Terahertz (THz) waves generally denote electromagnetic waves with a frequency of between 0.1 and 10 THz and a wavelength between 0.03 and 3 mm. This spectral region is the transition region from macro electromagnetic theory to microscopic quantum theory, and this is the only region of the electromagnetic spectrum that has not been studied well so far [1, 2]. As for emerging spectroscopic techniques, many theoretical studies have shown that the vibrational and rotational levels of most biomolecules are located in the THz band and can therefore be used to study the intramolecular modes of low-frequency chemicals [3-6]. Li et al. [7] proposed using contactless measurements of leaf water content based on THz spectrometry to verify the feasibility of the THz method based on different drought stress treatments, cultivation substrates and changes in stomatal conductance. Lian et al. [8] constructed a sub-interval PLS (sub-PLS) model based on a partial THz spectrum and established a method for quantitatively analyzing trans fatty acids in cooked soybean oil using THz spectroscopy. Nakajima et al. [9] measured the THz spectra of natural, amorphous, dry starch from maize and potato sources using the Fourier transform; the results suggested that a Fourier transform THz spectrometer could be used to measure the crystallinity of starch. Wang et al. [10] proposed a rapid, accurate, qualitative, and quantitative method for chiral drug research based on linear-polarization THz spectroscopy and imaging technology. The results had a recognition accuracy of 100% and are of great significance for the rapid, accurate, and non-destructive identification of chiral drugs. Feng et al. [11] used a THz time-domain spectroscopy (THz-TDS) system to obtain the spectra change of hog and sheep sausage casings before and after casing modification. The results showed that the THz-TDS system could be used to determine parameters such as the water content, swelling ratio and thickness of original and modified casings. The results of the above studies show that THz technology has the potential for application in a wide range of different fields. However, at present, THz technology is mainly used to study solid samples and is rarely used to study liquid samples.

Along with the development of THz technology, microfluidic technology has also developed rapidly. Microfluidic techniques can be used to transport, mix, separate, and detect liquids in a highly selective and sensitive way. The first microfluidic chip was manufactured by Terry et al. [12] in 1979. As the technology has continued to develop, various types of microfluidic devices have been produced, and these show great potential for application to the THz spectrum [13, 14].

Graphene oxide (GO) is a promising graphene-based material that has received much attention over the past two decades due to its simplicity and scalability and because it is inexpensive to synthesize [15]. This material has extensive applications in many fields, including biology and the production of heat dissipation materials, energy storage devices, and alloys. In addition, due to its unique electronic transport characteristics, GO has good prospects for application in the THz band. GO is a derivative of graphene; However, in contrast to the nearly perfect structure of graphene, GO contains a large number of aerobic groups. Through the interaction of hydroxyl and carboxyl groups with functional groups of other substances, it can obtain ordered composites perpendicular to the surface and functional composites with various suspension structures. The behavior of functional graphene changes greatly with the degree of oxidation. Due to its high specific surface area and because its surface is rich in functional groups, GO has wider application prospects than graphene and has become a new focus of research in the field of materials science [16, 17].

In this study, THz and microfluidic chip technology were combined to study the THz absorption characteristics of a GO aqueous solution. Using a THz-TDS system, an analysis of the THz time-domain spectrum and frequency-domain spectrum of a microfluidic chip injected with samples was made, and the association between oxygen-containing groups of GO molecules and water at different concentrations of the solution and under applied electric and magnetic fields was explored. The results form a basis for further studies of the physicochemical properties and reaction mechanisms of GO.

The molecular structure of GO is shown in Fig. 1. If a single layer of material is stripped away from the GO, the material can be stabilized in an aqueous solution or polar solvents due to the large number of oxygen-containing groups found on its surface and edges. Following oxidation, GO maintains the layered structure of graphite, but many oxygen-based functional groups are introduced into the graphene monolith in each layer. The introduction of these oxygen-based functional groups makes the simple graphene structure very complex. Because GO is easy to graft modification, it can be combined with other composite materials in situ, thus possibly enhancing the electrical and thermal conductivity of the composite material.

