Response Characteristics of Hydrogen Sensors Based on PMMA- Membrane-Coated Palladium Nanoparticle Films
Minrui Chen,
†,‡Peng Mao,
∥,⊥Yuyuan Qin,
‡Jue Wang,
‡Bo Xie,*
,†Xiuzhang Wang,
†Deyan Han,
§Guo-hong Wang,
§Fengqi Song,
‡Min Han,
‡Jun-Ming Liu,
‡and Guanghou Wang
‡†Institute for Advanced Materials, Hubei Normal University, Huangshi 435002, P. R. China
‡National Laboratory of Solid State Microstructures, Nanjing University, Nanjing 210093, P. R. China
§Hubei Collaborative Innovation Center for Rare Metal Chemistry, Hubei Normal University, Huangshi 435002, P. R. China
∥Nanoscale Physics Research Laboratory, School of Physics and Astronomy, University of Birmingham, Birmingham B15 2TT, United Kingdom
⊥College of Electronic Science and Engineering, Nanjing University of Posts and Telecommunications, Nanjing 210023, P. R. China
*S Supporting Information
ABSTRACT: Coating a polymeric membrane for gas separation is a feasible approach to fabricate gas sensors with selectivity. In this study, poly(methyl methacrylate)-(PMMA- )membrane-coated palladium (Pd) nanoparticle (NP) films were fabricated for high-performance hydrogen (H2) gas sensing by carrying out gas-phase cluster deposition and PMMA spin coating. No changes were induced by the PMMA spin coating in the electrical transport and H2-sensing mechanisms of the Pd NPfilms. Measurements of H2sensing demonstrated that the devices were capable of detecting H2
gas within the concentration range 0−10% at room temperature and showed high selectivity to H2due to thefiltration effect of the PMMA membrane layer. Despite the presence of the PMMA matrix, the lower detection limit of the sensor is less than 50 ppm. A series of PMMA membrane layers with different thicknesses were spin coated onto the surface of Pd NPfilms for the selectivefiltration of H2. It was found that the device sensing kinetics were strongly affected by the thickness of the PMMA layer, with the devices with thicker PMMA membrane layers showing a slower response to H2gas. Three mechanisms slowing down the sensing kinetics of the devices were demonstrated to be present: diffusion of H2gas in the PMMA matrix, nucleation and growth of theβphase in theαphase matrix of Pd hydride, and stress relaxation at the interface between Pd NPs and the PMMA matrix. The retardation effect caused by these three mechanisms on the sensing kinetics relied on the phase region of Pd hydride during the sensing reaction. Two simple strategies, minimizing the thickness of the PMMA membrane layer and reducing the size of the Pd NPs, were proposed to compensate for retardation of the sensing response.
KEYWORDS: hydrogen sensors, PMMA membrane layer, Pd nanoparticle films, H2-sensing kinetics, electrical transport mechanism
■
INTRODUCTIONHydrogen gas (H2), as the most important new alternative energy source, and one of the cleanest sources, has attracted considerable attention because of its high energy density, low cost, and renewability. However, H2 is highly flammable and explosive when its concentration exceeds the lower explosion limit in air of 4% at room temperature. Developing a reliable, accurate, fast, and wide-range hydrogen gas leakage detector is very important for hydrogen generation, storage, and utilization.
Many types of H2sensors have been reported, including sensors based on metal oxides,1,2 acoustic waves,3−5 and metal and semiconductor thin films.6−8 However, these sensors have a number of significant drawbacks, notably, their response time, selectivity, sensitivity, and repeatability, that make them unsuitable for many applications.
Recently, palladium (Pd) nanostructures have been used to fabricate high-performance H2sensors,9−18such as the sensor
devices we developed through the gas-phase deposition of Pd nanoparticle (NP) films with controlled coverage on gold interdigital electrodes,9,19−21and the H2 sensor based on Pd nanowire arrays.13,14,22−28 In these devices, the sensing mechanism is based on the tunneling and/or hopping of electrons between the nanogaps which are closed by hydrogen- induced lattice expansion (HILE) during H2 exposure.22 Changes in the gap size owing to lattice expansion or the dwindling of Pd during H2absorption or desorption results in changes in the electron barrier, leading to a significant and measurable change in the conductance of the nanostructures.22 Although Pd-nanostructure-based H2sensors exhibit excellent sensing performance at room temperature, cross-sensitivity with
Received: May 29, 2017 Accepted: July 25, 2017 Published: July 25, 2017
www.acsami.org
carbon monoxide (CO), water vapor, and methane (CH4) remains a critical issue. More importantly, among the many cross-sensitive gases, CO is highly toxic to the Pd catalyst. As a consequence, the H2-sensing ability of Pd is destroyed by CO, which is attributed mainly to a reduction in the available dissociation sites on the Pd metal surface caused by CO adsorption.29−32 It is fortunate that developments in H2 separation technologies based on polymeric membranes hold promise for smaller, cheaper, and more feasible H2sensors with high selectivity.33−36Recent reports have shown that polymeric membranes could act as a gas-selective polymer barrier in H2gas sensors, enabling them to absorb and release H2without cross- sensitivity and toxicity.33,37,38In this regard, Lee et al. reported the first demonstration of a H2 sensor based on a polymer- membrane-coated Pd NP/single-layer graphene (SLG) hybrid structure that achieved high selectivity toward H2,37 but the understanding of the sensing kinetics of the H2sensors is still quite limited. In the case of the nanostructured HILE-based H2 sensors, further investigations are required to confirm whether the electrical transport and sensing mechanisms have been changed or not because of the encapsulation of the PMMA matrix before the PMMA membrane layer would be used as a selective gas separator. Also, the fast response speed of the nanostructured HILE-based H2sensors may be lost because of the barrier effect of the selectivefiltration membrane. On the basis of the fact that rapid response is a major challenge in the development of H2 sensors, the effect of the thickness of polymeric membranes on the H2-sensing kinetics still requires close examination to establish the feasibility of the use of this method for the fabrication of nanostructured HILE-based H2 sensors with high selectivity and fast response.
