXX
Self-Organized Frameworks on Textiles (SOFT): Conductive Fabrics
for Simultaneous Sensing, Capture, and Filtration of Gases
Merry K. Smith and Katherine A. Mirica*
Department of Chemistry, Burke Laboratory, Dartmouth College, Hanover, New Hampshire 03755, United States
S
* Supporting Information
ABSTRACT: Wearable electronics have the potential to advance
personalized health care, alleviate disability, enhance communication,
and improve homeland security. Development of multifunctional
electronic textiles (e-textiles) with the capacity to interact with the
local environment is a promising strategy for achieving electronic
transduction of physical and chemical information. This paper describes a
simple and rapid approach for fabricating multifunctional e-textiles by
integrating conductive two-dimensional (2D) metal−organic frameworks
(MOFs) into fabrics through direct solution-phase self-assembly from
simple molecular building blocks. These e-textiles display reliable
conductivity, enhanced porosity, flexibility, and stability to washing.
The functional utility of these integrated systems is demonstrated in the context of chemiresistive gas sensing, uptake, and
filtration. The self-organized frameworks on textiles (SOFT)-devices detect and differentiate important gaseous analytes (NO,
H2S, and H2O) at ppm levels and maintain their chemiresistive function in the presence of humidity (5000 ppm, 18% RH). With
sub-ppm theoretical limits of detection (LOD for NO = 0.16 ppm and for H2S = 0.23 ppm), these constitute the best textile-
supported H2S and NO detectors reported and the best MOF-based chemiresistive sensors for these analytes. In addition to
sensing, these devices are capable of capturing and filtering analytes.
afforded by the diversity of available ligands and nodes can
impart rationally designed functionality into the assembled
framework at the monomer stage.36−38 Integration of
conductive MOFs into textiles to create SFSs that synergisti-
cally integrate the desirable features of conductive MOFs
(electronic response to stimuli and high surface area) with the
porosity and flexibility of the textile substrate can greatly
enhance the functional utility of e-textiles.
While integration of conductive MOFs into textiles for the
purposes of developing functional electronic sensors has not
been previously established, interfacing nonconductive MOFs
with fibers and fabrics,39−44 polymers,45,46 nanoparticles,47
biomaterials,48,49 and chemically modified supports41,45 is well
documented.50−52 Methods of solvothermal growth,50−52
liquid-phase epitaxy,50−52 spray-coating,42 film casting,51
postsynthetic attachment,40,51 electrochemical deposition,51,53
atomic layer deposition,46 and direct precipitation43,51 can
produce hybrid systems with enhanced function in selective
chemical separation, sequestration, and filtration,50−52,54−56
resistance to biofouling,57 stabilization of biomolecules,48,49 and
controlled release of catalysts.41 Moreover, introduction of
lithographic58−62 and inkjet printing techniques63 can create
patterns of surface-attached MOFs with controlled lateral
resolution with potential utility in solid-state devices. Many of
these systems, however, are not conductive and consequently
INTRODUCTION
■
Wearable electronics1−3 hold promise in enhancing health
monitoring,4,5 alleviating disability,6 and tracking of environ-
mental pollution.7 Electronic textiles (e-textiles) have the
potential to advance the impact of wearable electronics through
breathable devices and garments capable of electronically
transduced interactions with the local environment.8−10 This
electronic signal transduction can be harnessed to create Smart
Fabric Sensors (SFSs)11 that report applied pressure,12 strain,13
heart rate,14 or presence of a chemical.15,16 While physical SFSs
typically rely on the flexibility and tunability of the fabric
platform to enhance device performance,11−17 chemical SFSs
often employ fiber-integrated stimuli-responsive materials (e.g.,
conductive polymers, carbon nanotubes, graphene),15,16,18,19
that either do not affect or diminish the sensing response
compared to other substrates.17,18 A movement toward
developing SFS technology based on synergy rather than
compromise  wherein the sensor performance is enhanced
by the textile platform  is an essential step in expanding the
scope and practical viability of wearable chemical sensors.18−21
An emerging class of conductive multifunctional nanoma-
terials based on metal−organic frameworks (MOFs) holds
remarkable potential in electronic applications22−25 as chemir-
esistors,26−30 supercapacitors,31,32 and electrocatalysts.33−35
These porous, d−π conjugated, crystalline frameworks can be
prepared through modular bottom-up self-assembly of organic
ligands with metallic nodes to generate extended conductive
solid-state materials.23 The structural and functional tunability
Received: August 18, 2017
© XXXX American Chemical Society
A
DOI: 10.1021/jacs.7b08840
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
are incapable of direct electronic transduction of their function.
Inspired by these performance metrics with nonconductive
MOFs in other areas, we reasoned that developing effective
methods for integrating conductive MOFs with textiles can lead
to SFSs with enhanced performance in gas sensing and
filtration.
This paper describes a one-step fabrication of flexible, textile-
supported devices based on conductive MOFs (Figure
1A,B).64,65 The direct self-assembly of 2,3,6,7,10,11-hexahy-
droxytriphenylene (HHTP) or 2,3,6,7,10,11-hexaaminotriphe-
nylene (HATP) organic ligands with metallic nickel(II) nodes
from solution onto natural (cotton) and synthetic (polyester)
fabrics produces textiles that are conformally coated with
conductive MOF crystallites at the fiber-level: Ni3(2,3,6,7,10,-
11-hexahydroxytriphenylene)2 or Ni3(2,3,6,7,10,11-hexaimino-
triphenylene)2 (Ni3HHTP2 or Ni3HITP2), respectively (Figure
1C).64,65 To the best of our knowledge, this work constitutes
the first use of direct self-assembly to deposit conductive MOFs
on textiles.
The MOF-based e-textiles reported herein display at least
seven unique characteristics for functional materials design: (i)
They are the first example of using conductive MOFs as
functional components in e-textiles. (ii) They exhibit enhanced
porosity by combining mesoporosity of the textile and
microporosity of the conductive MOF. (iii) They represent
the first demonstration of simultaneous detection, capture,
preconcentration, and filtration of gases in MOF-based
chemiresistive device architectures. (iv) They are flexible,
washable (in water and acetone), and stable to heat. (v) They
retain their functional integrity, conductivity, and gas sensing
capability in the presence of humidity (at least 5000 ppm of
H2O). (vi) These systems show that ligand specific differences
(HITP vs HHTP) in MOF design can be useful attributes for
controlling the unique chemiresistive response of these
materials toward analytes. (vii) In terms of magnitude of
chemiresistive response, self-organized frameworks on textiles
(SOFT)-devices constitute the best SFS NO and H2S detectors
reported and the best MOF-based chemiresistive NO and H2S
sensors with theoretical limits of detection (LOD) of 0.16 ppm
for NO and 0.23 ppm for H2S.