Figure 1. The molecular structure of graphene oxide.

In this study, GO was prepared by the simplified Hummer method using graphite flakes. First, 4 mg of graphite flakes were added to a sulfuric acid/phosphoric acid solution. After the solution was stirred, 15 mg of solid potassium permanganate was added. The solution was then stirred again and left to stand for 3 days until the graphite was fully oxidized and the color of the mixture changed from dark purplish green to dark brown. Hydrogen peroxide solution was then added; this caused the color of the solution to change to bright yellow, indicating that the sample was fully oxidized. Finally, GO was formed by washing the sample with hydrochloric acid and distilled water several times [18]. The sample that was produced measured several microns across, and the thickness of the lamella was between 0.7 and 1.8 nm. This method of preparing GO has a high conversion rate; It is also safer than the conventional method, and there is little possibility of an explosion occurring. The prepared GO was mixed with water to produce aqueous solutions of GO with concentrations of 2, 6, 8, and 10 mg/ml. After the solutions were left standing for a long time, precipitation or agglomeration of GO occurred at the bottom of the container. This could be eliminated by using ultrasound for 15-30 min without affecting the properties of the product.

3.1. Experimental Optical Path

The THz–TDS system used in this experiment is shown in Fig. 2. The laser is a self-locked mode fiber femtosecond laser independently developed by Peking University with a central wavelength of 1,550 nm, a pulse width of 75 fs, an output power of 130 mW and a pulse repetition frequency of 100 MHz. The femtosecond laser pulse is divided into two beams after passing through a polarization beam splitter. One of the beams enters a fiber photoconductive antenna through a fiber coupler to generate THz waves. Another beam passes through a fiber optical coupler and enters a fiber photoconductive antenna to detect THz waves. The microfluidic chip with a liquid sample is placed between two off-axis parabolic mirrors. In order to avoid the influence of water vapor in the air on the experimental results, our experiment was conducted in an environment filled with nitrogen. The signal output by a THz detection antenna is amplified by a lock-in amplifier and then input to a computer for data processing to obtain the THz time domain spectrum, and then obtain the frequency domain spectrum after Fourier transform.

Figure 2. Diagram of the experimental optical path.

3.2. Preparation of the Microfluidic Chip

Cycloolefin copolymer (COC) is an ideal material for microfluidic chips because its transmission to THz waves is more than 90% [19], and it has good hydrophobicity and high mechanical strength. Microfluidic chips include a substrate, coverslips, and microchannels. In this study, two COC plates were used as the substrate of the microfluidic chips and cover plates with a size of 40 mm × 35 mm × 1.95 mm were used. A 50-m-thick strong sticky double-sided adhesive was used to bond the microfluidic chip, and a hollow concave-shaped microfluidic channel was carved in the middle of the double-sided adhesive. In addition, 2 mm inlet and outlet holes were made in the cover plate. The microfluidic chip fabrication process is shown in Fig. 3. In the experiment, the inlet and outlet holes of the microfluidic chip were sealed to prevent concentration changes caused by water evaporation in the sample.

Figure 3. Fabrication of the microfluidic chip.

3.3 External Electric Field Device

In the experiment, a DC high voltage power supply (dw-p153-05c51) was used to provide the voltage, and two metal plates (8 cm × 5 cm × 0.3 cm) were connected to the output end of the high voltage power supply. The microfluidic chip was fixed with a plastic card slot and the chip was placed between two metal plates. The microfluidic chip in the experiment is much smaller than the size of two metal plates, so the microfluidic chip is in a uniform electric field. The distance between the two parallel metal plates is 5 cm, the output voltage of the high voltage power supply is adjusted to 6,000 V, and the uniform electric field strength between the two metal plates is 1,200 V/cm. A schematic diagram of the experimental structural device of the applied electric field used in this experiment is shown in Figs. 4 and 5. Figure 4 shows the electric field system with the electric field direction perpendicular to the direction of THz transmission. Figure 5 shows the electric field system with the electric field direction parallel to the THz transmission direction. During the experiment in Fig. 5, the electrode plates on both sides of the microfluidic chip need to be removed to ensure the passage of THz waves.