In this report, Pd NPfilms were fabricated by the gas-phase deposition of Pd nanoclusters with controlled coverage on gold interdigital electrodes.20 A thin layer of poly(methyl meth- acrylate) (PMMA) was spin coated onto a Pd NP film, to fabricate a high-performance H2sensor with gas selectivity. The electrical transport and H2-sensing mechanisms of the Pd NP films before and after PMMA coating were investigated. Further, the sensing behavior of this device was examined, with a focus on the kinetic characteristics of the Pd NPfilm coated by the PMMA membrane when responding to H2. The sensors with the PMMA membrane layers clearly exhibited a prolonged response time
which is strongly dependent on not only the thickness of the PMMA membrane layer but also the phase region of Pd hydride during the sensing reaction. Through the analysis of experimental data, three mechanisms slowing down the sensing kinetics of the devices were proposed. Meanwhile, two simple strategies, minimizing the thickness of the PMMA membrane layer and reducing the size of the Pd NPs, were suggested to compensate for the prolonged response time in consideration of the retardation effect caused by the PMMA matrix.
■
EXPERIMENTAL SECTIONDevice Fabrication. Figure 1 is a schematic illustration of the preparation of the sensor devices. The Pd NPs were deposited on prefabricated gold interdigital electrodes by a cluster beam deposition system. The gold interdigital electrodes, 80 nm thick and with 4μm electrode separation, were patterned onto a silicon substrate with a 300 nm silicon dioxide (SiO2) insulating surface layer using a standard lithographic lift-offprocedure. Pd NPs were formed in a high-vacuum chamber by a magnetron plasma gas aggregation cluster source. Argon (100 sccm) was introduced into the aggregation tube to maintain a stable pressure of 120 Pa. A stable magnetron discharge was carried out with an input power of 47 W. Atoms were sputtered from the Pd target, and Pd nanoclusters formed through the aggregation process in the argon gas. During deposition, the current across the electrode gap was measured in real time with a source meter (Keithley 2601B) by applying a bias of 1 V, as shown inFigure 1a. The Pd NPfilm was formed after the deposition, which was stopped when the predetermined conductance values defined as the initial conductanceG0were attained (Figure 1b).
For the PMMA coating, 4 g of PMMA powder from Alfa Aesar with molar mass 400 000−550 000 g mol−1, according to manufacturer data, was dispersed in 100 mL of anisole. The mixture, agitated by a magnetic stirrer, was heated using a 60°C water bath for 6 h to accelerate the dissolution of the PMMA. The mixture was then coated on the surface of the sensor devices (predeposited Pd NPfilms) by spin coating (spin coater, KW-4A, Institute of Microelectrons of the Chinese Academy of Sciences) for 30 s. A PMMA membrane layer was formed on the surface of the Pd NPfilm deposited on the interdigital electrode substrate after drying in air for 24 h, as shown inFigure 1c,d. The thickness of the PMMA membrane layer could be varied by adjustment of the spin velocity.
Structure Characterization.The scanning transmission electron microscopy (STEM) investigation was performed using a JEOL instrument (JEM2100F) with a spherical-aberration corrector (CEOS GmbH). The images were acquired using high-angle annular darkfield (HAADF) and brightfield (BF) detectors. For facilitation of the STEM Figure 1.Schematic illustration of the procedure used to fabricate the hydrogen sensor based on a PMMA-membrane-coated Pd NPfilm. (a) Deposition of Pd NPs. (b) Pd NPfilm formed on the interdigital electrode substrate after the NP deposition. (c) PMMA-membrane-coated Pd NPfilm fabricated by spin coating. (d) Structure of the hydrogen sensor based on the PMMA-membrane-coated Pd NPfilm.
ACS Applied Materials & Interfaces
observation, Pd NPs were deposited on SiO2films supported by copper grids at the same time as sensor device fabrication. The thickness of the PMMA membrane layer was measured by scanning edge sections using atomic force microscopy (AFM; 175 NTEGRA Probe NanoLaboratory, NT-MDT Co.).
Electron Transport Measurement. The investigation of the electron transport property of Pd NPfilms was carried out by measuring temperature-dependence on resistance. Temperature-dependent dc resistance was measured using a closed cycle cryostat (Janis CCS450) with a temperature controller (Cryogenic Model 32B) over the range 20−300 K. The resistance was monitored using a source meter (Keithley 2601B).
Gas-Sensing Tests. The H2-sensing response of the fabricated sensors was studied at room temperature in a gas flow chamber fabricated in our laboratory. The tested target gases included pure H2 and mixtures of CO/nitrogen (N2), CH4/N2, H2/N2, and H2/air. For pure H2, since the sensor responds only to the H2partial pressure, it is easy to convert the data measured for pure H2gas to a H2concentration response for a gas mixture with a minor deviation by assuming 1 atm ambient pressure. Theflow of the target gases to the test chamber was regulated by a mass flow controller (Sevenstar D07-7A/ZM). The pressure of pure H2gas (PH2) in the test chamber was monitored by a piezo-resistive gauge. During the H2-sensing measurement, the current across the electrode gap was measured in real time with a source meter by applying a bias of 1 V. The relative change in conductance is defined as the sensing responseS:
= Δ = −
S G G/ 0 (G G0)/G0 (1) whereGis the conductance of the sensor exposed to the target gas, and G0 is the initial conductance of the Pd NP film without target gas loading.
A rapid loading of H2is necessary for measuring the response time of the sensors. As shown inFigure S1in the Supporting Information, the samples were therefore sealed inside a small measurement cell that was connected to a much larger chamber. The capacity of the large chamber and small measurement cell is about 22.4 L and 9.6 mL, respectively. For each H2pulse measurement,first, keeping the electromagnetic valve open, the large chamber and small measurement cell were pumped until the vacuum is below 1 Pa. Next, the electromagnetic valve was closed,
and then the large chamber was filled with H2 to a predetermined pressure. A H2 pulse could be formed and entered the small measurement cell as the electromagnetic valve was rapidly opened.
Because of the great difference in volume between the large chamber and measurement cell, the pressure changes in the large chamber are negligible during the formation of the H2pulse. Simultaneously, both PH2 and the conductance of the Pd NP films were recorded. The response time is defined as the time required to reach a conductance change (ΔG) of 90% at a given H2pulse pressure.
