Introduction

G) are allotropes of carbon materials with a 2D structure, which have a unique combination of properties such as excellent tribological property, large α (2600 m2/g), high electronic conductive property, extraordinary thermal conductivity, and superior mechanical property [1,2,3,4,5]. These properties suggest the outstanding versatility of G in fillers such as GO, rGO, and FLG for various applications of composites ranging from TIM, structure materials, anti-wear materials, electric conductive materials, and anti-corrosion materials [6,7,8,9,10]. The integration of G is an effective way to boost the performance of composites; however, in many cases, the actual performances of the resulting composites are noticeably lower than the theoretical predictions [11, 12].

The properties intensely respond to the architectures of composites. Therefore, the structural design of composites is a key factor for the improvement of various properties. Research on aligning anisotropic nanofillers (such as the arrangement of nanoscale zinc oxide [11], CNTs [12] and nanofibers [13]) in diverse matrices, thus enabling approximation of the properties of the composites toward the theoretical value, has drawn significant attention and has achieved notable progress in recent years. For example, G is used as an anisotropic 2D material whose properties at IPD are predominantly higher than those of TTD, and it has recently been verified that the orientation control of the distribution of G in matrices can facilitate the resultant properties, thus closing the gap to the theoretical value [14].

In this review, efforts have been devoted to summarizing the fabrications and properties as well as the applications of AGCs. A number of synthetic methods for AGCs have been reviewed, such as the liquid crystal method, the exerting electronic field method, and the exerting magnetic field method. The electrical, thermal, mechanical, anti-corrosion, tribological, and EMI shielding properties of AGCs are mainly discussed, and the reasons for the better performances of AGCs compared to RGCs are analyzed. This review provides guidance and stimulates the development for high-performance AGCs.

Preparations of AGCs

In recent years, a number of methods have been developed for the preparations of various matrices AGCs. The methods reported herein for the preparations of polymer-based AGCs can be divided into six categories: (1) LC, (2) exerting eternal field method (both magnetic field and electronic field), (3) VF, (4) LBL, (5) SEI, and (6) UFC. For metal-based AGCs, there are mainly two ways: (1) RPS and (2) VFSPS. The following provides a detailed discussion of those methods and their influencing factors.

Preparations of polymer-based AGCs

LC

Several of the materials after melting or dissolving, despite the loss of rigidity of solid materials, have obtained liquid flowability, retained the anisotropic orderly crystalline material, and formed an intermediate state with both crystal and liquid properties. The ordered fluid in the process of solid–liquid transformation is called liquid crystal. LC is widely applied to prepare AGCs due to its simplicity. In theory, the G crystalline liquid mainly originates from the ultra-high α and moderate viscosity of the mixtures [15]. Specifically, steric hindrance forms in LMPG and GAD with low viscosity when the flaky G overlaps, and the effect induces a transition from isotropy to a long-range-order liquid crystal structure. In general, researchers use the concentration of G to control the viscosity of mixtures due to the linear relationship between viscosity and concentration. The G critical concentration, which is the lowest concentration for the formation of a liquid crystal, is a central parameter to control the formation of liquid crystal [16]. Aboutalebi et al. quantitatively calculated the G critical concentration by assuming circular G sheets [17]. The G critical concentration can be calculated by Eq. (1):

$$ \rho D^{3} = \frac{{CD^{2} }}{{\frac{\pi }{4}td}} $$
(1)

where D, ρ, C, d, and t represent G diameter, quantity of G of per unit area, concentration of G (in g/m3), mass density of G (in g/cm), and thickness of G, respectively. Theoretical analysis shows that the liquid crystal phenomenon emerges when the ρD3 exceeds the critical value of 4.12. As we can see from Eq. (1), the higher α, the easier the liquid crystal can be formed. Figure 1 shows the optical images of the crystal liquid phase with different GO concentrations in water. With increasing GO concentrations, the GO liquid crystal increases. However, exorbitant concentration is not encouraged because high viscosity hampers G spontaneous stacking. Furthermore, the shape of G, the presence of functional groups attached on G, and the viscosity of solvents also influence the formation of the liquid crystal, because these factors can affect the steric hindrance behavior.

Figure 1
figure 1

(see Ref. [17])

Optical images of liquid crystals with a 0.15 wt%, b 0.2 wt%, c 0.5 wt%, and d 1.0 wt% GO in water.

Following the liquid crystal formation in diverse systems, LMPG curing is conducted by adding curing agents [18], casting [19], and other methods. For instance, a GO liquid crystal formed in chitosan solution at an incorporation of GO above 0.25 wt%; then, the mixtures were casted in molds to obtain AG/chitosan composites [19]. In general, despite the simple procedures of LC, several limitations still exist. (1) Although the macroscopic arrangement of G is perfect (Fig. 2b), the microcosmic distribution of G in matrices is inhomogeneous (Fig. 2d); (2) several parameters, especially the G concentration, are limited at a related narrow range, due to the requirements of critical concentration and viscosity of LMPG.