Figure 4. Diagram of the vertical electric field system.

Figure 5. Diagram of the parallel electric field system.

3.4. External Magnetic Field Device

Because many substances under the condition of an external magnetic field will show unique characteristics, the experiment adopted the diameter of an 8 cm circular electromagnet to provide the magnetic field, by WYJ-9b power supply to the electromagnet (voltage regulation range of 0-35 V), through the adjustment of the electromagnet power supply voltage to change the magnetic field. The experimental system is shown in Fig. 6. The microfluidic chip with the sample is placed parallel between two electromagnets, and THz waves pass through the hole in the middle. In order to prevent the influence of the heat generated by the electromagnets on the microfluidic chip, we used two fans to blow air at the two electromagnets to cool them.

Figure 6. Diagram of the magnetic field system.

4.1. THz Characteristics of GO Solutions with Different Concentrations

The aqueous solutions of GO with concentrations of 2, 6, 8, and 10 mg/ml that had been prepared using deionized water were injected into the microfluidic chip in separate experiments. A THz-TDS system was used to study the THz transmission characteristics of the solutions. The THz time-domain spectra and Fourier transform frequency-domain spectra, transmission spectra, and absorption coefficient spectra that were obtained are shown in Fig. 7. From these, it can be seen that as the concentration of the GO solution increases, its THz absorption becomes weaker.

Figure 7. The graphene oxide (GO) solutions with different concentrations. (a) Time-domain spectra, (b) frequency-domain spectra, and (c) absorption coefficients.

4.2. THz Characteristics of GO Exposed to an Electric Field

The GO solution with a concentration of 2 mg/ml was injected into the microfluidic chip to form a 50-μm liquid film. The chip was then placed into the THz-TDS system. Two parallel metal plates were placed on both sides of the microfluidic chip as shown in Fig. 4, and the output voltage was adjusted to produce an electric field of 1,200 V/cm, which was maintained for 5 min. The THz transmission spectrum was then measured. In order to restore the sample to its original state as far as possible, the power was turned off and the solution was left to stand for 5 min. The electric field was then applied again and the same measurements were made after 10 min. These steps were repeated with the electric field applied for 15 min and then 20 min. The THz time-domain spectrum, frequency-domain spectrum and absorption coefficient spectrum of the solution obtained using the THz-TDS system are shown in Fig. 8. With the vertical electric field applied for 5 min, the THz intensities of the time-domain and frequency-domain spectrum were significantly lower than for the case of no electric field. As the length of time for which the electric field was applied increases from 5-20 min, the intensity of the THz spectrum gradually increases but remains lower than the intensity when there is no applied electric field.

Figure 8. A graphene oxide (GO) solution in an applied vertical electric field. (a) THz time-domain spectra, (b) frequency-domain spectra, and (c) absorption coefficients.

We then carried out similar experiments with the electric field applied parallel to the microfluidic chip. To do this, the metal plates were placed at the two sides of the microfluidic chip as shown in Fig. 5. Because the metal plate then blocked the optical path, it was impossible to directly measure the THz spectrum. To make the measurements, the power supply was first disconnected and the metal plates removed so that the THz spectrum could be measured. The other steps were the same as those for the vertical electric field. The results show that the parallel electric field had basically the same effect on the THz absorption coefficients as the vertical electric field, and it was therefore concluded that the effect of an electric field on the transmission spectrum is independent of its direction.