■
RESULTS AND DISCUSSIONThe structural characterization of a typical Pd NPfilm sample is shown by the HAADF-STEM images inFigure 2a. As seen, Pd NPs are randomly distributed in the substrate and form numerous closely spaced particle-assembling areas. These randomly distributed Pd NPs in the substrate resulted in the random formation of a large number of nanogaps, which constitute numerous quantum transport39pathways based on percolation.40The charge transport in our sensor occurs through tunneling and/or hopping between the Pd NPs. The exponential dependence of the tunneling and/or hopping probability41on the size of the nanogaps allows the potential development of the high-performance sensor.22Figure 2b shows the HAADF-STEM and BF-STEM images of a single Pd NP. The lattice fringes have an interplanar spacing of 0.226 nm, corresponding to the (111) planes of the face-centered cubic (FCC) structure of metallic Pd, which clearly reveals the crystalline nature of the Pd NPs synthesized in the present study. According to the selected-area electron diffraction (SAED) pattern (Figure 2c), it can also be deduced that the Pd NPs were in a crystalline state. The main diffraction rings can be assigned to the Pd FCC phase corresponding to the (111), (200), (220), and (311) crystal planes. The morphology of the Pd NPfilm after PMMA coating is shown inFigure 2d. From the comparison between the images of Figure 2a,d, the Pd NPs exhibited the same organization before and after PMMA coating, indicating that PMMA has no Figure 2.(a) Typical HAADF-STEM image of Pd NPfilms. (b) Typical HAADF-STEM image (upper left) and BF-STEM image (lower right) of a single Pd NP. The inset shows the fringes of the Pd NP. (c) SAED pattern. (d) HAADF-STEM image of the Pd NPfilm after PMMA coating. (e) HAADF-STEM image of a few closely spaced Pd NPs after PMMA coating. (f) Size distribution of Pd NPs. The red line is the log-normalfitting of the distribution.
ACS Applied Materials & Interfaces
obvious effect on the surface morphology of the Pd NP film.
Figure 2e shows the details of a few closely spaced Pd NPs embedded in the PMMA matrix. This image ensures that the nanogaps between the closely spaced Pd NPs were not visibly disturbed. As mentioned above, the fabrication of Pd NPs was carried out by a cluster beam deposition system. This formation process of nanoclusters involves a gas expanding from a high pressure source into a vacuum, which makes the NPs deposit onto the substrate at a highflying speed. As a result, the NPs are adsorbed on the substrate with high adhesive force. The maintaining of the film morphology after PMMA coating, which can be attributed to the adhesive force between the NPs and the substrate, makes it possible for this nanostructure to act as a sensing device. We statistically analyzed the size of the Pd NPs, and the distribution fitting a log-normal dependence is presented inFigure 2f with a probable size of 10 nm.
We measured temperature-dependent resistance and H2- sensing response of a typical Pd NPfilm before and after PMMA coating to check whether the electrical transport mechanism and sensing behavior remain unchanged after PMMA coating. The thickness of the PMMA membrane layer is about 0.45μm.
Figure 3presents the temperature-dependence of resistance for the typical Pd NPfilm before and after PMMA coating. We
can clearly see that the resistanceRof the Pd NPfilm both before and after PMMA coating increases monotonically with decreasing temperatureTover the entire measurement temper- ature range 20−300 K, indicating that the Pd NPfilm exhibited the electrical transport phenomenon similar to semiconductors with a negative temperature coefficient. By plotting R(on log scale) versus T−1/4, we can see that the R−T curve is well- described by the Mott variable-range hopping (VRH) mecha- nism42because of the linear relationship between logRandT−1/4 as shown inFigure 3. The Mott VRH model gave a temperature- dependence conduction relationship:42
=
R RMexp[(TM/ )T 1/4] (2)
where RM is the resistance parameter, and TM is the Mott characteristic temperature whose value depends on the electronic density of states at the Fermi level and the localization length. The values ofTM of the Pd NPfilm before and after PMMA coating are 2.57×104and 1.33×104, respectively. In addition, the resistance of the Pd NPfilm increased after PMMA
coating. A detailed study of the physical mechanism that results in the value change ofRandTMby PMMA coating needs further investigation, but is beyond the aims of this study. However, whether or not the Pd NPfilm is coated by PMMA, it exhibited the electrical transport behavior dominated by the Mott VRH mechanism, indicating that the electronic transport mechanism of Pd NPfilm is essentially unchanged after PMMA coating. As a result, the unique quantum transport behavior of such a Pd NP film with the PMMA membrane layer presented an opportunity for fabricating high-performance HILE-based H2sensors.
Figure 4shows comparisons of the H2-sensing response for a typical Pd NPfilm before and after PMMA coating. As shown in Figure 4a,b, the sensor was exposed to several pure H2loading and deloading cycles; the measured gas pressures are noted at the peak of each response cycle. The positive sign of ΔG/G0
Figure 3.Resistance of the Pd NPfilm before and after PMMA coating plotted againstT−1/4. The solid curvesfit the experimental data withR∝ exp[(TM/T)1/4].
Figure 4.(a, b) Response curves for a typical Pd NPfilm exposed to pure H2loading and deloading cycles before and after PMMA coating. The pressure (in Pa) for each hydrogen loading is marked at the top of each response peak. (c) Linear−log calibration plots ofΔG/G0versusPH2for the same typical Pd NPfilm before and after the PMMA coating shown in parts a and b.
ACS Applied Materials & Interfaces
indicates that the conductance of the Pd NPfilm before and after PMMA coating increased and decreased quickly during H2 loading and deloading for each cycle. The H2-sensing response curves of the Pd NPfilm before and after PMMA coating are both similar to the results of our previous studies.9,20 The relative changes in conductance (ΔG/G0) versus PH2 are plotted in Figure 4c, in which the calibration curves for the sensors are shown. The calibration curves can be divided into three discrete response regimes corresponding to the three phase regions of Pd hydride, i.e., anαphase region, anα−βphase coexistence region, and a βphase region, in which the sensor exhibited different sensitivities to H2. AsPH2increased from 102to 105Pa, the sensor both before and after PMMA coating exhibited excellent response sensitivity, indicating that the devices were capable of detecting H2gas over a wide dynamic range (0−10%) at room temperature.