This technology fundamentally advances functional materials
design, MOF-based electronics, and methods for integrating
electrically conductive stimuli-responsive materials into solid-
state devices. We believe that these characteristics are of value
for the researchers interested in bottom-up design and
synthesis of stimuli-responsive materials, the design of wearable
electronics with enhanced capabilities in chemical detection,
personal protection, and gas sequestration.
RESULTS AND DISCUSSION
■
Introducing continuous conductivity while maintaining the
functional integrity of the textile platform requires integration
of conductive materials at the fiber level.9 Multifunctional
conductive nanomaterials have served as productive targets for
developing SFSs,8−18 but their deposition strategies are often
limited in continuous fiber-scale integration or require
specialized instrumentation. A strategy that maintains the
simplicity of bottom-up deposition while achieving conformal
fiber-scale integration of conductive sensor materials, such as
direct self-assembly from molecular components, is highly
desirable.
We chose two MOFs with high bulk electrical conductivities
(0.01−2.0 S/cm):64,65 metal catecholate and imino cross-linked
frameworks, Ni3HHTP2 and Ni3HITP2, respectively. The
reactive starting materials readily assemble into conductive
d−π conjugated scaffolds in solvothermal reaction conditions
(Figure 1B). We adapted procedures from our earlier report
detailing the direct assembly of conductive MOFs onto
chemiresistive devices on the surface of polymeric films
equipped with carbon-based electrodes.29 This paper demon-
strates that the method of direct assembly can be generalized to
achieve the integration of conductive MOFs into textiles with
unique SOFT-device architecture.
Figure 1. Fabrication of MOF devices and sheet resistance of SOFT-
sensor devices. (A) Starting materials for SOFT-sensors from organic
triphenylene-based ligand (HHTP or HATP), metallic node (Ni-
(OAc)2 or NiCl2), and fabric swatch (cotton shows no conductivity).
Photograph of cotton swatch is shown. (B) Solvothermal con-
densation of these reagents in water produces conductive SOFT-
devices: textiles coated with nanoporous MOF. Macroscopic through
molecular level detail is shown, from a photograph of cotton SOFT-
sensor postreaction (top left), to scanning electron micrographs
detailing MOF coating on fibers (bottom left), and characteristic MOF
nanorod texture (bottom right), to space-filling model of MOF (top
right). (C) Sheet resistance measurements for SOFT-swatches.
Fabrication of Devices. We chose textiles for the
immobilization of functional MOF-based materials because
fabrics are chemically modifiable (as shown by dyeing
processes),66 physically tunable (roughness and thickness),67
B
DOI: 10.1021/jacs.7b08840
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
chemically diverse (natural and synthetic),67 often low-cost and
renewable, and compatible with straightforward integration
directly into the existing infrastructure of cloth garments,
coverings, and commercial goods.8−10,67,68 To prepare SOFT-
devices, commercial textiles were cut into swatches (1 cm × 2
cm). We focused our efforts on natural woven cotton fabric
(0.243 mm thick), due to the precedent for cotton serving as an
effective platform for e-textiles.8−11 To display the generality of
the method, we also tested synthetic nonwoven polyester
batting (0.643 mm thick, compressed) and found these devices
to perform indistinguishably from the cotton-based SOFT-
devices (see sections SVIII and SXIII).
To achieve the assembly of MOFs on fabrics, we added solid
molecular precursors (organic ligand and metallic node)
directly to a vial containing the fabric swatch (one swatch per
vial), and added water such that all reagents and textiles were
fully dampened (0.033 M with respect to triphenylene ligand).
For Ni3HITP2, ammonium hydroxide (6 equiv with respect to
HITP) was also added to the reaction vessel.65 The mixtures
were subjected to sonication (5 min) and stirred overnight at
80 °C. We allowed the resulting SOFT swatches to cool in the
reaction mixture, then isolated, washed thoroughly with water
(1 × 2 mL) and acetone (1 × 2 mL), and drip dried the devices
in air (18 h).
(section SIII). The crystallization of MOF coating did not
compromise the thermal stability of the cotton substrate, which
is similarly stable up to 350 °C under nitrogen (section SIII). A
mass loss around 100 °C corresponding to the loss of trapped
or bound water commonly observed in porous materials. A loss
of only 1% mass was observed for both devices in this range,
suggesting that devices were fully dry after the process of
washing and drip drying. These results suggest that
conductivity results from the MOF itself, not from trapped
water.
Enhanced Porosity. We used Brunauer−Emmett−Teller
(BET) measurements to gain insight into the porosity of the
SOFT-devices as compared to their MOF and cotton
constituents. For Ni3HHTP2, the BET surface area for the
bulk MOF was found to equal 421 m2/g in argon (77 K), in
excellent agreement with reported values (425 m2/g).64
Integration of MOFs into SOFT-devices increased the BET
surface area from 0.3 m2/g for bare cotton to 3.6 m2/g for
SOFT-devices, an order of magnitude increase in porosity
(section SVII). If all MOF pore openings remained unblocked,
and full internal surface area of the MOFs on textiles was
available, the BET surface area would be equal to approximately
50 m2/g. The nature of the templation likely caused a certain
percentage of MOF surface to be unavailable (the surface that
templates on the fiber), leading to a deviation from the
theoretical maximum surface area. Ni3HITP2 exhibited similar
trends, with a surface area of 340 m2/g for pure MOF and a
surface area equal to 6.8 m2/g for the SOFT-device (argon: 77
K).
SOFT-devices comprised an average of 7−8% MOF (by
weight), corresponding to 4−6 mg of MOF per cm3 of textile
(see Supporting Information section SII for full experimental
details). Textile-supported MOF constituted 16−30% of the
total mass of MOF prepared (section SX), taking into account
both the SOFT-device and the precipitated MOF powder
isolated from the reaction solution. The characterization of bulk
MOF powder samples matched reported analysis (See sections
SII−SVI in the Supporting Information).64,65
Functional Performance and Yield. Conductive MOF-
coated textiles were classified as functional in this study if they
exhibited measurable resistance when probed with a multimeter
over a 1 cm distance once fully dried. The crystallization of
conformal MOF coatings was found to be an essential feature
for the electrical and chemiresistive performance. Control
experiments that involved soaking fabric swatches in solutions
containing individual starting materials, or repeatedly dip-
coating swatches (see section SXI) in suspensions of
presynthesized MOFs in varied solvents did not generate a
conductive pathway, suggesting that direct crystallization is
essential for SOFT-fabrication. As a further control, we soaked
swatches of fabric substrate in solutions containing both
starting materials, as we would to produce SOFT-devices, but
prevented the large-scale crystallization of MOF by limiting the
reaction time and keeping the samples at room temperature (5
min soaking with 1 min sonication: see sections SIV for XRD,
SV for SEM, and SVI for EDX). This method did not produce a
conductive pathway in fabric swatches, implying that
crystallization of MOF is required for SOFT fabrication.
Direct solution-phase crystallization of MOFs was capable of
generating a large surface area of conductivity, up to 25 cm2 of
conductive surface (measured diagonally across a 5 cm × 5 cm
swatch, largest area tested) for Ni3HHTP2 (see section SX).
