Graphene oxide dielectric permittivity at GHz and its applications for wireless humidity sensing

Graphene oxide relative dielectric permittivity, both its real and imaginary parts, have been measured under various humidity conditions at GHz. It is demonstrated that the relative dielectric permittivity increases with increasing humidity due to water uptake. This electrical property of graphene oxide was used to create a battery-free wireless radio-frequency identification (RFID) humidity sensor by coating printed graphene antenna with the graphene oxide layer. The resonance frequency as well as the backscattering phase of such graphene oxide/graphene antenna become sensitive to the surrounding humidity and can be detected by the RFID reader. This enables batteryless wireless monitoring of the local humidity with digital identification attached to any location or item and paves the way for low-cost efficient sensors for Internet of Things applications.


Results and Discussions
Extraction of GO relative dielectric permittivity under various humidity conditions through full electromagnetic wave simulation and experimental measurements. The electrical property of GO can be completely characterized by its relative dielectric permittivity, = − [6,13]. There are several classical methods to measure relative permittivity in microwave band, including the transmission line (TL) method, free space method, resonator cavity, etc. [14]. However, all these methods do not suit permittivity measurement for small and thin piece of GO under different humidity environments.
Here, to measure the relative permittivity of the GO layer under various humidity conditions, a resonator circuit was designed (Fig. 1a,) with GO (thickness 30μm ± 2μm) printed on the top of the capacitor area (15mm×8mm) of the resonator (see Method for the details of GO preparation and sample fabrication). In order to extract the relative permittivity, a calibration circuit with exactly the same parameters was prepared, where GO layer was mimicked by a thin dielectric layer of exactly the same thickness as GO with known relative permittivity (see Supporting Materials, Fig. S1b).
Both the GO and the calibration circuits were placed in a hermetic container (2 litre in volume, see Supporting Materials, Fig. S3) in which constant humidity conditions were achieved by placing various saturated salt solutions inside the container. Three phase (vapour-liquid-solid) saturated salt solutions made of different salts were used to create different humid environments with constant RH values as these systems produce a constant vapour pressure over a long period of time [15,16]. The saturated salt solutions used were LiCl (RH-11%), K 2 CO 3 (RH-43%), Mg (NO 3 ) 2 (RH-55%), NaCl (RH-75%) and K 2 SO 4 (RH-98%) aqueous solutions prepared by dissolving excess amount of salts in deionised water.
Before each measurement with each particular salt, the humidity was set to be stabilised for at least 48 hours. All measurements are done at 24°C. When the electrical property of GO (such as its permittivity) changes with humidity, it alters the loading of the resonator and results in a shift of the resonance frequency as well as change of the backscattering phase.
The measured transmission coefficients (S 21 ) for the samples with and without printed GO layer are displayed in Fig. 1b water uptake. For the resonator with GO layer, it can be observed that the resonance shifts to lower frequency and its fractional bandwidth increases as the humidity rises. This reveals that both the real ( ) and the imaginary ( ) parts of the relative permittivity of GO increase as GO absorbs more water.
The resonance frequency of the GO covered resonator, as well as the extracted , and the loss tangent ( = / ) are presented in Fig. 2a 12], whereas the change is much smaller at high frequency. This is probably due to the orientation polarization of absorbed water. At low frequency, the polarization of the water can follow the electrical field direction and hence large permittivity changes as humidity varies. At high frequency, the electrical field direction changes fast so that the polarization of the water can't catch up and hence the dielectric permittivity has relatively smaller change with humidity [17]. Water has dielectric permittivity of ~ 80. As humidity increases, more water will be absorbed by the GO hence higher permittivity [17].