The response of the Pd NPfilm to a low concentration (50− 600 ppm) H2/air mixture after PMMA coating was tested to obtain the lower detection limit of the sensor with the PMMA membrane layer. The sensor exhibited a significant response to 50, 150, 250, 400, and 600 ppm concentrations of H2with about 0.37%, 0.62%, 0.86%, 1.25%, and 1.64% relative changes in conductance, respectively (Figure S2 in the Supporting Information). In this measurement, lower concentrations of the H2/air mixture were not tested, because of experimental limitations. However, on the basis of the fact that the response signal to 50 ppm of H2 is much higher than the conductance noise (Figure S2 in the Supporting Information), it can be concluded that the lower detection limit of the sensor with the PMMA membrane layer is less than 50 ppm.
As seen inFigure 4c, the calibration curve of the Pd NPfilm after PMMA coating is almost identical to that before PMMA coating untilPH2reached 3000 Pa, when Pd hydride entered theβ phase region. Thereafter, the response of the Pd NPfilm after PMMA coating dropped slightly. The results indicate that the Pd NPfilm maintained its response sensitivity to some extent even after PMMA coating. Previous studies have shown that Pd NPs undergo a large lattice expansion and form aβphase hydride43,44 when the H2concentration is increased above a certain value. We thus hypothesize that the response drop (inFigure 4c) of the Pd NPfilm after PMMA coating is due to the hindering effect of the PMMA layer on the large volume expansion of Pd hydride.45−47 Collectively, the data inFigure 4support the conclusion that the sensing mechanism of Pd NPfilms remains based on the hopping of electrons after PMMA coating, and that the PMMA coating had no effect on the mechanism. In fact, the maintenance of the sensing mechanism is essential for the remarkable H2-sensing performance of Pd NPfilms.
For inspection of the selective H2filtration effect of the PMMA membrane layer, the responses to gas mixtures of CO/N2, CH4/ N2, H2/N2; and the mixture of H2, CO, and CH4in N2[(H2+ CO + CH4)/N2] for Pd NP films with and without the spin coated PMMA membrane layer were compared, and the results are shown inFigure 5. The thickness of the PMMA membrane layer is about 0.45μm. For a test of the response to (H2+ CO + CH4)/N2, the mixtures of CO/N2, CH4/N2, and H2/N2were simultaneously fed into the test chamber through three massflow controllers at a 1.5 L min−1flow rate. As shown inFigure 5, the response to H2 of sensors with and without the PMMA membrane layer is approximately 9.1%, indicating little negative influence of the PMMA membrane layer on the response sensitivity of Pd NPfilms to H2. The responses of the sensor
without the PMMA membrane layer to CO and CH4are−3.4%
and 4.5%, respectively. However, the responses of the sensor with the PMMA membrane layer to CO and CH4have decreased by roughly an order of magnitude. Such a huge decrease of response to CO and CH4 observed for the sensor with the PMMA membrane layer is due to the filtration effect of the PMMA matrix. In addition, the responses of sensors with and without the PMMA membrane layer to (H2+ CO + CH4)/N2 are 2.9% and−0.7%, respectively. The negative response of the sensor without the PMMA membrane layer to (H2 + CO + CH4)/N2indicates that the CO adsorbs on the Pd metal surface more effectively than the other two kinds of gas, which can be attributed to the higher binding energy of CO on Pd. In the presence of the PMMA membrane layer, the sensor exhibited a positive response to (H2+ CO + CH4)/N2that is less than the response to H2/N2because of the dilution of H2in (H2+ CO + CH4)/N2. This at least proves that CO and CH4adsorption on the surface of Pd NPs was prevented while the sensors were exposed to (H2+ CO + CH4)/N2because of thefiltration effect of the PMMA membrane layer. Hence, these measurements revealed the remarkable fact that the PMMA membrane layer, as a selective gas filter for H2, can protect Pd NP films from responding to CO and CH4without significant degradation of their response sensitivity to H2.
For a study of the effect of the thickness of the PMMA membrane layer on the H2-sensing kinetics of Pd NPfilms, four samples with the PMMA membrane layer spin coated at rotational speeds of 2000, 3000, 4000, and 5000 rpm (denoted samples a, b, c, and d, respectively), and one sample without a PMMA membrane layer (denoted sample e), were prepared. The initial conductanceG0of allfive samples was 0.2μA during the Pd nanocluster deposition, to eliminate the influence of NP coverage on sensing performance.20,21
Figure 6a shows the AFM three-dimensional (3D) image of the edge section of the PMMA membrane layer on the surface of sample b. We can obtain the thickness of the PMMA layer by quantitatively measuring the step height of the PMMA edge section in the image. In the same way, the thickness of three other samples was measured. Figure 6b shows the thickness of the PMMA layer versus the rotational speed of the spin coater for the four samples. In Figure 6b, the thickness of the PMMA membrane layer of samples a, b, c, and d is 0.55, 0.49, 0.41, Figure 5.Response of sensors with and without a PMMA membrane layer to target gas mixtures including CO/N2, CH4/N2, H2/N2, and (H2 + CO + CH4)/N2. The concentration of CO/N2, CH4/N2, and H2/N2 was 1000 ppm. The tested mixed gas of H2, CO, and CH4in N2was formed by filling the three kinds of gas into the test chamber simultaneously.
ACS Applied Materials & Interfaces
and 0.36μm, respectively. It can be seen that there is a decrease in the thickness of the PMMA layer with increasing rotational speed for the four samples. In principle, gas permeability decreases with increased polymer membrane thickness. For a thicker PMMA membrane layer, the filtration effect is better, but the permeability of H2is lower, leading to the reduction of sensitivity and response speed during H2sensing. For qualification of the PMMA membrane layer to act as a gasfilter, the Pd NPs have to be fully sealed into the PMMA matrix. Therefore, the thickness of the PMMA membrane layer should be at least larger than the size of the Pd NPs in our devices. From the STEM images of Pd NPs inFigure 2and the data fromFigure 6, since the thickness of the PMMA membrane layer is much larger than the size of Pd NPs, it was deduced that the Pd NPs were fully embedded in the PMMA matrix. This full encapsulation enabled the PMMA membrane layer coated on the top of the Pd NP films to act as a gas separator, allowing H2to penetrate.34−36,48
Generally speaking, response time is one of most crucial technical parameters for a H2safety sensor, for obvious reasons.
Although the PMMA membrane layer provided enhanced performance in selectivity for H2 sensing, on the basis of the importance of the response time, it is necessary to evaluate the effect of the PMMA membrane layer on the sensor response time. Therefore, the response time of the aforementionedfive samples a−e was measured.