Figure 2
figure 2

SEM and TEM images of a, c RGCs and b, d AGCs

We summarize the critical concentrations of multifold G forming liquid crystal in diverse systems (in Table 1). In the available reports, the lowest critical concentration (0.1 wt%) of GO was obtained in water [17], due to the small viscosity of the water medium and the large size of GO.

Table 1 Critical concentration of reported G crystalline liquids

EMF and EEF

G is featured as having diamagnetism, a ubiquitous property of materials due to the magnetic response of orbital electrons [27]. This property enables researchers to align G via the theory of Landau diamagnetism [28]; however, relevant studies are still rare because of the requirement of exerting an enormous magnetic field to rotate G. It is an imperfect technique that cannot be applied and engineered at a large scale. To date, anchoring paramagnetic nanoparticles, especially Fe3O4, onto the surface of G is a prevalent practical way to align G under a relatively small external magnetic field.

The EMF (here, the magnetic field is defined as static magnetic field) normally consists of two steps for the preparation of AGCs. The first procedure mainly anchors paramagnetic nanoparticles onto G, then dispersing the magnetic G in liquid polymer solutions to obtain LMPG. Next, the LMPG undergoes curing under the external magnetic field. A schematic of EMF is shown in Fig. 3. The initial G-Fe3O4 would rotate parallel to the orientation of the magnetic induction line. One advantage of the strategy is that is imposes few limitations of G concentration and LMPG viscosity, for which it can be potentially applied for preparations of AGCs with ultra-low and ultra-high G contents. A further method allows very convenient control of the alignment direction of G by regulating the orientation of the magnetic induction line. Moreover, it is preferable for G to load nanoparticles than other anisotropic materials, such as CNTs [30] and Al2O3 [31], due to their ultrahigh α. EMF can not only be used for the preparation of 2D film materials but is also suitable for the synthesis of 3D bulk materials. Typically, G-Fe3O4 hybrid nanoparticles were in situ grown onto G by a coprecipitation method, and the AG/bismaleimides composite was prepared by casting under a 2.0 T external magnetic field [32]. The resultant composites showed excellent tribological properties (friction coefficient as low as 0.07 when the addition of G-Fe3O4 content was 0.6 wt%). In another case, Yan et al. deposited Fe3O4 (about 8 nm) onto G by a modified coprecipitation method, and the liquid mixture of G-Fe3O4 and EP monomer was placed to a 300 mT magnetic field [33]. Finally, AG/EP was obtained by adding a curing agent. The AG/EP composites exhibited obvious anisotropy in thermal conductivity. The increase in thermal conductivity along with IPD was far higher than that in vertical direction. Renteria et al. applied EMF to prepare AG/EP TIM, which led to a decreasing temperature rise by as much as 10 °C in response to the addition of 1 wt% AG in AG direction [34]. In addition to the mentioned applications, the AGCs prepared by EMF were applied in magnetic-controlled switches [35], transparent and flexible electrodes [36], and thermal conductive materials [37].

Figure 3
figure 3

(see Ref. [29])

Mechanism of aligning G-Fe3O4 by the statics magnetic field.

In recent years, nonlinear alignment dynamics have been demonstrated as a strategy for arraying magnetic nanosheets in suspensions at a biaxial orientation under a rotating magnetic field (Fig. 4b) [38, 39]. The composites possess higher anisotropic properties than those prepared by EMF. The rigorous demands for the strategy shown in Fig. 4a, such as the frequency of the magnetic field and the rheological properties of the LMPG, require research into the alignment of G by rotating magnetic field; however, no such studies have been published to date.

Figure 4
figure 4

(in Ref. [38])

a Universal curve shows the dependence of the critical frequency on magnetic field strength, magnetic susceptibility, and fluid viscosity. b The cross-section morphology of biaxial alumina platelet composites.

Nanoparticles with high electron mobility are easily polarized and generate orientation torsion under the induction of an electric field, followed by alignment along the direction of parallel to electric field [40]. The essential metrics of G and rGO, high electron mobility, and large α, are available for preparing AGCs by EEF. Both DC or AC can achieve the alignment of G in liquid; however, G is inclined to electrophoretic alignment in DC electric field resulting in their aggregation near electrodes [41]. Therefore, AC is often used to induce G orientation. Wang presented an analytical model to predict the electric field-induced alignment of cantilevered G nanoribbons and showed that G can be easily arrayed than CNTs under electric fields [42]. Experimentally, AG/EP composites were prepared by exerting an AC electric field (25 V/mm, 10 kHz), and the as-prepared composites exhibited 7–8 orders of magnitude improvement in the electrical conductivity and achieve an increase in the mode I fracture toughness of nearly 900% [41]. Pang et al. prepared AG/polystyrene composites by EEF in the process of annealing and verified that G formed the conductive network easier than CNTs [43]. Compared to EMF, EEF is a relatively rare technique for preparing AGCs, because the high experimental equipment and energy consumption requirements restrict its large-scale development and application.