This fabrication method yielded functional swatches with
Characterization of Devices. To characterize the
successful integration of MOFs into textiles, bulk MOF
(isolated from the SOFT-device reaction solution) was
compared to MOF incorporated into devices using powder
X-ray diffraction (PXRD: section SIV), scanning electron
microscopy (SEM: section SV), and energy-dispersive X-ray
spectroscopy (EDS: section SVI). Key PXRD peaks corre-
sponding to Bravais planes oriented perpendicular to the 2D
plane of the nanomaterial were present for both bulk and
SOFT-samples, suggesting long-range “honeycombed” order in
a slipped parallel packing motif, as expected.64,65 The PXRD
traces for the MOF-coated fabrics, however, were convoluted
by the fabric substrate, which presents a series of sharp peaks
and broad amorphous regions, and while suggestive of MOF
presence, were not conclusive. Identity and morphology of the
MOFs was confirmed by EDS and SEM (Figure 1C and
sections SV and SVI). Electron micrographs of the MOF-
coated fabrics indicated tightly packed MOF-nanorods forming
a continuous coating over each fiber. Cross sections of the
SOFT-devices indicated conformal coating of the textile by the
MOF (see Supporting Information SV). If the swatches
remained fully submerged in the reaction solution throughout
assembly, the conformity of the MOF assembly was uniform
throughout the swatch. Swatches that stuck to the sides of the
vial, or remained partially submerged were considered non-
functional devices, even if the submerged portion displayed
conductivity. The elemental composition of the MOF-nanorods
matched bulk MOF by EDS (section SVI).
moderate sheet resistance (Ni3HHTP2 2.8
Ni3HITP2 5.6
0.5 MΩ/cm2,
2 MΩ/cm2, Figure 1C) and high yields of
functional devices. For each device, we performed a minimum
of n = 10 fabrication trials, yielding 94% functional devices for
Ni3HHTP2 on cotton (n = 30), 87% for Ni3HHTP2 on batting
(n = 13), and 96% for Ni3HITP2 on cotton (n = 23). The small
number of nonfunctional devices either failed to display
conductivity or did not have a uniform distribution of MOF
across the swatch.
Thermal gravimetric analysis (TGA) of SOFT-devices
showed both devices to be stable up to 350 °C under nitrogen
C
DOI: 10.1021/jacs.7b08840
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Two-point conductivity for cotton SOFT-sensors was
calculated20 to be 1.6 × 10−4 S/cm for Ni3HHTP2 and 2.6 ×
10−3 S/cm for Ni3HITP2. The devices were semiconductive
within their usable temperature range (room temperature to
∼80 °C, see section SX), as evidenced by their linear increase in
conductivity with increasing temperature. Gold-coated pins
exhibited Ohmic contacts with the SOFT-sensors in the range
from −2.0 to 2.0 V at 0.1 V intervals (see Supporting
Information SX).69
Mechanical Stability. The SOFT-devices retained the
physical properties of the textile and were capable of being
bent, twisted, and wrapped around curved surfaces without a
significant perturbation or loss of conductivity (Figure 2A).
the resistance greatly: a swatch stretched from 4 to 5 cm (25%
increase) showed a resistance change of 0.06−0.05 MΩ/cm2,
respectively (see section SX). The SOFT-devices were stable to
washing with water or with organic solvent (acetone) and
resistant to disassembly through mechanical handling.
When cut and punctured (e.g., sewing), SOFT-devices lost
up to 20% of their mass via fraying of the fabric and some MOF
shedding but still retained their conductivity, with no significant
change in sheet resistance or chemiresistive performance. Most
mass lost was attributable to fraying of the fabric, as the mass of
the MOF comprised only 7−8% of the full mass of the device.
Fraying occurred primarily during cutting: when the devices
were prepared as large swatches (4.5−25 cm2) and then cut
into small sections for testing of device properties (i.e., sensing
capabilities).
Directed Patterning of Conductive Areas within
Devices. The strategy for incorporating conductive nanoma-
terials into e-textiles through self-assembly makes it possible to
fabricate unique device architectures that would be otherwise
difficult to achieve. The sensor fabrication was tolerant of an
assortment of prepatterned architectures: MOF assembly
proceeded on top of and between existing electrodes, including
sewn-in conductive threads and painted silver electrodes.
Generating patterns of the MOFs on fabrics was accomplished
by masking specific areas of the fabric with a hydrophobic
barrier (e.g., petroleum jelly). Loading a syringe with petroleum
jelly enabled manual patterning of this hydrophobic barrier
directly onto the surface of the fabric with reasonable control
( 0.5 cm). Gentle heating (40−50 °C) melted the jelly into
the bulk of the textile, and masked the patterned area from
MOF assembly in aqueous solution. Isolation of the swatch
from the reaction mixture and washing with water, acetone, and
with petroleum ether removed any residual petroleum jelly
yielding a clean, patterned SOFT-device (section SXII). This
approach enabled the preparation of a range of device
architectures, including arrays of SOFT-devices mounted on a
single swatch, equipped with prepatterned (or sewn) electro-
des.
Assessment of SOFT-Devices as Chemiresistive
Sensors. We demonstrated the multifunctional utility of
SOFT devices in the context of chemiresistive sensing of
toxic gases, as well as capture, filtration, and preconcentration
of gaseous analytes. The sensing performance of the SOFT
devices was tested in the configuration of a chemiresistor
against toxic gaseous analytes with low permissible exposure
limits (PELs): H2S (PEL = 10 ppm) and NO (PEL = 25 ppm).
As target analytes, H2S and NO are also relevant for breath
analysis, as their relative concentrations in exhalate have
implications in endogenous processes related to disease.70
Furthermore, H2S and NO are common toxic emissions from
human industry and considered dangerous pollutants.71 NO,
produced primarily in vehicle exhaust, is a reactive compound
that contributes to the formation of ozone and smog.71,72 H2S
is a corrosive poison released from refineries and animal farms
that is acutely dangerous to aquatic life.71 As such, development
of materials and devices capable of maximizing the real-time
detection, uptake, or controlled release of NO and H2S is
broadly important in human health and safety,73 environmental
remediation,71,72 and health monitoring.70 As controls, we
tested the response of the SOFT devices to water and a range
of organic vapors (section SXIII).
Figure 2. Selected mechanical and electronic properties of SOFT-
sensors. (A) Photographs of mechanical manipulation of as-
synthesized SOFT-devices (4 cm × 1 cm cotton swatch shown for
clarity). SOFT-devices are capable of withstanding twisting, bending,
and wrapping around objects without significant changes in resistance,
as demonstrated in the plots in parts B and C. (B) Sheet resistance
values of cotton SOFT-swatch (2 cm × 1 cm) upon mechanical
manipulation for Ni3HHTP2 and (C) Ni3HITP2, illustrating their
consistent electronic performance under physical stress.