Figure 7a displays the response time of thefive samples versus PH2. The five samples all exhibited strong dependence of the response time onPH2, similar to the behavior of H2sensors based on Pd nanostructures.9,14,19,49However, two remarkable features of the data are shown inFigure 7a. First, the response time of all
samples was found to rapidly fall with an increase inPH2, except for a slow-response peak for the PH2 range 1−3 kPa, which corresponds to theαtoβphase transition of Pd hydride with the coexistence of theαandβphases of Pd hydride. This slow sensor response can be attributed mainly to the large lattice expansion caused by nucleation and growth of theβphase in theαmatrix because of H2absorption.44,50,51Second, samples a−d exhibited a significant prolonging of the response time compared with sample e. This slowing of the sensor response strongly depended on not only the thickness of the PMMA membrane layer but also PH2. As the thickness of the PMMA membrane layer increased, the response time was significantly prolonged. Moreover, we noticed that this slowing of the response time induced by the PMMA membrane layer became increasingly prevalent asPH2 was increased, especially in theβphase region (PH2≥3000 Pa).
To obtain further insights into the effect of the PMMA membrane layer on the response time of the sensors in the three hydride phase regions, we plotted the response time as a function of the thickness of the PMMA membrane layer at threePH2values Figure 6. (a) AFM 3D image of the edge section of the PMMA
membrane layer coated on the surface of sample b. (b) Thickness of PMMA membrane layers versus rotational speed.
Figure 7.(a) Response time as a function ofPH2for samples a−e on a linear−log scale. The dashed lines indicate the boundaries of the three hydride phase regions. (b) Plots of the response time versus the thickness of the PMMA membrane layer for samples a−e atPH2= 1, 3, and 2.3 kPa. The solid lines are the linearfitting of the dependence of the response time on the thickness of the PMMA membrane layer from 0.36 to 0.55μm. The slopes of the linearfitting (ux, wherexis the value of PH2) are indicated. The dashed lines serve to guide the eye.
ACS Applied Materials & Interfaces
(1, 3, and 2.3 kPa), corresponding to the start and end points of theαtoβphase transition and the point of lowest response in the α−βphase coexistence region, respectively (Figure 7b). It is clear that the linear relationship between the response time and the thickness of the PMMA membrane layer is in the range 0.36− 0.55μm for the threePH2values. As quantitatively analyzed in Figure 7b, the slopes of the threefitting curves, which represent the slowing effect of the PMMA membrane thickness on the response time, are 22.84, 69.34, and 111.18 sμm−1atPH2= 1, 2.3, and 3 kPa, respectively, indicating that, for higherPH2 values, especially in theβphase region, the retardation effect becomes stronger.
In the case ofPH2= 1 kPa, the response time of sample e, which was not coated with a PMMA membrane layer, coincides with the linear relationshipfitted within samples a−d (the red dashed line inFigure 7b). In the process of H2sensing, H2molecules diffused through the PMMA filter before adsorption on the surface of Pd NPs. The reduction in the response speed is mainly induced by the diffusion process of H2 molecules within the PMMA matrix. Therefore, the response timetis composed of two terms: the diffusion time through the PMMA layertD=Dh and the hydrogenation reaction time of the Pd NPfilmt0, where the coefficientDrefers to the diffusibility of H2in the PMMA matrix, andhis the thickness of the PMMA membrane layer, which is essentially the diffusion distance of H2molecules in the PMMA matrix. As a consequence, in theαphase region (PH2≤1 kPa), the response timetconsists of the hydrogenation reaction timet0and the diffusion time through the PMMA layertD.
The retardation of the response time in theα phase region mainly results from the diffusion of H2molecules through the PMMA layer, but this does not appear to be the case in theβ phase region, as indicated by the deviation of the response time of sample e from the linearfitting curve of samples a−d as shown inFigure 7b (the green dashed line). In the case ofPH2= 3000 Pa, the starting point of the βphase region, the Pd hydride has undergone a large lattice expansion of more than 3.47% of the lattice parameter (3.89 Å), which is much larger than the expansion in the α phase region (less than 0.13%) during absorption of H2.52In addition, because of the presence of the PMMA matrix, the large volume expansion of Pd NPs in theβ phase region suffers from the suppression induced by mechanical stress on the interface between the PMMA matrix and Pd NPs.53−55 Consequently, once the large volume expansion occurs in Pd NPs embedded in the PMMA matrix, the response time will be retarded because of stress relaxation. Therefore, we attribute the slowing of the response time in theβphase region to stress relaxation induced by the lattice expansion, in addition to the diffusion process of H2molecules within the PMMA matrix.
On the basis of this analysis, in theβphase region (PH2≥3 kPa), the response timethas three terms: the hydrogenation reaction timet0, the diffusion time through the PMMA layertD, and the prolonged response time induced by the mechanical stress relaxationtR.
This phenomenon agrees with the result that the response of the Pd NP film after PMMA coating dropped in the βphase region as shown inFigure 4c, indicating that the suppression of the volume expansion of Pd NPs in theβphase region not only prolonged the response time but also lowered the sensitivity. It is clear that the PMMA membrane inhibits the expansion of the Pd NPs by H2absorption. As a result, in these devices coated by PMMA, the conductance changes induced by H2 absorption
decrease with the increase inPH2. This means that a quantitative determination of the PH2 from the conductance measurement would be unreliable at a higher range of hydrogen pressures.
Accordingly, it is reasonable to conclude that the upper detection limit of the devices with PMMA membrane layers is lower than that of the devices without PMMA membrane layers because of the inhibitory action of the PMMA matrix on the expansion of Pd NPs by H2absorption. For the same reason, the lower detection limit of the devices is also affected by the PMMA matrix.
However, the PMMA matrix affects the upper detection limit more strongly than the lower detection limit because of the much larger volume expansion of Pd NPs in theβphase region than that in theαphase region.