VF

G and CNTs inherit the major feature of 2D and 1D structures, respectively, which equip them with facile fabrication of anisotropic stacked structure by the flowing of solvents [44, 45]. For G, the VF is not only suitable for the preparation of 2D AG membrane composites, but is also available for preparing 3D bulk AGCs. VF consists of two approaches shown in Fig. 5. Approach (i) obtains G paper by filtrating GAD and then penetrating aimed polymers monomers into G layers of G paper by soaking the G paper in the polymer monomer liquid. Finally, ACGs are obtained by polymer monomer polymerization. Typically, G was dispersed in a mixture of water and ethanol (at a volume ratio of 1:4), and then, the mixture was vacuum-filtrated with a PTFE membrane to obtain the G paper. Next, the G paper was immersed in a mixture of EP monomer and curing agent to obtain AG/EP composites [44]. With the same procedures, Xia et al. prepared aligned G-diamine composites with excellent performance for the removal of natural organic matter [47]. For the steps of approach (ii), the LMPG is prepared, and following the mixture is filtrated through a porous filter membrane to obtain AGCs. In comparison with approach (i), approach (ii) is widely applied due to its simple preparation steps. For example, Liu et al. prepared AG/PVA by filtrating the mixture of G dispersion and PVA solution [48]. Song et al. applied approach (ii) to obtain flexible film of AG/cellulose composite with excellent anisotropic thermal conductivity [49]. A notable difference between approaches (i) and (ii) is that AGCs with wider content range of G prepared by approach (i), while approach (ii) only can prepare AGCs with ultra-high G content.

Figure 5
figure 5

(see Refs. [44] and [46], respectively)

a Approach (i) and b approach (ii) for the preparation of AGCs by VF

Theoretical studies revealed the influencing factors of aligning G by VF. The higher α, the better the alignment through the “excluded volume” interaction will be. In other words, a high α value is a main factor for the alignment of G [49]. Moreover, the ππ stacking and the van der Waals forces between adjacent G lead to their stacking [50]. The functional degree of G also affects the preparation of AGCs by filtration due to the interaction between adjacent amphiphilic groups on the basal plane [18]. Table 2 lists the recent studies on AGCs prepared by VF. It is worth noting that AGCs with high G contents can be prepared by VF, which is a complementary method of LC.

Table 2 AGCs prepared by VF

LBL

LBL is a simple and environment-friendly preparation technique and has been widely applied for the preparation of 1D, 2D, and 3D nanomaterials [53,54,55]. It offers a powerful tool for the preparation of tailored nanostructures with controlled thickness and functionality. Due to the inherent advantages of LBL, the as-obtained AGCs by LBL have been used for sensors, heterojunction, inkjet printing, hybrid lubricating films, and as anode material for lithium ion batteries [53, 55,56,57,58].

There are mainly two mechanisms for conducting LBL: non-covalent interactions and covalent interactions. Figure 6 shows both mechanisms. (1) A non-covalent interaction forms via intermolecular force between adsorbate and adsorbents, such as ππ stacking and formation of hydrogen bonding. Interestingly, ππ interaction between π-conjugated organic materials and the 2D sp2 carbon network G is a non-modified method in theory, while the poor dispersity and inevitable gathering of unmodified G in solvents would impel G to modify. (2) Covalent interaction is the transfer, exchange, or coexistence of electrons. In contrast to approach (i), a covalent bond is formed between atoms by sharing electron pairs. Due to the good dispersity of graphene derivative in various solvents, LBL, GO, rGO, and FLG are widely applied in LBL to fabricate AGCs. However, depending on the applications of the AGCs, diverse types of G are chosen for LBL. Given the moderate combined electrics of thermotics and mechanic properties, rGO is preferably used in preparation of AGCs for high electric, thermal conductive, and mechanical applications, such as anode materials for lithium ion batteries, sensors, and TIM [55, 58]. For the improvement in mechanical properties of matrices alone, GO, which possesses similar mechanics to rGO and inexpensive cost, is in the leading material in use. The procedures of approach (i) and (2) are identical. Substrate is circularly dipped into GAD and matrix solvents and the cycle times correspond to the assembled layers of G. Therefore, the thickness and performances of the AGCs are controllable by adjusting the assembled times. For instance, the AG/PVA membranes were fabricated by circularly dipping hydroxyl group modified quartz substrates into PVA solution and GO solution [59]. The resulting 300-bilayer AGCs achieved improvements of 98.7% and 240.4% in elastic modulus and hardness, respectively.

Figure 6
figure 6

(see Refs. [54] and [59], respectively)

Mechanisms of a covalent interactions and b non-covalent interactions.

Approaches (i) and (ii) can be used simultaneously in many cases. PDDA and PSS, two types of polyelectrolytes, are prevalently used in terms of LBL [60]. In light of the charge distribution (positive or negative) of the G surface, PDDA (PSS) was selected to adsorb on the surface via electrostatic interaction. In the meantime, the functional groups on G would interact with the macromolecule long chain via non-covalent interactions (dispersion force, induction force, hydrogen bond, and ππ bond). Kirschner et al. demonstrated that the strongest driving forces in LBL are covalent interactions [61].