4.3. THz Characteristics of GO Exposed to a Magnetic Field

In the next experiment, the microfluidic chip was placed in a magnetic field that was varied by changing the applied voltage; the chip was left in the field for five different lengths of time. First, 2 mg/ml of the GO solution were added to the microfluidic chip. The chip was then placed between the two magnets of the THz-TDS system and the voltage was adjusted to produce a magnetic field with a strength of 10 mT. The chip was left in this position for 5 min, and the THz spectrum was measured. The power supply was then turned off and the chip left for a further 5 min to allow the sample to return to as close to its original state as possible. The magnetic field was reapplied, this time for 10 min. The experiment was repeated with the magnetic field applied for 15 and then 20 min. The THz time-domain spectrum, frequency-domain spectrum, and absorption coefficient spectrum were obtained for each case, and the results are shown in Fig. 9. It can be seen that the THz absorption increases as the length of time that the microfluidic chip is left in the magnetic field increases.

Figure 9. A graphene oxide (GO) solution in an applied magnetic field. (a) THz time-domain spectra, (b) frequency-domain spectra, and (c) absorption coefficients.

In order to confirm the reliability of the results described above, we repeated the above experiment three times and analyzed the error when the magnetic field was applied for 5 min and 10 min. As shown in Fig. 10, for a magnetic field applied for either 5 or 10 min, the standard deviation in the repeated measurements is less than the difference between the results for the two lengths of time. Therefore, it was concluded that the error in the measurements had little influence on the experimental results and that the experimental data are valid. This applies to all of the experimental data obtained in this study.

Figure 10. Error analysis for repeated experiments using the same sample placed in a magnetic field.

The intermolecular hydrogen bond is one of the key factors affecting the THz spectral absorption of solutions [20]. The surface and edge of GO include large numbers of oxygen-containing functional groups, such as carboxyl, carbonyl, epoxy, and hydroxyl groups. The unshared lone electrons in the outer orbital of the oxygen atoms in these groups can combine with the partially positively charged hydrogen in water molecules by forming hydrogen bonds through electrostatic attraction [21]. The THz spectra and absorption spectra of GO solutions with different concentrations show that the transmission of the THz waves increased as the concentration of the solution increased. This is because when the concentration of the solution is low, the interaction between the oxygen-containing groups in the GO and the water molecules is strong, and a large number of bonds is formed. As the concentration of the GO solution increases, the molecular layers become intertwined due to being insufficiently extended; This increases the interaction between the molecular layers and to some extent blocks the association between the oxygen-containing groups and water, leading to a reduction in the number of hydrogen bonds between these groups and the water. In addition, as the concentration of the solution increases, the proportion of deionized water decreases, which also reduces the absorption of THz waves to some extent. Thus, as a result of the effects of intermolecular hydrogen bonding, the absorption of THz by a GO solution decreases as the concentration of the solution increases.

In the experiments described above, a vertical electric field was applied to a GO solution with a fixed concentration for different lengths of time. The THz spectra of GO solution obtained shows that the absorption of THz reaches the maximum after applying an electric field for 5 minutes, that is, the transmission intensity reaches the minimum. The THz absorption gradually decreased as the length of time that the solution was placed in the electric field increased from 5 to 20 min but remained higher than the absorption when there was no electric field. When the direction of the applied electric field was parallel to the THz transmission direction, similar results were obtained. In this experiment, the THz absorption is affected by the molecular spacing and the existence of hydrogen bonds, so the obtained spectra can be interpreted from these two perspectives. Ding et al. [22] demonstrated that an applied electric field can reduce the amount of hydrogen bonding in water molecules and increase the diffusion coefficient. Under the action of an external electric field, water molecules become polarized; the dipole moment changes along the direction of the external electric field, and some hydrogen bonds in the water will be destroyed. According to this theory, the THz absorption should then decrease gradually with time throughout the experiment. However, the experiments described above showed that the THz absorption increased significantly within 5 min of the electric field being applied, indicating that this phenomenon is due to the effect of the electric field on GO and that the influence of the molecular spacing is dominant. When there is no electric field, GO molecules are in equilibrium, the molecular spacing is large, and THz waves can easily penetrate. The oxygen-containing functional groups of GO and the defects in the basic plane of graphene cause GO to be divided into clusters of different sizes [23]. When the electric field is applied, the polar sp³ carbon clusters will produce a polarization effect [24], which will change the charge distribution in the polar groups and generate an induced dipole moment:

(1)$${\mu }_{ind}=-\alpha \cdot E$$

Where μ_ind is the induced dipole, E is the applied electric field strength, and α is the induced polarizability. The larger the induced polarizability is, the more obvious the polarization effect is [25]. In order to balance the electric field force, the electrostatic repulsion, and the electric dipole effect between the particles in the GO molecules, the atoms move closer to each other. As a result, the clustering between molecules rapidly increases, which also leads to an increase in the electrostatic repulsion between the molecules, and a new equilibrium state is quickly reached. Once this state has been reached, the distance between the atoms decreases, which means that the THz waves cannot pass through the solution as easily; This explains the sharp increase in the THz absorption just after the electric field is applied. If the applied electric field intensity remains unchanged, the applied force also remains unchanged, and the atoms in the GO molecules will not move because they have reached a new equilibrium. After 5, 10, 15, or 20 min, the molecular spacing thus remains unchanged, and the influence of the electric field on the hydrogen bonds in the water molecules becomes the dominant effect. According to Ding et al [22], the amount of hydrogen bonding between water molecules in the solution gradually decreases, the absorption of THz waves by the solution also gradually decreases, and the transmission increases.

In the experiment where a magnetic field was applied to a GO solution with a fixed concentration, the THz absorption increased with the increase of the standing time of the solution. Mohammad et al. [26] showed that when water is placed in a magnetic field, the magnetic force does not directly affect the water molecules; However, the internal energy of the molecules increases and their movement intensifies, which leads to the breaking of some hydrogen bonds. As a result, the absorption of THz waves by the water decreases, and the transmission increases. However, our experiment shows that the absorption of THz waves increases under the influence of a magnetic field. This is because under the influence of a magnetic 27, GO exhibits clear magneto-optical effects [27]. The electrons exhibit strong cyclotronic motion, which increases the movement of the electrons and the number of electrons that can move freely. The conjugated π-bonds between the GO layers are then destroyed to a certain degree, causing the original layered, orderly arrangement of GO molecules to become disordered, possibly even leading to the formation of a single-layered structure. As a result, the transmission of THz waves is blocked, that is, the absorption of THz waves is enhanced. As the length of time that the GO remains in the magnetic field increases, the activity of the electrons increases, the GO molecules become more disordered, and the absorption of THz waves increases even more.

In this paper, the THz absorption characteristics of a GO aqueous solution with different concentrations and the effects of applying electric and magnetic fields to a GO solution were studied using a THz-TDS system and microfluidic technology. In the range 0.1-0.6 THz, the THz absorption was found to decrease as the concentration of the GO solution increased. There are two main reasons for this. First, the enhanced interaction between molecular layers prevents the combination of oxygen containing groups in GO with water. Second, the concentration increases, the proportion of water decreases, and the number of hydrogen bonds between water molecules decreases. When an electric field or magnetic field was applied for different lengths of time, the THz absorption showed regular changes that were mainly a result of the response of water molecules and electrons to the external field. This affected the order and clustering ability of the GO molecules, causing changes in the THz absorption. The results of this study provide some guidance for the identification and application of GO. It is also concluded that THz technology can be used to identify GO. The results of this study will thus help to expand the applications of THz technology and lay a foundation for the use of THz technology to study the properties of solutions.

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data underlying the results presented in this paper are not publicly available at the time of publication, but may be obtained from the authors upon reasonable request.

The authors would like to thank Enago for providing English proofreading.

National Natural Science Foundation of China (NSFC 61575131).