Battery-free Wireless GO sensing enabled by printed graphene RFID
technology. It's well known that GO is sensitive to humidity [7,12]. However, the sensing mechanism proposed here is different to those published works. In this work, the GO layer was directly coated on the graphene radio-frequency identification (RFID) antenna. Instead of using GO capacitor to sense the humidity [7, 12], the phase shift of backscattering signal due to the humidity change was detected by the RFID reader. The GO sensor is battery-free, wireless and fully printable. Battery-free wireless sensing is in the heart of IoTs technology [18,19], allowing collection of information about the immediate state of the object without the need of batteries. Below we demonstrate a battery-free RFID humidity sensor by combining printable graphene RFID antenna with GO coating.
where is the power available at the antenna port, is the antenna impedance. The switching between the two input impedance states Z C1 and Z C2 generates two different currents at the antenna port, which can be calculated as [22]: When the humidity changes, the GO layer on RFID antenna changes its dielectric property.
At high humidity, the ionic conductivity due to the intercalated water increases and even pristine GO becomes conductive but only poorly (mega ohms resistance at 100% RH and  From Fig. 4a, it can be seen that the humidity has clear effects on the backscattered signal phase at typical RFID frequency spectrum from 880 MHz to 920 MHz, which experimentally proves that the backscattered signal contains humidity information. Together with the ID information of the sensing tag, a printed graphene enabled battery-free RFID GO humidity sensing system is presented. As it can be seen from Fig. 4b   ) can be observed. Unambiguously demonstrating the effectiveness of wireless printed graphene enabled battery-free RFID GO humidity detection.
It is worth noticed that the technique used here to detect the humidity change is very different to that employed in other reported printed battery-free UHF RFID sensors [26][27][28]. In those reported works, the minimum power-on-tag was measured and the resonance frequency was then extracted from the minimum power-on-tag. This technique requires the reader to scan the whole allocated UHF RFID frequency spectrum and after-measurement data process to find out the minimum power-on-tag and resonance frequency. In this work, the backscattered signal phase was measured. The advantage of measuring backscattered signal phase is that there is no need for the reader to scan the whole allocated frequency spectrum in order to find out the resonance frequency. As it can be seen in Fig. 4, the humidity change can be simply detected at a single frequency point, which greatly simplifies and speeds up the measurement.

Conclusions
We have experimentally extracted the GO relative dielectric permittivity under various humidity conditions at GHz. The measurement results clearly reveal that the GO dielectric property (relative permittivity, or dielectric constant and loss tangent) changes with the humidity but in a different manner as it does in a couple of MHz or lower frequency. Most distinguishingly, the relative dielectric permittivity does not have large changes (from ~ten to a few thousands [7]) and decreases with decreasing humidity at GHz. Furthermore the dielectric property has been used to design and build a RFID sensing tag which can act as a battery-free wireless humidity sensor, by coating GO layer on top of the printed graphene RFID antenna. Such combination can form bases for future energy harvesting enabled RFID sensors for IoTs applications. Furthermore, backscattered signal phase rather than minimum power-on-tag or resonance frequency has been used to detect the humidity change, which can significantly simplify and speed up the monitoring process. As it can be seen, the three-layer structure is obvious and clear -GO layer, printed and compressed graphene layer and paper substrate, stacked in sequence from top to bottom.  To extract the relative permittivity of the GO under various humidity conditions, GO was mimicked by thin dielectric layer, which has exactly the same size, thickness and location as that shown in Fig. S1 (b). The transmission coefficients of the resonator of 5 different sets of relative permittivity ( = ′ − i ′′ ) were simulated and shown in Fig. S2. Each colour set of the curves in the figure contains the same real part ( ′ ) but various imaginary part ( ′′ ) of the relative permittivity. It can be observed that for the same ′ , the resonance frequency changes little with ′′ . This is because ′′ , which is related to material loss tangent ( = ′′ / ′ ), mainly affects the Q factor of the resonator. The simulations reveal that the changes of relative permittivity pose obvious alteration on the resonator's transmission performance. GO permittivity can be extracted by comparing the experimental measurements and full electromagnetic wave simulations.    GO samples were completely dried inside a glove box before exposing to different humidity for 5 days. Mass uptake monotonically increases with increasing humidity and it varies from ~ 5% to 50%.  Relative humidity (a.u.) Fig. S6 shows the durability test of the resonator coated with GO for 43% RH to 98% RH.

Method
The first measurement was taken before drafting the manuscript in December 2015. The second measurement was made in January 2017. The time span is 13 months. All the measurement data collected were after 96 hours of humidity equilibration time. The x axis is the resonance frequency which indicates the frequency response to the GO coated resonator.
The y axis is the transmission coefficient which indicates the propagation of the GO coated resonator. It can be seen that the measured results agree well with the previous data in both frequency response and the propagation level.