Table 3 summarizes the various ACGs prepared by LBL in recent years. It shows that preparing 2D films by covalent interactions is the mainstream technique due to the facile control of the thickness by times of replication. The complex 1D and 3D structures of AGCs limit the development of preparing AGCs by LBL. LBL mainly offers a facile way to prepare controllable properties of AG film composites that can be applied in diverse fields.

Table 3 Summary of various AGCs prepared by LBL

SEI

Inspired by the preparation of G paper with layer-by-layer stacked structure [72,73,74], SEI, a method of aligning G by induction of solvents evaporation, has been developed for the synthesis of AGCs. From this point, under the effect of solvent evaporation, G would be subjected to a uniform upward orientation force, which leads to stacks at a near horizontal or vertical manner in the air–liquid interface. The most remarkable advantage of SEI is that the properties, such as thermal conductivity and electric conductivity, can be regulated by precisely controlling the AGCs thickness. Furthermore, the thickness of AGCs can be regulated via the added quantity of G. The facile maneuverability is a further merit of SEI. Based on both advantages, these types of AGCs were widely applied in electrode [75], sensors [76], and FETs [77]. The AGCs prepared by SEI consists of two categories: (1) AGLCs and (2) MGACs.

For the preparation of AGLCs, the procedures start with uniformly coating the substrates with DAG. Then, an AG film forms via evaporation of the solvents. The compression of the AG film follows this stage. Several parameters influence the formation and properties of the AG film. In general, the order and degree of the stacked G layer structure and the integrity of the film increase with increase in addition of G. Additionally, the usages of surfactants are a central element for the formation of AG film on substrates. Research verified that no G films can be obtained when CTAB and SDS are employed as surfactants of G, and small-size AG films formed when SLS is used as surfactant [78]. For instance, Malekpour et al. deposited AG film on the PET substrate by using nonionic polymer-type surfactant dispersed G ink and then compressing the AG film with a compression roller. The resulting composites showed potential application as a photoelectricity technique due to the significant improvements in thermal conductivity along the IPD [77]. However, in many cases, due to the good dispersity of GO, rGO, and FLG in solvents, no surfactants are added. Using similar procedures, Kim et al. synthesized flexibly aligned rGO FETs. The FETs were operated at low voltages of < 2 V with a hole and electron mobility of 214 and 106 cm2/V s, respectively [79]. With regard to the properties of G film, the compression procedure will improve the thermal and electric conductivity because this step can enhance the contact of G. (The detailed mechanisms are discussed in "Properties of AGCs" section.) Compared to the preparation of AGLCs by CVD, the SEI offers large-scale production capability and allows for designing G film directly onto substrates without transfer steps; however, the drawback is the relative low quality of G film due to surface modification, which decreases the intrinsic properties of G.

Similar to the VF, the preparation steps of the MGAs by SEI also include two steps. Preparing the LMPG and evaporating the solvents are both required to obtain the MGAs. Current reports showed that the G content is an important parameter for the control of the formation of AG [78]. G and GD tend to be arranged in disorder when the added amount is low. When the G content is high, the steric hindrance and excluded volume are more obvious, making G tend to stack orderly at the air-liquid interface. In addition to that, the evaporation rate of solvents is vital for the array of G. Low evaporation speed corresponds to a low induction force, which can align G parallel to substrates, whereas a high evaporation rate aligns G vertical to substrates [80]. Experiments showed that AG (parallel to substrate) formed in the G/SBR system at a curing temperature of 30 °C and concentrations of G above 1 wt%; AG (vertical to substrate) formed at a curing temperature above 60 °C [80]. Fan et al. prepared AG/PVDF (AG parallel to substrate) composites by evaporating the LMPG solvent (N,N-dimethylacetamide) under 60 °C. The prepared composites showed a high dielectric constant (7940) when the cooperation of rGO was 0.0177 vol% [81].