With regard toFigure 7a, the slow-response peak indicates a huge delay in the response time for theα−βphase coexistence region of Pd hydride for all five samples. Associated with the diffusion of H2 molecules in the PMMA matrix and stress relaxation at the interface between Pd NPs and the PMMA membrane layer, the slow-response peak can be mainly attributed to the nucleation and growth of theβphase in theα phase matrix of Pd hydride, which plays a key role in the sensing kinetics. In theα−βphase coexistence region, the response timet is composed of four terms: the hydrogenation reaction timet0, the diffusion time through the PMMA layer tD, the prolonged response time induced by the mechanical stress relaxationtR, and the prolonged time induced by the nucleation and growthtN.
The response time of a sensor is mainly determined bytN, as it is larger than the other terms. The microscopic mechanism leading to the slow-response peak in Figure 7a remains unexplained.
Although all of the Pd NPfilms with a PMMA membrane layer in our study clearly exhibited up to tens of seconds of retardation of the sensing response, in view of the improvement of selectivity toward H2, this retardation is acceptable. Langhammer et al.
studied the hydriding and dehydriding kinetics of Pd NPs in different sizes.56It was proven that smaller Pd NPs are faster in the hydriding and dehydriding kinetic processes. In our previous study, the Pd NP films with different mean diameters were obtained by controlling the NP deposition. The response time of the devices was found to depend strongly on the Pd NP size: a smaller Pd NP shows faster response to H2.9In fact, the response time depends partly on the hydrogen diffusion in NPs. The diffusion length can be shortened by reducing the NP size.
Therefore, we suggest that it is feasible to compensate for the retardation of the sensing response by reducing the size of the Pd NPs. Minimizing the thickness of the PMMA membrane layer, which results in shortening of the diffusion length in the PMMA matrix, is also crucial for maintaining the high-speed response of the sensor. However, on the basis of the fact that the permeability increases with decreased polymer membrane thickness, the sensors lose gas selectivity once the PMMA membrane thickness is reduced to a critical dimension. On the basis of this point, the PMMAfiltration membrane layers should be as thin as possible on the premise of the selectivefiltration of H2.
■
CONCLUSIONSIn summary, we fabricated H2sensors by depositing a Pd NPfilm on interdigital electrodes, and then a PMMA membrane, which acted as a selectivefiltration layer, was spin coated on the surface of the Pd NP film. The sensing behavior of this sensor was investigated.
ACS Applied Materials & Interfaces
We demonstrated that no changes to the electric transport and sensing mechanisms of Pd NPfilms resulted from the presence of the PMMA membrane layer. In addition, the sensors with a PMMA membrane layer showed good selectivity for H2with little decrease in the H2response sensitivity in theβphase region of Pd hydride. It was observed that the sensors exhibited retardation of the sensing response induced by the PMMA membrane layer. As the thickness of the PMMA membrane layer increased, the response time exhibited a tendency toward significant prolongation. On the basis of our experimental data, the retardation of the response time caused by the PMMA membrane layer differs significantly in the three phase regions of Pd hydride. We ascribe the increased response time to three mechanisms: diffusion of H2molecules in the PMMA matrix, nucleation and growth of theβphase in theαphase matrix of the Pd hydride, and stress relaxation at the interface between Pd NPs and the PMMA membrane layer.
■
ASSOCIATED CONTENT*S Supporting Information
The Supporting Information is available free of charge on the ACS Publications websiteat DOI:10.1021/acsami.7b07641.
Schematic diagram and response curve (PDF)
■
AUTHOR INFORMATION Corresponding Author*E-mail:[email protected]. Phone: +86-0714-6576185.
ORCID
Bo Xie:0000-0002-3862-0411
Jun-Ming Liu:0000-0001-8988-8429 Author Contributions
The paper was written through contributions of all authors. All authors have given approval to thefinal version of the paper.
M.C. and P.M. contributed equally to this work.
Notes
The authors declare no competingfinancial interest.
■
ACKNOWLEDGMENTSWe acknowledgefinancial support from the National Natural Science Foundation of China (Grants 61301015, 11604161, and 11627806), the Training Program of the Major Research Plan of the National Natural Science Foundation of China (Grant 91622115), the Natural Science Foundation of Jiangsu Province (Grant BK20160914), and the Promotion Program of Hubei Normal University.
■
(1) Aroutiounian, V. Metal Oxide Hydrogen, Oxygen, and CarbonREFERENCES Monoxide Sensors for Hydrogen Setups and Cells. Int. J. Hydrogen Energy2007,32(9), 1145−1158.(2) Seiyama, T.; Kato, A.; Fujiishi, K.; Nagatani, M. A New Detector for Gaseous Components Using Semiconductive Thin Films.Anal. Chem.
1962,34(11), 1502−1503.
(3) Christofides, C.; Mandelis, A. Operating Characteristics and Comparison of Photopyroelectric and Piezoelectric Sensors for Trace Hydrogen Gas-detection 0.2. Piezoelectric Quartz-crystal Microbalance Sensor.J. Appl. Phys.1989,66(9), 3986−3992.
(4) Dong, S. X.; Bai, F. M.; Li, J. F.; Viehland, D. Sound-resonance Hydrogen Sensor.Appl. Phys. Lett.2003,82(25), 4590−4592.
(5) Dong, S. X.; Bai, F. M.; Li, J. F.; Viehland, D. A Piezoelectric-sound- resonance Cavity for Hydrogen Gas Detection.IEEE Trans. Ultrason.
Ferroelectr. Freq. Control2003,50(9), 1105−1113.
(6) Shivaraman, M. S.; Lundstrom, I.; Svensson, C.; Hammarsten, H.
Hydrogen Sensitivity of Palladium-thin-oxide-silicon Schottky Barriers.
Electron. Lett.1976,12(18), 483−484.
(7) Ito, K. Hydrogen-sensitive Schottky-barrier Diodes.Surf. Sci.1979, 86(JUL), 345−352.
(8) Kim, J.; Gila, B. P.; Chung, G. Y.; Abernathy, C. R.; Pearton, S. J.;
Ren, F. Hydrogen-sensitive GaN Schottky Diodes.Solid-State Electron.
2003,47(6), 1069−1073.
(9) Xie, B.; Zheng, M. Y.; Liu, F.; Peng, X.; Wang, G. H.; Han, M. Fast Response Characteristics of Hydrogen Sensors Based on Pd Nano- particle Films with Controlled Coverage.J. Nanopart. Res.2013,15(6), 1746−1754.