Gentle handling (i.e., installing into custom enclosures,
bending) did not cause measurable mass loss of the MOF.
Repeated deformation did not impact performance: devices
bent or folded up to and beyond 10 cycles (typical handling for
multiple sensing experiments) retained consistent conductivity.
The direct bottom-up integration process allows for the
coverage of large areas of fabric in different aspect ratios.
Squares of material up to 25 cm2 displayed conductivity
throughout the SOFT-device, including along the diagonal
(∼5.5 cm). The swatches were cut along the grain, and
consequently could be stretched along the diagonal bias. For
the large SOFT swatch, stretching deformation did not impact
General Procedure for Chemiresistive Measurements.
SOFT swatches were placed into custom Teflon enclosure
D
DOI: 10.1021/jacs.7b08840
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
equipped with spring-loaded gold pins (serving as electrodes
and immobilizing textile swatches) and gas inlet and outlet
ports (Figure 3A) and sealed tightly for leak-free gas flow.
decreased with analyte exposure for Ni3HHTP2 (saturation =
−49% 10%, n = 3), and increased for Ni3HITP2 (saturation =
81%
6%, n = 3), demonstrating the importance of the
identity of the heteroatom in the framework cross-linker on
sensing ability. When exposed to H2S, an increase in resistance
was observed for both Ni3HHTP2 (saturation = 98% 9%, n =
3) and Ni3HITP2 (saturation = 97% 2%, n = 3). The density
of MOF crystallites in the SOFT-sensors, coupled to the
intrinsic porosity of the fabric substrate, likely contributed to
enhancing the response sensitivity toward the analyte as
compared to previous reports.29 These dosimetric responses
were highly reproducible, as evidenced by repeated saturation
studies (see section SXIII). For both Ni3HHTP2 and
Ni3HITP2, exposure to H2S demonstrated >10% variance
between batches (80 ppm, n = 3, see section SXIII).
The SOFT-sensors were capable of differentiating certain
analyte concentrations semiquantitatively in the first 5 min of
exposure (Figure 3C). Plotting the slope (m) of the device
trace over the first 5 min of exposure versus the concentration
of analyte (5−80 ppm range) provided a linear response for
Ni3HHTP2 toward decreasing concentrations of NO, and for
both devices toward H2S (summarized in Figure 3C). For
Ni3HITP2 SOFT-devices exposed to NO, concentrations as low
as 10 ppm saturated SOFT-devices at the same rate as higher
concentrations.
Detection Limits in N2. The responses of Ni3HITP2
SOFT-devices exposed to NO suggested a low theoretical
limit of detection (LOD). We determined the theoretical LODs
for each analyte/SOFT-device pair (for full details on
calculation, see section SXIII) by exploiting the maximum
percent response (within 1 h of exposure) versus concentration
of analyte.74 The theoretical LODs of Ni3HITP2 and
Ni3HHTP2 toward NO (0.1−80 ppm) were 0.16 and 1.4
ppm, and those toward H2S (1−80 ppm) were 0.52 and 0.23
ppm, respectively. For NO, this constitutes one of the best
MOF-based sensors.75 For H2S, Eddaoudi and co-workers
reported a MOF-based capacitive sensor with an LOD of 5.4
ppb, orders of magnitude lower than our SOFT-devices.76
Their system showed a high selectivity for H2S, while ours is
capable of detecting and differentiating NO from H2S with
excellent LODs.
Figure 3. Chemiresistive device performance for SOFT-sensors. (A)
Custom enclosure for dosing SOFT-sensors with analytes. (B)
Representative response for Ni3HITP2 (blue) and Ni3HHTP2 (red)
SOFT-sensors when exposed to (from left to right) NO or H2S. (C)
Slope (m) of first 5 min of sensor response versus concentration of
analyte, in response to NO (left) and H2S (right).
Chemiresistive Responses of SOFT-Sensors in the
Presence of Water. The capacity to differentiate the signal
corresponding to H2O from relevant analyte signal is essential
for practical applicability. Dosing devices with concentrations of
H2O vapor much higher than the concentrations of analytes (at
least 5000 ppm, ∼18% relative humidity) caused a decrease in
Constant voltage was applied across the electrodes using a
portable potentiostat (1.0 V). The current was monitored
across the exposure (through saturation) and recovery (10
min) cycles of analyte dosing. Gaseous analytes were delivered
from custom-ordered tanks containing each analyte at 1%
concentration balanced with N2, were further diluted in dry
nitrogen (N2), and regulated using a system of dual mass flow
controllers.
Characterization of the Sensitivity and Dynamic
Range of SOFT-Sensors in N2. SOFT devices exhibited
dosimetric responses to NO and H2S, detailed in Figure 3B,C
and section SXIII, that were contingent upon MOF identity.
Percent response is shown in terms of normalized conductance
(−ΔG/Go, see section SXIII for equation), and compared for
SOFT-sensors (1.5 cm × 0.5 cm). While MOFs employing the
HHTP organic ligand have been previously used to sense these
analytes,29 the study of the chemiresistive responses of HITP-
based MOFs to these gases is has not been previously
examined.27,65 Representative saturation traces for NO and
H2S are shown in Figure 3B. For NO, SOFT-device response
response for Ni3HITP2 (−8%
2%) and an increase for
Ni3HHTP2 (26% 1%). Principle component analysis (PCA)
enabled the straightforward differentiation of sensor response
toward H2O from sensor response toward NO and H2S (both
at 80 ppm diluted in dry N2) for arrays featuring Ni3HHTP2
and Ni3HITP2 SOFT-sensors at saturation (Figure 4A).
Encouraged by these results, we tested the ability of the
SOFT-devices to detect and differentiate NO and H2S in a
humid environment (18% relative humidity, 5000 ppm).
Mixing a humid vapor stream delivered by gas generator with
the diluted gaseous stream delivered by mass flow controller
upstream of the sensing apparatus produced a humidified
analyte stream with controlled concentrations of gas (80 ppm)
and water vapor (5000 ppm). Allowing the devices to
equilibrate in the presence of H2O and then dosing with
analyte produced signal indistinguishable (within error) from
E
DOI: 10.1021/jacs.7b08840
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
H2O for 5 min), followed by complete drying in air (12 h) fully
restored the function of the devices within error (see section
SXIV). Heating the swatches was less effective: after exposure
to H2S, gentle heating of the swatch in air (70 °C, 1 h) partially
restored the performance of the sensors to >60% recovery for
Ni3HHTP2 and >70% for Ni3HITP2 (see section SXIV).