UFC

The freeze-casting technique with traits of versatility, facile accessibility, and capability of producing materials with complex shapes and structures, is extensively used to fabricate 3D GA and GA composites [82, 83]. In addition to the afore-mentioned advantages of the freeze-casting technique, the resulting materials with porous structure, ultra-low density, and promising interface connection between G attract tremendous attention [84]. By regulating experimental parameters such as freeze temperature and the degree of super-cooling, a peculiar morphology and structure can be engineered [85]. In particular, the porosity, ultra-low density, and robust AG network and AGCs were designed by a modified freeze-drying process named UFC. A critical step of the UFC is the control of orientated growth of dispersing medium crystallization via high temperature gradient. The formation mechanism of the AG 3D network by UFC is shown in Fig. 7. An ultra-high degree of super-cooling can promote crystal growth along the orientation of the maximum temperature gradient, and can restrain the possible growth of secondary dendrites [86]. Furthermore, the rapid growth of dispersing medium crystallization can remove G and polymer solutions from crystals, and the removal reaction forces G to array parallel to the growth direction of the crystal. In recent years, liquid nitrogen has been employed to generate nearly 200 °C degree of super-cooling. Typically, the bottom of LMPG (DAG) is placed onto liquid nitrogen, while the roof is in contact with room-temperature air. The degree of super-cooling of the dispersing mediums is the difference between the theoretical crystallization temperature and the temperature of liquid nitrogen (~196 °C). Moreover, the concentrations of G also affect the construction of 3D AG structure [87]. Only if the concentration overtops the percolation threshold, G forms an architecture of continuous, well-interconnected, and aligned pores [88]. It has been reported that an AG network would form when the GO concentration of DAG is above 1 mg/mL [89]. The UFC also comprises two approaches to synthesize AGCs. Approach (i): frozen GAD are freeze-dried to obtain an AG aerogel into which the aimed matrices are then filled. Li et al. followed approach (i) prepared AG/EP composites with excellent EMI shielding efficiency [90]. Approach (ii): freezing the LMPG directly followed by freeze-drying to obtain AGCs. In contrast to bulk AGCs from approach (i), AGCs with porous morphology can be prepared by approach (ii). With approach (ii), Vickery et al. prepared porous and ultra-low-density AG/PVA composites with excellent mechanical properties [91], while Wang et al. prepared AG/PVA composites with analogous structures with benign dielectric properties [92].

Figure 7
figure 7

(see Ref. [87])

Mechanism of the UFC method for constructing 3D AG.

Recently, the distinctive 3D AG structure has attracted extensive attention in the preparation of high-performance electrodes because the porous AG structure can provide rapid ion diffusion and high electric conductivity and can buffer the expansion of electrode materials (such as silicon and SnO2) during the transversion of alloying [93]. To date, vertical AG (vertical to the collector) electrodes were synthesized by high-cost PECVD [94]. The AGCs prepared by approach (ii) possess high conductivity, orientation degree of G, and mechanical stability. Therefore, UFC is a promising alternative to PECVD for the preparation of vertical 3D AGCs. Despite the afore-mentioned advantages of UFC, the obvious drawback of the requirement of ultra-high super-cooling in the progress hinders its large-scale application.

Preparations of metal-based AGCs

The above-mentioned methods are for the preparation of AG/polymer composites. In recent years, researchers have developed effective methods to synthesize metal-based AGCs due to their high thermal and dimensional stability and the superior strength of metals compared to polymers. Inspired by nature, biomimetics implement the insights into researchers to array G in metals. The “brick and mortar” structure of natural nacre with nearly 95% hard mineral aragonite as “mortar” and about 5% soft protein as ‘brick’ was simulated to construct metal-based AGCs. Until now, both RPS and VFSPS have been developed mainly to synthesize metal-based AGCs with a “brick and mortar” structure [95, 96]. The procedures of RPS are shown in Fig. 8a. Metals are used to replicate layered porous structure materials, adsorb the G inside of porousm and press the embodied G porous metal structure to obtain AGCs. Xiong et al. employed copper to chemically replicate fir wood with a layered porous structure and the prepared AG/copper composites significantly enhanced the obtained mechanical properties [95]. For the steps of VFSPS, G was adsorbed on the surface of metal micron particles by filtrating a mixture of metal powder and GAD, and then consolidating the bulk powder mixture by SPS (see Fig. 8b). Following the VFSPS, AG/copper composites were obtained with excellent mechanical, electronic, and tribological performance [96,97,98]. In addition to PRS and VFSPS methods, complicated CVDS which consist of two steps was developed to prepared AG/metal composites. For example, Kim et al. and Gao et al. prepared AG/copper composites with excellent mechanics and electronic conductivity [99, 100]. However, CVDS is a rarely used method for the preparation of metal-based AGCs. Most metals (excluding Cu, Ni, and their alloys) are incapable of directly growing G, suggesting that CVDS can prepare Cu-based, Ni-based, and Cu/Ni alloy-based AGCs without transferal of G to the surface of other metals. Moreover, it is difficult to prepare metal-based AGCs with uniformly distributed nano-level AG because the large-size G film forms by CVD.

Figure 8
figure 8

Mechanism of a PRS (Ref. [95]) and b FSPS (Ref. [96]) progresses

In summary, for the preparation of metal-based AGCs reported so far, copper is only used for the matrix due to the following two reasons: Copper powder possesses high chemical stability in the air environment, and it is favorable that copper powder does not form a copper oxide film on the surface of copper particles in the procedures. Furthermore, when using copper oxide powder as precursor, it reduces easily. Due to the disadvantages of the copper matrix, such as high density and poor corrosion resistance in the alkaline environment, more metal matrices should be developed to meet the requirements of applications and industries. Furthermore, the discussion of influencing factors and the establishment of a theoretical model should also be achieved to enable the precise control of preparation procedures.

