High‑Quality Few‑Layer Graphene Produced by a Simple Needle‑Valve Hydrodynamic Exfoliation

Introduction

Since its isolation in 2004, graphene has attracted sustained interest due to its exceptional electrical, mechanical, and chemical properties [1,2]. Its promise spans electronics [3], photonics [4], catalysis [5,6], energy conversion and storage [7–9], and polymer nanocomposites [10,11]. To harness these advantages, it is critical to develop methods that yield high‑quality graphene at industrial scale.

Current exfoliation strategies include micromechanical cleavage [12], chemical vapor deposition [13,14], solvothermal synthesis [15], chemical exfoliation [16,17], and liquid‑phase exfoliation [18,19]. Liquid‑phase exfoliation is the most practical for large‑scale production, typically driven by ultrasonication. Unfortunately, ultrasonication suffers from limited yield, long processing times, scale‑up difficulties [20,21], and often produces graphene with higher defect densities [22].

Fluid‑dynamics‑assisted liquid‑phase exfoliation offers a promising alternative. Jet cavitation devices [29], high‑shear mixers [26], high‑pressure homogenizers [31], and even kitchen blenders [27,30] have demonstrated the ability to exfoliate graphite with lower defect content. However, these methods usually require intense operating conditions and extended processing times, resulting in graphene with higher Raman D/G ratios (0.14–0.78) and lower aspect ratios [26–33]. Consequently, a more efficient, scalable technique that delivers both high concentration and high aspect ratio graphene is needed.

In this work, we introduce a simple hydrodynamic exfoliation approach that utilizes a needle valve to generate the necessary cavitation and shear forces. Using an 80 wt % N‑methyl‑pyrrolidone (NMP) aqueous solution, we produce few‑layer graphene with superior structural integrity. The flakes are characterized by SEM, TEM, AFM, Raman spectroscopy, and XPS, and the effects of operating parameters on concentration are systematically explored.

Materials and Methods

Materials

NMP (purity 99.5 %) and graphite powder (≤ 325 mesh, purity 99.9 %) were sourced from Aladdin Industrial Corporation (Shanghai, China). Deionized water was purified using a laboratory water purification system (SZ‑97A, Shanghai, China).

Exfoliation of Graphite into Few‑Layer Graphene Flakes

The exfoliation apparatus is illustrated in Fig. 1. As graphite‑laden liquid passes through the narrow gap of the needle valve, abrupt velocity and geometric changes create cavitation and velocity gradients, which impart normal and shear forces that delaminate graphite layers. By adjusting the valve opening, the working pressure can be precisely controlled.

The typical procedure involves dispersing graphite powder in 80 wt % NMP aqueous solution to a concentration of 10 mg mL⁻¹. The suspension is pumped through the needle valve at 20 MPa for 16 cycles using a plunger pump (model 2‑JW, Zhijiang Petrochemical). After exfoliation, the dispersion is centrifuged at 500 rpm for 60 min (SC‑3610, USTC Zonkia) to remove unexfoliated graphite, and the supernatant is collected for further analysis.

A schematic view of the needle valve

Schematic diagram of the hydrodynamic‑assisted exfoliation process

Characterization

Morphology and size were examined by SEM (VEGA3, TESCAN) after gold coating in an argon atmosphere. TEM (Tecnai G2 F30 S‑Twin, 300 kV) images were obtained by depositing droplets on holey carbon grids. AFM (Bruker Dimension Icon, tapping mode) images were captured on freshly cleaved mica substrates. Raman spectra were recorded with a Lab RAM HR800 (λ = 532 nm) at room temperature. XPS (ESCALAB 250Xi) quantified oxidation states. UV‑Vis absorption (Lambda 35, PerkinElmer) at 660 nm was used to determine graphene concentration.

Results and Discussion

Quality of Graphene Flakes

Figure 3 compares SEM images of bulk graphite and the exfoliated product. Bulk graphite exhibits flake‑like morphology with lateral dimensions of 5–20 µm and thicknesses around 10 µm. In contrast, the exfoliated powder consists of much thinner flakes (lateral size 1–7 µm) and sub‑micron thicknesses, confirming effective delamination. Some flakes display folded edges, indicative of mono‑ or few‑layer graphene [26].

SEM images of a bulk graphite and b exfoliated graphene powder

Transmission electron microscopy provides direct evidence of layer number. Figure 4a shows a mono‑layer flake with a clearly folded edge; its high‑resolution image (Fig. 4b) reveals a single dark line, characteristic of mono‑layer graphene. Selected‑area electron diffraction (Fig. 4c) displays a hexagonal pattern with intense {1100} spots, confirming high crystallinity. Additional images (Fig. 4d–f) illustrate bilayer, trilayer, and five‑layer flakes. Statistical analysis of over 100 flakes (Fig. 4h) shows that ~71 % contain fewer than five layers, with an average layer count of five. Although the smallest flakes may be under‑represented due to TEM grid limitations, the data underscore the high quality of the product.