The nature of the observed recovery by washing suggests that
the host−guest interaction of metal to analyte is consistent with
a Lewis acid binding site at the metal center. It is possible that
axial water ligands bound to the metal center (shown to be
present in an isoreticular MOF crystal structure)64 are displaced
by more favorable binding with incoming analyte molecules
upon analyte dosing. Displacement of adsorbed oxygen on the
surface is also plausible. The electron donation provided by the
bidentate chelation of catecholate or imino moieties to the
metal center is known to increase the lability of the interaction
between axial ligands and the metal center in octahedral Ni(II)
complexes.77 We hypothesize that soaking of the analyte-
saturated device in water can release the bound analyte through
ligand displacement in water and restore the sensing ability of
the SOFT-device.
SOFT-devices also displayed very good shelf lives. After
standing exposed to ambient air for over a month (50 days),
the resistance of the devices increased by 32% (average sheet
resistance L = 5 cm, fresh = 0.81 MΩ/cm2, aged = 1.1 MΩ/
cm2). However, the chemiresistive sensing performances of the
devices did not diminish with time, in spite of the increase in
resistance (See section SXIII).
Figure 4. Performance of SOFT-devices in competition with water.
(A) Principle component analysis for arrays of (one each) Ni3HHTP2
and Ni3HITP2 SOFT-sensors (triangle and diamond each represent
single array) at saturation, showing capability for differentiating NO,
H2S, and H2O. (B) Summarized saturation sensor response for
Ni3HHTP2 (red) and Ni3HITP2 (blue) exposed to analytes (80 ppm
of NO and H2S, left to right) in dry nitrogen (solid bars) and in the
presence of 5000 ppm water (with water droplet), demonstrating
consistent function between dry and humid (18% RH) environments.
Summarized Fundamental Insight and Performance
Advances from Chemiresistive Devices. The sensing
results summarized in Figures 3 and 4 highlight the novelty
in the performances of these devices and shed light on
fundamental advances in understanding the charge transport
properties and host/guest interactions of conductive MOFs.
Important differences in the direction of the chemiresistive
response were observed for Ni3HHTP2 and Ni3HITP2 in
response to NO: a critical observation in elucidating ligand-
specific differences in the fundamental sensing mechanisms for
charge perturbation in structurally similar MOFs. We propose
that the chemiresistive responses for devices equilibrated in dry
and humid N2 suggest that Ni3HHTP2 behaves as a p-type
semiconductor, while Ni3HITP2 could be a mixed-type (p/n)
semiconductor.78 However, the charge transport mechanisms of
these materials are highly complex, and ligand-specific differ-
ences in MOF structure, defects, host−guest interactions, and
charge transfer kinetics may exist. It is possible that the ligand-
specific responses we observe are consequences of these
differences.
dosing in dry N2 (Figure 4B, section SXIII). Many
chemiresistors can be sensitive or unstable to the presence of
water. However, in one example, a related MOF-based
chemiresistor isoreticular to those used in our SOFT-sensors
(Cu3HITP2) was shown to detect NH3 in humidity levels up to
60% RH.28 This work is the first example of using Ni3HITP2
and Ni3HHTP2 to detect NO and H2S in the presence of
humidity, and our results appear consistent with the promising
precedent for related systems. Due to limitations of the gas
delivery apparatus (see section SXIII for details), sensing of
analytes in environments with higher humidity was not
examined.
In terms of performance, the magnitude of device responses
greatly exceeds the best previously reported results for this class
of materials.29 Moreover, the direction of device response
allows for facile differentiation of analytes, and finally, the
capacity for the devices to detect analytes in the presence of
H2O is highly desirable,79 since many chemiresistive devices are
either sensitive to H2O80 or sensitive to pretreatment
conditions in the presence of H2O.81
Analyte Breakthrough in Flow-Through SOFT-Devi-
ces. The intrinsic porosity of cloth is one of the notable
advantages to utilizing fabric as a substrate in e-textiles.8,9 The
capacity for chemical adsorption of gases hierarchically within
the fabric and the integrated porous material42,82 offers an
opportunity for preconcentration of analytes and simultaneous
filtering and detection of gases. Related flexible systems, such as
Influence of Applied Voltage on Device Performance.
The chemiresistive performance of devices was not affected by
applied voltage: SOFT-devices performed indistinguishably at
1.0 V versus 0.1 V applied voltage (section SXIV). Since the
measurements of current resulted in reliably and conveniently
measurable numbers (∼0.5−1.0 μA for Ni3HHTP2 and 0.1−1.0
mA for Ni3HITP2) when 1.0 V was applied, chemiresistive
sensing experiments were performed at an applied voltage of
1.0 V. Devices maintained their Ohmic character in the range of
−2.0 to 2.0 V while saturated with gaseous analytes (section
SXIV).
Device Recovery and Long-Term Stability. Despite
exhibiting dosimetric responses toward NO and H2S, the
SOFT-devices were fully recoverable postsaturation by washing.
A simple wash step with water (suspending devices in deionized
F
DOI: 10.1021/jacs.7b08840
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
mixed matrix membranes (MMMs)83,84 or hybrid MOF/
polymer composites,52 are effective functional components in
gas sequestration and separations.50−52,54−56 These systems,
however, have not yet been developed into effective, electrically
accessible sensors that detect capture and breakthrough. Since
porous MOF-based membranes are widely known as functional
systems capable of gas separation and sequestration,85 our
SOFT-devices are excellent candidates for integration into an
electronic sensor with filtration properties.
second membrane (SOFT-device) was not significantly
attenuated upon dosing (See section SXVI).
For breakthrough systems with two identical sensors (i.e.,
Ni3HHTP2/Ni3HHTP2 or Ni3HITP2/Ni3HITP2), analyte
breakthrough showed attenuated response between the first
and second sensor. For the sake of quantitative comparison,
breakthrough responses at 10 min of analyte exposure were
considered, and full representative traces are shown in Figure
5B. At 10 min of exposure to NO (10 ppm), Ni3HITP2/
Ni3HITP2 increased in response by 11% for the first sensor, and
only 2% for the second, corresponding to uptake of 0.4 mmol
of NO (∼12.9 mmol of NO/mg of Ni3HITP2) before
breakthrough (accounting for the volume of exposed SOFT-
device). For Ni3HHTP2/Ni3HHTP2, 10 min of exposure to
NO (20 ppm) caused a decrease in resistance of −3% for the
first sensor and −1% for the second, equivalent to uptake of 0.8
mmol of NO (∼22.9 mmol of NO/mg of Ni3HHTP2). With
exposure to H2S (10 ppm), Ni3HITP2/Ni3HITP2 devices
increased in the response measured by the first sensor by 65%,
and 16% for the second, corresponding to an estimated uptake
of 0.4 mmol of H2S before breakthrough, equivalent to ∼12.3
mmol of H2S/mg of Ni3HITP2. For Ni3HHTP2/Ni3HHTP2
exposed to H2S (20 ppm), an increase in response of 7% for the
first sensor and 2% for the second was recorded after 10 min of
exposure: an estimated uptake of 0.7 mmol of H2S before
breakthrough (∼23.8 mmol of H2S/mg of Ni3HHTP2).
We performed breakthrough studies employing two SOFT-
sensors layered in a custom enclosure, shown in Figure 5A.