Properties of AGCs

Electrical conductivity

Based on the electric conductive percolation theory, the conductive network formation of fillers is an important condition for enhancing the electric conductivity of composites. In general, the value of increasing volume fraction of conductive fillers to a critical value and an insulator-to-conductor transition occurs (Pc) and σ are two parameters used for the evaluation of the electric performances of polymer composites. A desirable lower Pc and higher σ at lower addition of G mean that lower consumption of the expensive G can achieve higher electric conductivity performances. For metal-based composites, the percent increase of σ is used as evaluation criterion.

As a conductive nanofiller with ultra-high σ and α, G has been considered as an ideal conductive filler. In particular, the arrangement of G in matrices affects the construction of the conductive network and the conductivity of composites. Because the electric conductivity of G at IPD is higher than that at TTD, the array can fully utilize their electric conductivity [23]. Furthermore, a large body of work indicated that the AG would facilitate the formation of a conductive network. Various matrices, such as PU [22], EP [18], PVDF-HFP [25], cellulose [68], SBR [80], PVDF [81], and copper [99] have boosted electric conductive properties via the alignment of G, and the resulting AGCs demonstrated lower Pc and higher σ in the orientation of the aligned G. For example, the Pc (0.22 vol%) of AG/EP composites was twofold higher than that of RG/EP composites, and the σ of AG/EP composites increased by nearly two orders of magnitude compared to that of RG/EP composites [41]. In addition, experiments showed that the electric properties of RG/EP composites and AG/EP composites at TTD are almost identical [41]. An effective-medium theory of σ for AG/polymer composites has been developed [14], demonstrating that aligning G is a better way to improve the resulting electric properties, as well as achieving the conductive applications of AP/EP composites. The comprehensive electric property of AGCs outperforms that of RGCs.

Several intrinsic properties of G significantly affect the electric properties of AGCs: (1) the presence of functional groups on G. The more functional groups there are, the lower σ will be. Basically, the introduction of groups onto G can damage the inherent sp2 structure, which then rapidly decreases the σ of G. For example, Yousefi et al. showed that aligned GO composites functioned as an insulator even if the addition of GO was as high as 3 wt% [18]. (2) The dispersity of G in matrices. Good dispersity implies that the formation of a conductive network is possible even at low G addition. However, it is well known that G inevitably gathers in matrices without modification. By adsorbing surfactants and anchoring functional groups onto G, G obtains good dispersity, while it decreases their σ. To address this contradiction, rGO, with the have combined performances of moderate dispersity and σ, is widely used in conductive AGCs as a compromising strategy [25, 78, 80]. (3) Size of G. The σ of large-area G is higher than that of small-area G because large-area G possess smaller inter-sheet contact resistance in matrices [101]. Kumar et al. exhibited large-area reduced-AG/PVDF-HFP composites, the σ of the composites was obviously higher than that of small-area reduced-AG/PVDF-HFP composites [25].

Here, we summarize and compare the electric properties of RGCs and AGCs for different directions. As shown in Table 4, the results mostly correspond to the presented viewpoints, although the Pc and σ are strongly influenced by a number of factors. Aligning G into matrices is a more effective way to enhance the electric performance of composites.

Table 4 Electric properties of RGCs and AGCs

Mechanical properties

G is considered to be one class of the emerging enhanced phases for various matrices due to their incomparable mechanics and ultra-high α, that can effectively boost the mechanical performance of matrices when the additional content is low. Figure 9 shows the AG enhancement mechanisms for both polymer-based and metal-based composites. It is reasonable to assume that the introduction of AG is a more effective way to improve the mechanical performances for matrices than RG because flake-like G can bear loads only at the IPD. Furthermore, the arranged G can effectively inhibit the slip of the polymer chains in AG/polymer composites [107] and the plane slip in AG/metal composites [99].

Figure 9
figure 9

(Refs. [107] and [100], respectively)

a AG enhancement mechanism for polymer-based composites and b AG inhibits the plane slip mechanism in metal-based composites.

Therefore, theoretically, the orientation distribution of G in matrices has a pivotal effect on the improvement of mechanics [110]. A large number of studies investigated the array of G in various matrices, such as PI [22], chitosan [19], PU [20, 22], EP [18], Tween [51], and copper [99], to enhance mechanics. To date, because most of the presented studies focused on the increase in tensile strength and elastic modulus, we, therefore, discuss the reinforcement mechanisms of AGCs for tensile strength and elastic modulus.

Based on the Halpin–Tsai model [108], the moduli of AGCs are theoretically calculated by Eqs. (23):

$$ E = E_{\text{m}} \left( {\frac{{1 + \gamma_{1} \beta V_{\text{f}} }}{{1 - \gamma_{1} V_{\text{f}} }}} \right) $$
(2)
$$ \gamma_{1} = \frac{{\frac{{E_{\text{f}} }}{{E_{\text{m}} }} - 1}}{{\frac{{E_{\text{f}} }}{{E_{\text{m}} }} + \beta }},\;\;\beta = \frac{2\alpha }{3} $$
(3)

Here, E, Em, and Ef represent elastic moduli of AGCs, matrices, and fillers respectively, and Vf represent the volume fraction of G in Eqs. (23). Since Ef and α are directly proportional to E, the ultra-high elastic modulus and α of G have intrinsic superiority as oriented enhanced phases. In addition to the afore-mentioned influencing factors, the interfacial interaction between G and matrices also plays a key role in the improvement of mechanics. The stronger interfacial interaction leads to better reliable transfer of the loads between matrices and G. The formations of covalent bonds, hydrogen bonds, and ππ bonds have been applied between matrices and G to enhance the interfacial interaction. Establishing chemical bonds between functional groups of G and polymer monomers during the procedures of polymerization of polymer normally reduces GO. Yousefi et al. expounded the detailed reaction mechanism of GO and EP monomers in procedures of in situ polymerization, and using a modulus as high as 800 MPa resulted in AG/EP with 2 wt% GO, which demonstrated a strong interfacial interaction between GO and EP [18]. The contained oxygen, nitrogen, and fluorine functional groups on G and polymer matrices probably form strong hydrogen bonds. The prime formation mechanism of hydrogen bonds mainly includes highly polar polyalcohols and G in a G/polymer system. Huang et al. explained the formation of hydrogen bonds between hydroxyl of PVA and oxygenated groups of GO by a shift of vibration peak position [109]. G and many polymers with abundant aromatic rings are an important prerequisite for the improvement of the interfacial bonding by ππ stacking, as has been reported by PEDOT and GO [110]. As an unavoidable problem, the dispersity of G in matrices also significant affects the improvement of the resulting mechanics. With increasing G content, the moduli and strength of AGCs showed a first increasing and then decreasing trend in most relevant studies, as a result of the unavoidable gathering at high contents, while leads to a decrease of load transfer.

Table 5 summarizes and compares the elastic modulus and tensile strength of AGCs and RGCs with different G contents by various preparation methods. Although the mechanic improvement in composites is strongly influenced by the matrices, G types, and preparation methods, AGCs possess higher mechanics than RGCs.

Table 5 Elastic moduli and strength of AGCs and RGCs

Thermal properties

In addition to the high phonon λ of honeycomb-like G as a result of the strong covalent bonds, the super-high α and thermal stability, which can minimize the thermal interface resistance, indicate G as an ideal thermal conductive enhanced filler. Similar to the improvement in σ, the G network plays the role of thermal conductive channels in composites. AG can facilitate the formation of the thermal conductive network and noticeably decrease the thermal interfacial resistance between G and matrices. The thermal conductive mechanism is shown in Fig. 10. Renteria et al. reported that the thermal properties of AG/EP composites far exceeded those of RG/EP composites [37]. Moreover, aligning G with anisotropic λ, ultra-high 2000–5000 W/mK along the IPD or 10–20 W/mK along TTD [119], can make the best use of G as a 2D material as well as can provide the possibility for designing AGCs with extraordinary λ at the aligned G orientation. The anisotropic thermal conductive AGCs are hopefully applied in TIM, dissipating materials [37, 52, 120,121,122,123], which have a strong requirement for oriented heat conduction. The thermal properties of AGCs and RGCs are listed in Table 6. The summarized reasons indicate that AGCs have better thermal properties.

Figure 10
figure 10

(Ref. [118])

Thermal conductive mechanism of pure matrix, RGCs, and AGCs.

Table 6 Thermal properties of AGCs and RGCs

Anti-corrosion properties

In the natural environment, H2O, O2, and Cl- are the nemesis of corrosion in most materials. Therefore, isolating the metal interface from the electrolyte is a rational way to protect materials from corrosion by introducing organic coating at the surface of materials [132]. However, the anti-permeability properties of pure organic coatings are poor. The combination with fillers for polymer coatings is an effective method to enhance their anti-corrosion properties. G with high α, chemical inertness, and excellent anti-permeability (only protons can pass through) have inspired considerable work to utilize G/polymer composites as anti-corrosion coatings [133,134,135]. A proven theory indicates that the cooperation of G strongly increases the diffusion pathway of the electrolyte from the surface of coatings to the metal interface [136, 137]. Moreover, experiments have shown that the alignment of G would strongly affect the anti-corrosion properties of coatings because AG (vertical to the direction of electrolyte diffusion) can effectively increase the diffusion pathway. The anti-corrosion mechanisms of AGCs and RGCs are shown in Fig. 11. Luo et al. reported that the AG/EP anti-corrosion property far exceeded that of RG/EP coating [138]. Li et al. synthesized an AG/PU anti-corrosion coating via the LC method, and verified that the anti-corrosion performance of AGCs was higher than that of RGCs vis EIS analysis [21]. Because the anti-permeabilities of G and GO are nearly identical, GO with high processability and dispersibility in aqueous media was widely applied in anti-corrosion coatings [132, 139, 140]. However, reported studies on anti-corrosion of AGCs are limited, due to the undeniable difficulty for the preparation of AGCs on the surface of materials by the afore-mentioned methods.