Typical TEM images and electron diffraction of the prepared graphene flakes. a Mono‑layer graphene flake with folded edge, b magnified image of the blue box in image (a), c electron diffraction of the selected black box in image (a), d a bilayer graphene flake, e a trilayer graphene flake, f a five‑layer graphene flake, g several individual graphene flakes, h distribution of number of layers (obtained from TEM analysis of at least 100 graphene flakes)

Atomic force microscopy quantifies thickness and lateral dimensions. Figure 5a shows mono‑layer flakes with a height of ~1 nm, consistent with literature values for mono‑layer graphene measured by AFM [38]. Few‑layer flakes (Fig. 5b) exhibit a thickness of ~3.6 nm and lengths up to 5 µm. Analysis of >200 flakes (Fig. 5c–d) reveals that ~90 % are thinner than 4 nm, while ~50 % have lengths between 1–7 µm. The average thickness (2.3 nm) and length (1.9 µm) confirm the high aspect ratio of the material.

Representative AFM images of a mono‑layer graphene flakes and the corresponding height profiles, b few‑layer graphene flakes and the corresponding height profiles, c thickness distribution of flakes, and d length distribution of flakes (c and d were obtained from AFM analysis of at least 200 graphene flakes)

Raman spectroscopy assesses defect density. Figure 6 shows spectra of bulk graphite and the exfoliated graphene. The D band (~1350 cm⁻¹), G band (~1580 cm⁻¹), and 2D band (~2700 cm⁻¹) are present for both samples. The 2D band in the graphene spectrum is a single symmetric peak, and the G band FWHM is 13 cm⁻¹, matching values reported for thin graphene [39]. Crucially, the D/G intensity ratio for the exfoliated graphene is 0.10, markedly lower than ratios reported for ultrasonication (0.29), shear‑force (0.17–0.37), and other fluid‑dynamic methods (0.21–0.78) [24,26,31,32], confirming the minimal defect content.

Raman spectroscopy of the bulk graphite and graphene

X‑ray photoelectron spectroscopy confirms the absence of oxidation. Figure 7 compares the C 1s spectra of bulk graphite and the exfoliated graphene, showing identical binding‑energy peaks and compositions, indicating that the hydrodynamic process does not introduce chemical modifications or oxidation.

XPS spectra of the bulk graphite and graphene

Effects of Operating Parameters on Graphene Concentration

We investigated the influence of working pressure (P), number of cycles (N), and initial graphite concentration (C_i) on the final graphene concentration (C). Figure 8a shows that increasing P from 1 to 20 MPa raises the concentration from near zero to 0.40 mg mL⁻¹. Beyond 20 MPa, the concentration plateaus, likely due to temperature‑induced restacking counteracting the enhanced cavitation stress.

Effects of operating conditions on the concentration of few‑layer graphene. a Working pressure, b number of cycles, and c initial concentration of graphite

Figure 8b demonstrates that increasing the number of cycles enhances concentration, reaching a maximum at 16 cycles (0.40 mg mL⁻¹). Additional cycles provide diminishing returns because further fragmentation of graphite brings particle sizes close to cavitation bubble dimensions, limiting new exfoliation.

Figure 8c illustrates that raising the initial graphite concentration from 2 to 10 mg mL⁻¹ boosts the graphene concentration, due to increased particle‑particle collisions that promote self‑exfoliation. When C_i exceeds 10 mg mL⁻¹, the concentration slightly declines, consistent with observations by Liang et al. [29] and Arao et al. [32], where overly concentrated dispersions hinder effective exfoliation.

Comparison with Other Fluid‑Dynamic Methods

Table 1 summarizes key metrics for few‑layer graphene produced by various fluid‑dynamic techniques. The needle‑valve method achieves a concentration of 0.40 mg mL⁻¹—higher than most reports—and a production rate of 0.40 g h⁻¹. Compared to Varrla et al. [30] (1 mg mL⁻¹) and Arao et al. [32] (7 mg mL⁻¹), our flakes exhibit longer average lengths (1.9 µm versus 0.63 µm and 1.41 µm) and a lower Raman D/G ratio (0.10 versus 0.21–0.78). These results confirm that hydrodynamic‑assisted exfoliation by a needle valve is a highly efficient, scalable route to high‑quality few‑layer graphene.

Possible Exfoliation Mechanisms

The superiority of the needle‑valve approach stems from two key mechanisms. First, as the suspension passes through the narrow gap, the total pressure drops below vapor pressure, initiating turbulent jets that generate extensive cavitation bubbles. Bubble collapse produces microjets and shock waves that shear graphite layers apart [41]. Second, the velocity gradient across the valve gap creates viscous shear forces that further promote delamination. Additionally, particle collisions during repeated passes enhance self‑exfoliation [26], all contributing to the high yield and quality observed.

Conclusions

We have demonstrated a facile, scalable hydrodynamic exfoliation method that yields high‑quality few‑layer graphene. Approximately 71 % of the flakes contain fewer than five layers, with average thickness and length of 2.3 nm and 1.9 µm, respectively. The Raman D/G ratio of 0.1 attests to the negligible defect and oxidation content. A laboratory‑scale run produced 0.40 mg mL⁻¹ at 20 MPa over 16 cycles, corresponding to a production rate of 0.40 g h⁻¹. These findings establish the needle‑valve technique as a promising platform for industrial graphene production.

Abbreviations

AFM:

Atomic force microscopy

C__i:

Initial concentration of bulk graphite

FWHM:

Full width at half maximum

I__D/I__G:

Raman D/G intensity ratio

N:

Number of cycles

NMP:

N-methyl pyrrolidone

P:

Working pressure

SEM:

Scanning electron microscopy

TEM:

Transmission electron microscopy

XPS:

X‑ray photoelectron spectroscopy