The results shown in Figure 5B reflect a remarkable capacity
for SOFT-devices to simultaneously detect and capture analytes
below their permissible exposure limits. The practical
implications of such results are highly promising with respect
to applications of SOFT-sensors in novel portable filtering and
sensing technology. A gas mask that employed one or more
layers of SOFT-sensor would allow for a warning to be
transduced before user exposure to toxins, with a predeter-
mined breakthrough rate based on analyte concentration.
CONCLUSION
Figure 5. Breakthrough studies for simultaneous detection and capture
of analytes. (A) Custom Teflon enclosure used to determine
membrane breakthrough for SOFT-devices. Two devices are loaded
such that the diameter of the inner bore for gas flow is completely
covered by SOFT-devices. Spring loaded gold pins immobilize the
sensors and serve as electrodes. The enclosure is then completely
sealed for leak-free flow. (B) Representative breakthrough sensing
traces for SOFT-devices. The first device exposed to analyte is
represented by a solid line, the second device by a dashed line. The
delayed response of the second device reflects gas capture by the first
device.
■
This paper describes the first example of flexible, textile-
supported chemiresistive sensors employing conductive metal−
organic frameworks (MOFs) as functional materials. The use of
direct assembly produces robust e-textiles over large areas, and
their preparation from direct assembly of monomers to produce
conformal coatings of conductive material at the fiber level is
efficient, straightforward, and requires no specialized equip-
ment. The SOFT-device arrays represent a novel class of smart
fabric sensors that are mechanically robust and capable of
simultaneous, quantitative detection, uptake, and differentiation
of hazardous gases.
The SOFT-devices reported herein serve as the first example
of e-textiles employing MOFs as the functional component and
possess at least three unique performance characteristics of
broad potential utility in gas detection and capture. First,
Ni3HITP2 and Ni3HHTP2 SOFT-devices serve as the first
single-platform e-textile sensor capable of detecting both NO
(LOD = 0.16 and 1.4 ppm, respectively, 0.1−80 ppm) and H2S
(LOD = 0.52 and 0.23 ppm, respectively, 1−80 ppm) and
constitute the best MOF-based chemiresistors for NO and H2S
detection. Second, these devices retain their performance in
18% humidity and are fully recoverable and washable. Third, a
single SOFT-device effectively sequesters NO or H2S in
simultaneous uptake and detection step, providing predictable
rates of breakthrough based on analyte concentration.
This experiment demonstrated potential for quantitative real-
time detection and filtration of NO and H2S (Figure 5B,
section SXVI) using SOFT-devices. The experiment was
prepared such that the pressurized gas stream (gas flow =
∼0.5 L/min) must pass through a first SOFT-sensor, followed
by free space (∼0.20 cm3), then a second SOFT-sensor (Figure
5A). The chemiresistive response for each sensor was
monitored continuously, and the devices were allowed to
equilibrate in N2 until a stable baseline current was observed.
Devices were monitored for 10 min in dry N2, then dosed with
analyte for a span of 60 min before recovering in N2. To ensure
that any capture of analyte was specific to the SOFT-device,
and not to the cotton, we performed experiments wherein the
first SOFT-device was replaced by unmodified cotton fabric. In
these control experiments, breakthrough of analytes to the
G
DOI: 10.1021/jacs.7b08840
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
While these systems are robust, they may lose mass upon
handling that involves cutting or puncturing of the fabric due to
fraying or MOF shedding. This limitation could be overcome
by the introduction of a porous polymeric coating, which would
add long-term stability to the device but still be permeable
enough to allow perfusion of gaseous analytes or water to the
SOFT swatch. Technical optimization of factors influencing
wearability, long-term stability, and performance in the
presence of high humidity and in complex sensing environ-
ments that include mixtures of potentially interfering gases will
need to be addressed before practical application of SOFT-
devices in the state-of-the-art wearable systems.
The behavior of SOFT-devices provides fundamental insight
on properties of this class of conductive MOFs, including
mechanistic details of analyte interaction and change transport.
The dosimetric responses of SOFT-devices to analytes are
entirely recoverable by washing, which suggests metal−analyte
binding interactions that are consistent with labile Lewis acid
site coordination. The direction of the dosimetric responses
implies that Ni3HITP2 is a mixed semiconductor, while
Ni3HHTP2 may be primarily p-type in nature. We believe
that these SOFT-devices have the potential to help elucidate
further fundamental properties of conductive MOFs. In the
future, this technology may have broad utility in sensing
applications ranging from real-time gas detection in wearable
systems, to electronically accessible adsorbant layers in
protective equipment like gas masks. Finally, our fabrication
method has the potential to be extended into other systems,
producing a range of novel, multifunctional e-textiles with
highly tunable properties, and a host of unprecedented
applications in wearable and portable devices.
REFERENCES
■
(1) Rogers, J. A.; Someya, T.; Huang, Y. Science 2010, 327, 1603.
(2) Kim, D.-H.; Lu, N.; Ma, R.; Kim, Y.-S.; Kim, R.-H.; Wang, S.; Wu,
J.; Won, S. M.; Tao, H.; Islam, A.; Yu, K. J.; Kim, T.; Chowdhury, R.;
Ying, M.; Xu, L.; Li, M.; Chung, H.-J.; Keum, H.; McCormick, M.; Liu,
P.; Zhang, Y.-W.; Omenetto, F. G.; Huang, Y.; Coleman, T.; Rogers, J.
A. Science 2011, 333, 838.
(3) Someya, T.; Bao, Z.; Malliaras, G. G. Nature 2016, 540, 379.
(4) Steinhubl, S. R.; Muse, E. D.; Topol, E. J. Sci. Transl. Med. 2015,
7, 283rv3.
(5) Gao, W.; Emaminejad, S.; Nyein, H. Y. Y.; Challa, S.; Chen, K.;
Peck, A.; Fahad, H. M.; Ota, H.; Shiraki, H.; Kiriya, D.; Lien, D.-H.;
Brooks, G. A.; Davis, R. W.; Javey, A. Nature 2016, 529, 509.
(6) Chortos, A.; Liu, J.; Bao, Z. Nat. Mater. 2016, 15, 937.
(7) Snyder, E. G.; Watkins, T. H.; Solomon, P. A.; Thoma, E. D.;
Williams, R. W.; Hagler, G. S. W.; Shelow, D.; Hindin, D. A.; Kilaru, V.
J.; Preuss, P. W. Environ. Sci. Technol. 2013, 47, 11369.
(8) Weng, W.; Chen, P.; He, S.; Sun, X.; Peng, H. Angew. Chem., Int.
Ed. 2016, 55, 6140.
(9) Zeng, W.; Shu, L.; Li, Q.; Chen, S.; Wang, F.; Tao, X.-M. Adv.
Mater. 2014, 26, 5310.
(10) Zhang, L.; Baima, M.; Andrew, T. L. ACS Appl. Mater. Interfaces
2017, 9, 32299.