(10) Zeng, X. Q.; Wang, Y. L.; Xiao, Z. L.; Latimer, M. L.; Xu, T.; Kwok, W. K. Hydrogen Responses of Ultrathin Pd Films and Nanowire Networks with a Ti Buffer Layer.J. Mater. Sci.2012,47(18), 6647−
6651.
(11) Li, X. W.; Liu, Y.; Hemminger, J. C.; Penner, R. M. Catalytically Activated Palladium@Platinum Nanowires for Accelerated Hydrogen Gas Detection.ACS Nano2015,9(3), 3215−3225.
(12) Yang, F.; Jung, D.; Penner, R. M. Trace Detection of Dissolved Hydrogen Gas in Oil Using a Palladium Nanowire Array.Anal. Chem.
2011,83(24), 9472−9477.
(13) Yang, F.; Taggart, D. K.; Penner, R. M. Joule Heating a Palladium Nanowire Sensor for Accelerated Response and Recovery to Hydrogen Gas.Small2010,6(13), 1422−1429.
(14) Yang, F.; Kung, S. C.; Cheng, M.; Hemminger, J. C.; Penner, R. M.
Smaller is Faster and More Sensitive: The Effect of Wire Size on the Detection of Hydrogen by Single Palladium Nanowires. ACS Nano 2010,4(9), 5233−5244.
(15) Lee, W.; Lee, J.; Shim, W.; Lee, E.; Noh, J. S. Highly Mobile Palladium Thin Films on an Elastomeric Substrate: Nanogap-based Hydrogen Gas Sensors.Angew. Chem., Int. Ed.2011,50(23), 5301−
5305.
(16) Kiefer, T.; Villanueva, L. G.; Fargier, F.; Favier, F.; Brugger, J. Fast and Robust Hydrogen Sensors Based on Discontinuous Palladium Films on Polyimide, Fabricated on a Wafer Scale.Nanotechnology2010,21 (50), 505501−505506.
(17) Favier, F. Nanogaps for Sensing.Procedia Chem.2009,1(1), 746−749.
(18) Mubeen, S.; Yoo, B.; Myung, N. V. Fabrication of Nanoelectrodes and Nanojunction Hydrogen Sensor.Appl. Phys. Lett.2008,93(13), 133111−133114.
(19) Sun, L.; Chen, M.; Peng, X.; Xie, B.; Han, M. The Effects of Ni Contents on Hydrogen Sensing Response of Closely Spaced Pd-Ni Alloy Nanoparticle Films.Int. J. Hydrogen Energy2016,41(2), 1341−
1347.
(20) Xie, B.; Liu, L. L.; Peng, X.; Zhang, Y.; Xu, Q.; Zheng, M. Y.;
Takiya, T.; Han, M. Optimizing Hydrogen Sensing Behavior by Controlling the Coverage in Pd Nanoparticle Films.J. Phys. Chem. C 2011,115(32), 16161−16166.
(21) Xie, B.; Zhang, S.; Liu, F.; Peng, X.; Song, F.; Wang, G.; Han, M.
Response Behavior of a Palladium Nanoparticle Array Based Hydrogen Sensor in Hydrogen-Nitrogen Mixture.Sens. Actuators, A2012,181(0), 20−24.
(22) Favier, F.; Walter, E. C.; Zach, M. P.; Benter, T.; Penner, R. M.
Hydrogen Sensors and Switches from Electrodeposited Palladium Mesowire Arrays.Science2001,293(5538), 2227−2231.
(23) Walter, E. C.; Favier, F.; Penner, R. M. Palladium Mesowire Arrays for Fast Hydrogen Sensors and Hydrogen-actuated Switches.
Anal. Chem.2002,74(7), 1546−1553.
(24) Yun, M.; Lee, C.; Vasquez, R. P.; Penner, R.; Bangar, M.;
Mulchandani, A.; Myung, N. V. Electrochemically Grown Single Nanowire Sensors.Proc. SPIE2004,5593, 200−206.
(25) Yun, M. H.; Myung, N. V.; Vasquez, R. P.; Lee, C. S.; Menke, E.;
Penner, R. M. Electrochemically Grown Wires for Individually Addressable Sensor Arrays.Nano Lett.2004,4(3), 419−422.
(26) Im, Y.; Lee, C.; Vasquez, R. P.; Bangar, M. A.; Myung, N. V.;
Menke, E. J.; Penner, R. M.; Yun, M. H. Investigation of a Single Pd Nanowire for Use as a Hydrogen Sensor.Small2006,2(3), 356−358.
ACS Applied Materials & Interfaces
(27) Yang, F.; Kung, S. C.; Taggart, D. K.; Penner, R. M. Hydrogen Sensing with a Single Palladium Nanowire.Sens. Lett.2010,8(3), 534− 538.
(28) Yang, Y. A.; Taggart, D. K.; Cheng, M. H.; Hemminger, J. C.;
Penner, R. M. High-throughput Measurement of the Seebeck Coefficient and the Electrical Conductivity of Lithographically Patterned Polycrystalline PbTe Nanowires.J. Phys. Chem. Lett.2010, 1(20), 3004−3011.
(29) Miguel, C. V.; Mendes, A.; Tosti, S.; Madeira, L. M. Effect of Co And Co2on H-2 Permeation through Finger-like Pd-Ag Membranes.
Int. J. Hydrogen Energy2012,37(17), 12680−12687.
(30) Hulme, J.; Komaki, M.; Nishimura, C.; Gwak, J. The Effects of Gas Mixtures on Hydrogen Permeation through Pd-Ag/V-Ni Alloy Composite Membrane.Curr. Appl. Phys.2011,11(4), 972−975.
(31) Chabot, J.; Lecomte, J.; Grumet, C.; Sannier, J. Fuel Cleanup System-poisoning of Palladium-Silver Membranes by Gaseous Impurities.Fusion Technol.1988,14(2), 614−623.
(32) Wang, D.; Flanagan, T. B.; Shanahan, K. L. Permeation of Hydrogen through Pre-oxidized Pd Membranes in the Presence and Absence of Co.J. Alloys Compd.2004,372(1−2), 158−164.
(33) Rego, R.; Caetano, N.; Mendes, A. Hydrogen/Methane and Hydrogen/Nitrogen Sensor Based on the Permselectivity of Polymeric Membranes.Sens. Actuators, B2005,111-112, 150−159.