Figure 11
figure 11

(Ref. [21])

Anti-corrosion mechanisms of a RGCs and b AGCs.

Trinological property

Graphite carbon including G has been found to have outstanding tribological properties [141]. Furthermore, many studies indicated that AG can fully utilize the self-lubricating property of 2D material [32, 142]. Several mechanisms have been developed to explain the excellent tribological properties of AGCs: (1) AGCs can boost the contact area between AG and friction counterparts. (2) AG facilitate the formation of G transform film on friction counterparts under a moderate content of G addition. The schematic is shown in Fig. 12. Both phenomena transfer the wear from between the friction counterparts and matrices to between counterparts and AG. For instance, Liu et al. prepared AG/bismaleimides composites via EMF, and the obtained composites have a friction coefficient of as low as 0.07 for an addition of G-Fe3O4 of 0.06 wt% [32].

Figure 12
figure 12

(Ref. [142])

Tribological schematic of AGCs.

Friction coefficient and volume wear rate are used to evaluate the friction properties. We summarized and compared the friction coefficient and volume wear rate of AGCs and RGCs. As shown in Table 7, the tribological performances of AGCs far exceed those of the RGCs although there are limited reports on the tribological properties of AGCs. Interestingly, additions of rGO and FGS into SSBR-BR have a side effect on the tribological properties of composites [143]. It is plausible that the existence of a disordered alignment of rGO (FSG) in rubber matrices can greatly induce tear, curling, and peeling off on worn surfaces of the composites, emphasizing the importance of AG to enhance the wear performance.

Table 7 Tribological property of AGCs and RGCs

EMI shielding property

Traditional metallic coatings and cases of EMI shields with heavy weight are undesirable for their applications. G (including partial graphene derivatives) with low density, ultra-high electronic and thermal conductivity, and mechanics are indicated as ideal fillers for EMI shielding composites. The EMI shielding behavior of G primary stems from the strong electrical loss of microwaves [146, 147]. Normally, the EMI shielding performance is proportional to the σ value of G composites. In "Electrical conductivity" section, the advantages of AGCs have been discussed in electrics. However, the design of G composites with high dielectric permittivity is a further way to boost the EMI shielding property [148]. In microstructures of polymer-based AGCs, any pair of adjacent conductive G separated by an insulating polymer thin film can serve as a nanoscale capacitor. The enormous AG network, containing millions of capacitors, can hold an extremely large capacity for the storage of electric charges, thus equipping the AGCs with superior dielectric permittivity [149]. A schematic of nanoscale capacitors is shown in Fig. 13a. AGCs combine both merits of excellent electronic conductivity and the formation of numerous nanoscale capacitors, leading to the following two enhancements for EMI shielding performance: (1) capability of absorbing the incidental electromagnetism waves via polarization in numerous nanoscale capacitors; (2) the G aligned in IPD contributes positively to shielding the electromagnetic waves that emanate through the thickness direction by increasing the reflective area (Fig. 13b) [150]. Yousefi et al. prepared an AG/EP composite with 3 wt% of rGO via LC, and the composites exhibited a dielectric constant exceeding 14000 at 1 kHz and an EMI shielding efficiency of about 40 dB in 500–4000 MHz [149], which was far higher than similar RGCs [9, 151, 152]. Song et al. prepared an AG/wax composite via the hot-press method and the obtained AGCs presented above 30% increment in shielding effectiveness based on unaligned G/wax composite [150]. The afore-mentioned studies indicate that the cooperation of AG in polymers is a more effective means to boost the EMI shielding property of G composites.

Figure 13
figure 13

a Schematic of the formation of numerous nanoscale capacitors in G composites by AG. b Enhanced electromagnetic waves reflect and absorb AGCs

Conclusions and prospects

In summary, the methods for preparation and the resulting properties of AGCs have been summarized. With regard to the preparation methods, each has its own merits, such as the simple procedures of LC, VF, and SEI, which limit the G content of EMF, EEF, and VFSPS little, as well as the porous structure of UFC, energy conservation of LC, VF, LBL, and SEI. Researchers can choose the most suitable method for their experiments. With regard to the properties of AGCs, the enhanced mechanisms of AGCs and influencing factors have been discussed in detail. Aligning G in matrices is an ideal method to fully utilize its excellent properties. Based on these outstanding multifunctional properties, AGCs are promising as structural and functional materials (sensors, heterojunction, inkjet printing, hybrid lubricating films, and anode material for lithium ion batteries).

However, several existing issues still need to be addressed. (1) The studies on the influencing factors of the G orientation and alignment are not thorough, especially lacking systematic mathematical models for methods such as VF, SEI, UFC, and VFSPS. (2) The macroscopical order degree of AGCs is imperfect. (3) It remains difficult to prepare more metal-based and ceramic-based AGCs. (4) The current methods for preparing AGCs need to be studied in depth, existing methods need to be constantly improved, and new technologies for the large-scale industrial production need to be developed.