(11) Castano, L. M.; Flatau, A. B. Smart Mater. Struct. 2014, 23,
053001.
(12) Rothmaier, M.; Luong, M. P.; Clemens, F. Sensors 2008, 8, 4318.
(13) Amjadi, M.; Kyung, K.-U.; Park, I.; Sitti, M. Adv. Funct. Mater.
2016, 26, 1678.
(14) Diamond, D.; Coyle, S.; Scarmagnani, S.; Hayes, J. Chem. Rev.
2008, 108, 652.
(15) Han, J.-W.; Kim, B.; Li, J.; Meyyappan, M. Appl. Phys. Lett. 2013,
102, 193104.
(16) Hong, K. H.; Oh, K. W.; Kang, T. J. J. Appl. Polym. Sci. 2004, 92,
37.
(17) Yetisen, A. K.; Qu, H.; Manbachi, A.; Butt, H.; Dokmeci, M. R.;
Hinestroza, J. P.; Skorobogatiy, M.; Khademhosseini, A.; Yun, S. H.
ACS Nano 2016, 10, 3042.
(18) Bandodkar, A. J.; Jeerapan, I.; Wang, J. ACS Sens. 2016, 1, 464.
(19) Singh, E.; Meyyappan, M.; Nalwa, H. ACS Appl. Mater. Interfaces
2017, 9, 34544.
(20) Potyrailo, R. A.; Surman, C.; Nagraj, N.; Burns, A. Chem. Rev.
2011, 111, 7315.
(21) Potyrailo, R. A. Chem. Rev. 2016, 116, 11877.
(22) Stassen, I.; Burtch, N.; Talin, A.; Falcaro, P.; Allendorf, M.;
Ameloot, R. Chem. Soc. Rev. 2017, 46, 3185.
ASSOCIATED CONTENT
* Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/jacs.7b08840.
■
S
Experimental methods, characterization of MOFs and
devices, electronic characteristics, mechanical properties,
and customization of SOFT-devices, comprehensive
chemiresistive sensing performance and recovery of
SOFT-devices, and breakthrough control experiments
(PDF)
(23) Sun, L.; Campbell, M. G.; Dinca,
2016, 55, 3566.
̆
M. Angew. Chem., Int. Ed.
(24) Kobayashi, Y.; Jacobs, B.; Allendorf, M. D.; Long, J. R. Chem.
Mater. 2010, 22, 4120.
(25) Darago, L. E.; Aubrey, M. L.; Yu, C. J.; Gonzalez, M. I.; Long, J.
AUTHOR INFORMATION
Corresponding Author
*katherine.a.mirica@dartmouth.edu
■
R. J. Am. Chem. Soc. 2015, 137, 15703.
(26) Campbell, M. G.; Dinca,
(27) Campbell, M. G.; Liu, S. F.; Swager, T. M.; Dinca,
Chem. Soc. 2015, 137, 13780.
(28) Campbell, M. G.; Sheberla, D.; Liu, S. F.; Swager, T. M.; Dinca,
M. Angew. Chem., Int. Ed. 2015, 54, 4349.
̆
M. Sensors 2017, 17, 1108.
ORCID
̆
M. J. Am.
Katherine A. Mirica: 0000-0002-1779-7568
Notes
The authors declare no competing financial interest.
̆
(29) Smith, M. K.; Jensen, K. E.; Pivak, P. A.; Mirica, K. A. Chem.
Mater. 2016, 28, 5264.
ACKNOWLEDGMENTS
■
(30) Ko, M.; Aykanat, A.; Smith, M. K.; Mirica, K. A. Sensors 2017,
17, 2192.
(31) Sheberla, D.; Bachman, J. C.; Elias, J. S.; Sun, C.-J.; Shao-Horn,
Y.; Dinca, M. Nat. Mater. 2017, 16, 220.
̆
(32) Yan, Y.; Gu, P.; Zheng, S.; Zheng, M.; Pang, H.; Xue, H. J.
Mater. Chem. A 2016, 4, 19078.
(33) Clough, A. J.; Yoo, J. W.; Mecklenburg, M. H.; Marinescu, S. C.
J. Am. Chem. Soc. 2015, 137, 118.
(34) Miner, E. M.; Fukushima, T.; Sheberla, D.; Sun, L.;
Surendranath, Y.; Dinca, M. Nat. Commun. 2016, 7, 10942.
̆
K.A.M. acknowledges support from startup funds provided by
Dartmouth College, from the Walter and Constance Burke
Research Initiation Award, and from the Army Research Office
Young Investigator Program Grant No. W911NF-17-1-0398.
The authors thank Michael Ko for contributions towards
Ohmic behavior of SOFT-devices in the presence of analytes
and assistance with characterizing the control samples in the
SOFT-device fabrication and Aylin Aykanat for help with BET
measurements.
H
DOI: 10.1021/jacs.7b08840
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
(35) Dong, R.; Pfeffermann, M.; Liang, H.; Zheng, Z.; Zhu, X.;
Zhang, J.; Feng, X. Angew. Chem., Int. Ed. 2015, 54, 12058.
(36) Stock, N.; Biswas, S. Chem. Rev. 2012, 112, 933.
(37) Zhou, H.-C.; Long, J. R.; Yaghi, O. M. Chem. Rev. 2012, 112,
673.
(38) Kreno, L. E.; Leong, K.; Farha, O. K.; Allendorf, M.; Van Duyne,
R. P.; Hupp, J. T. Chem. Rev. 2012, 112, 1105.
(68) Gregory, R. V.; Kimbrell, W. C.; Kuhn, H. H. Synth. Met. 1989,
28, 823.
(69) Liu, B.; Chen, L.; Liu, G.; Abbas, A. N.; Fathi, M.; Zhou, C. ACS
Nano 2014, 8, 5304.
(70) Konvalina, G.; Haick, H. Acc. Chem. Res. 2014, 47, 66.
(71) Graedel, T. E.; Weschler, C. J. Rev. Geophys. 1981, 19, 505.
(72) Calvert, J. G.; Lazrus, A.; Kok, G. L.; Heikes, B. G.; Walega, J.
G.; Lind, J.; Cantrell, C. A. Nature 1985, 317, 27.
(73) Mustafa, A. K.; Gadalla, M. M.; Snyder, S. H. Sci. Signaling 2009,
2, re2.
(39) Kusgens, P.; Siegle, S.; Kaskel, S. Adv. Eng. Mater. 2009, 11, 93.
̈
(40) Chen, T.-Y.; Huang, Y.-J.; Li, C.-T.; Kung, C.-W.; Vittal, R.; Ho,
K.-C. Nano Energy 2017, 32, 19.
(74) Ammu, S.; Dua, V.; Agnihotra, S. R.; Surwade, S. P.; Phulgirkar,
A.; Patel, S.; Manohar, S. K. J. Am. Chem. Soc. 2012, 134, 4553.