(34) Kumar, S.; Sharma, A.; Tripathi, B.; Srivastava, S.; Agrawal, S.;
Singh, M.; Awasthi, K.; Vijay, Y. K. Enhancement of Hydrogen Gas Permeability in Electrically Aligned MWCNT-PMMA Composite Membranes.Micron2010,41(7), 909−914.
(35) Patricio, P. S. O.; de Sales, J. A.; Silva, G. G.; Windmoller, D.;
Machado, J. C. Effect of Blend Composition on Microstructure, Morphology, and Gas Permeability in PU/PMMA Blends.J. Membr. Sci.
2006,271(1−2), 177−185.
(36) Tung, K. L.; Lu, K. T. Effect of Tacticity of PMMA on Gas Transport through Membranes: MD and MC Simulation Studies.J.
Membr. Sci.2006,272(1−2), 37−49.
(37) Hong, J.; Lee, S.; Seo, J.; Pyo, S.; Kim, J.; Lee, T. A Highly Sensitive Hydrogen Sensor with Gas Selectivity Using a PMMA Membrane-coated Pd Nanoparticle/Single-layer Graphene Hybrid.
ACS Appl. Mater. Interfaces2015,7(6), 3554−3561.
(38) Kim, W.; Jang, B.; Lee, H.-S.; Lee, W. Reliability and Selectivity of H-2 Sensors Composed of Pd Film Nanogaps on an Elastomeric Substrate.Sens. Actuators, B2016,224, 547−551.
(39) Werman, Y.; Kivelson, S. A.; Berg, E. Non-quasiparticle Transport and Resistivity Saturation: a View from the Large-N Limit.npj Quant.
Mater.2017,2(1), 7.
(40) Suvakov, M.; Tadic, B. Modeling Collective Charge Transport in Nanoparticle Assemblies. J. Phys.: Condens. Matter 2010, 22 (16), 163201−163223.
(41) Beloborodov, I. S.; Lopatin, A. V.; Vinokur, V. M.; Efetov, K. B.
Granular Electronic Systems.Rev. Mod. Phys.2007,79(2), 469−518.
(42) Mott, N. F. Conduction in Glasses Containing Transition Metal Ions.J. Non-Cryst. Solids1968,1(1), 1−17.
(43) Cizek, J.; Melikhova, O.; Vlcek, M.; Lukac, F.; Vlach, M.;
Prochazka, I.; Anwand, W.; Brauer, G.; Mucklich, A.; Wagner, S.;
Uchida, H.; Pundt, A. Hydrogen-induced Microstructural Changes of Pd Films.Int. J. Hydrogen Energy2013,38(27), 12115−12125.
(44) Pundt, A.; Kirchheim, R. Hydrogen in Metals: Microstructural Aspects.Annu. Rev. Mater. Res.2006,36, 555−608.
(45) Wagner, S.; Pundt, A. Quasi-thermodynamic Model on Hydride Formation in Palladium-Hydrogen Thin Films: Impact of Elastic and Microstructural Constraints.Int. J. Hydrogen Energy2016,41(4), 2727−
2738.
(46) Lukac, F.; Vlcek, M.; Vlach, M.; Wagner, S.; Uchida, H.; Pundt, A.;
Bell, A.; Cizek, J. Stress Release during Cyclic Loading of 20 nm Palladium Films.J. Alloys Compd.2015,645, S450−S453.
(47) Vlcek, M.; Lukac, F.; Vlach, M.; Wagner, S.; Uchida, H.; Baehtz, C.; Shalimov, A.; Pundt, A.; Cizek, J. Influence of Microstructure and Mechanical Stress on Behavior of Hydrogen in 500 nm Pd Films.J.
Alloys Compd.2015,645, S446−S449.
(48) de Sales, J. A.; Patricio, P. S. O.; Machado, J. C.; Silva, G. G.;
Windmoller, D. Systematic Investigation of the Effects of Temperature and Pressure on Gas Transport through Polyurethane/Poly- (Methylmethacrylate) Phase-separated Blends. J. Membr. Sci. 2008, 310(1−2), 129−140.
(49) Yang, F.; Taggart, D. K.; Penner, R. M. Fast, Sensitive Hydrogen Gas Detection Using Single Palladium Nanowires That Resist Fracture.
Nano Lett.2009,9(5), 2177−2182.
(50) Suleiman, M.; Jisrawi, N. M.; Dankert, O.; Reetz, M. T.; Bahtz, C.;
Kirchheim, R.; Pundt, A. Phase Transition and Lattice Expansion during Hydrogen Loading of Nanometer Sized Palladium Clusters.J. Alloys Compd.2003,356-357, 644−648.
(51) Ingham, B.; Toney, M. F.; Hendy, S. C.; Cox, T.; Fong, D. D.;
Eastman, J. A.; Fuoss, P. H.; Stevens, K. J.; Lassesson, A.; Brown, S. A.;
Ryan, M. P. Particle Size Effect of Hydrogen-induced Lattice Expansion of Palladium Nanoclusters.Phys. Rev. B: Condens. Matter Mater. Phys.
2008,78(24), 245408−245413.
(52) Manchester, F.; San-Martin, A.; Pitre, J. The H-Pd (Hydrogen- Palladium) System.J. Phase Equilib.1994,15(1), 62−83.
(53) Wagner, S.; Pundt, A. Electrical Resistivity and Hydrogen Solubility of PdHcThin Films.Acta Mater.2010,58(4), 1387−1394.
(54) Wagner, S.; Pundt, A. Combined Impact of Microstructure and Mechanical Stress on the Electrical Resistivity of PdHcThin Films.Acta Mater.2011,59(5), 1862−1870.
(55) Wagner, S.; Uchida, H.; Burlaka, V.; Vlach, M.; Vlcek, M.; Lukac, F.; Cizek, J.; Baehtz, C.; Bell, A.; Pundt, A. Achieving Coherent Phase Transition in Palladium-Hydrogen Thin Films.Scr. Mater.2011,64 (10), 978−981.
(56) Langhammer, C.; Zhdanov, V. P.; Zoric, I.; Kasemo, B. Size- dependent Kinetics of Hydriding and Dehydriding of Pd Nanoparticles.
Phys. Rev. Lett.2010,104(13), 135502−135505.
ACS Applied Materials & Interfaces