(75) Desai, A. V.; Samanta, P.; Manna, B.; Ghosh, S. K. Chem.
Commun. 2015, 51, 6111.
(76) Yassine, O.; Shekhah, O.; Assen, A. H.; Belmabkhout, Y.;
Salama, K. N.; Eddaoudi, M. Angew. Chem., Int. Ed. 2016, 55, 15879.
(77) Wilkins, R. G. Acc. Chem. Res. 1970, 3, 408.
(78) Williams, D. E. Sens. Actuators, B 1999, 57, 1.
(79) Achmann, S.; Hagen, G.; Kita, J.; Malkowsky, I. M.; Kiener, C.;
Moos, R. Sensors 2009, 9, 1574.
(41) Neufeld, M. J.; Harding, J. L.; Reynolds, M. M. ACS Appl. Mater.
Interfaces 2015, 7, 26742.
(42) Lo
́
Cervantes, S. D. A.; Lozano-Per
́
Navarro, J. A. R. Angew. Chem., Int. Ed. 2015, 54, 6790.
(43) Ozer, R. R.; Hinestroza, J. P. RSC Adv. 2015, 5, 15198.
(44) Tian, H.; He, J. Langmuir 2016, 32, 12269.
(45) Meilikhov, M.; Yusenko, K.; Schollmeyer, E.; Mayer, C.;
Buschmann, H.-J.; Fischer, R. A. Dalton Trans. 2011, 40, 4838.
(46) Zhao, J.; Losego, M. D.; Lemaire, P. C.; Williams, P. S.; Gong,
B.; Atanasov, S. E.; Blevins, T. M.; Oldham, C. J.; Walls, H. J.;
Shepherd, S. D.; Browe, M. A.; Peterson, G. W.; Parsons, G. N. Adv.
Mater. Interfaces 2014, 1, 1400040.
(47) Darder, M. d. M.; Salehinia, S.; Parra, J. B.; Herrero-Martinez, J.
̀
M.; Svec, F.; Cerda, V.; Turnes Palomino, G.; Maya, F. ACS Appl.
Mater. Interfaces 2017, 9, 1728.
(80) Konvalina, G.; Haick, H. ACS Appl. Mater. Interfaces 2012, 4,
317.
(81) Morimoto, T.; Nagao, M.; Tokuda, F. J. Phys. Chem. 1969, 73,
243.
́
(82) Alzate-Sanchez, D. M.; Smith, B. J.; Alsbaiee, A.; Hinestroza, J.
P.; Dichtel, W. R. Chem. Mater. 2016, 28, 8340.
(83) Dechnik, J.; Sumby, C. J.; Janiak, C. Cryst. Growth Des. 2017, 17,
4467.
(48) Li, S.; Dharmarwardana, M.; Welch, R. P.; Ren, Y.; Thompson,
C. M.; Smaldone, R. A.; Gassensmith, J. J. Angew. Chem., Int. Ed. 2016,
55, 10691.
(84) Aroon, M. A.; Ismail, A. F.; Matsuura, T.; Montazer-Rahmati, M.
M. Sep. Purif. Technol. 2010, 75, 229.
(85) Qiu, S.; Xue, M.; Zhu, G. Chem. Soc. Rev. 2014, 43, 6116.
(49) Doonan, C.; Ricco,
Chem. Res. 2017, 50, 1423.
(50) Bradshaw, D.; Garai, A.; Huo, J. Chem. Soc. Rev. 2012, 41, 2344.
̀
R.; Liang, K.; Bradshaw, D.; Falcaro, P. Acc.
(51) Liu, J.; Woll, C. Chem. Soc. Rev. 2017, 46, 5730.
̈
(52) Denny, M. S., Jr.; Moreton, J. C.; Benz, L.; Cohen, S. M. Nat.
Rev. Mater. 2016, 1, 16078.
(53) Li, M.; Dinca, M. J. Am. Chem. Soc. 2011, 133, 12926.
̆
(54) Shah, M.; McCarthy, M. C.; Sachdeva, S.; Lee, A. K.; Jeong, H.-
K. Ind. Eng. Chem. Res. 2012, 51, 2179.
(55) Gascon, J.; Kapteijn, F.; Zornoza, B.; Sebastian
́
, V.; Casado, C.;
Coronas, J. Chem. Mater. 2012, 24, 2829.
(56) Li, J.-R.; Sculley, J.; Zhou, H.-C. Chem. Rev. 2012, 112, 869.
(57) Zirehpour, A.; Rahimpour, A.; Arabi Shamsabadi, A.; Sharifian
Gh., M.; Soroush, M. Environ. Sci. Technol. 2017, 51, 5511.
(58) Dimitrakakis, C.; Marmiroli, B.; Amenitsch, H.; Malfatti, L.;
Innocenzi, P.; Grenci, G.; Vaccari, L.; Hill, A. J.; Ladewig, B. P.; Hill,
M. R.; Falcaro, P. Chem. Commun. 2012, 48, 7483.
(59) Lu, G.; Farha, O.; Zhang, W.; Huo, F.; Hupp, J. T. Adv. Mater.
2012, 24, 3970.
(60) Shekhah, O.; Wang, H.; Paradinas, M.; Ocal, C.; Schupbach, B.;
̈
Terfort, A.; Zacher, D.; Fischer, R. A.; Woll, C. Nat. Mater. 2009, 8,
481.
̈
(61) Hermes, S.; Schroder, F.; Chelmowski, R.; Woll, C.; Fischer, R.
̈
̈
A. J. Am. Chem. Soc. 2005, 127, 13744.
(62) Carbonell, C.; Imaz, I.; Maspoch, D. J. Am. Chem. Soc. 2011,
133, 2144.
(63) Zhuang, J.-L.; Ar, D.; Yu, X.-J.; Liu, J.-X.; Terfort, A. Adv. Mater.
2013, 25, 4631.
́
(64) Hmadeh, M.; Lu, Z.; Liu, Z.; Gandara, F.; Furukawa, H.; Wan,
S.; Augustyn, V.; Chang, R.; Liao, L.; Zhou, F.; Perre, E.; Ozolins, V.;
Suenaga, K.; Duan, X.; Dunn, B.; Yamamto, Y.; Terasaki, O.; Yaghi, O.
M. Chem. Mater. 2012, 24, 3511.
(65) Sheberla, D.; Sun, L.; Blood-Forsythe, M. A.; Er, S.; Wade, C.
R.; Brozek, C. K.; Aspuru-Guzik, A.; Dinca,
136, 8859.
̆
M. J. Am. Chem. Soc. 2014,
(66) Wang, H.; Lewis, D. M. Color. Technol. 2002, 118, 159.
(67) Hearle, J. W. S.; Moreton, W. E. Physical Properties of Textile
Fibres, 4th ed.; Woodhead Publishing Limited: Cambridge, England,
2008.
I
DOI: 10.1021/jacs.7b08840
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX