Introduction
The mid-infrared (mid-IR) spectral region (2-20 μm) is attractive for a broad range of basic science, industry, energy, and medicine. In this region, many important molecules undergo strong and distinctive fundamental vibrational and rotational transitions, such that mid-IR absorption lines are about three orders of magnitude larger than overtones situated in the visible light or near-IR wavelengths [
1]. The mid-IR spectroscopy provides a powerful tool for nonintrusive measurement and biochemical structure analysis. For example, the large absorption cross-section renders detection of trace gases at the parts-per-billion (ppb) and parts-per-trillion (ppt) levels, which is of great interest in a wide range of studies such as breath analysis for disease diagnosis, greenhouse gas monitoring, and toxic vapor detection. In biologic applications, the “fingerprint” absorption characteristics enable the possibility of tracing out the biologic sample’s chemical and physical information, which can be used for disease diagnosis, drug development, and biologic imaging. Ultrafast lasers and highly sensitive photodetectors are critical components in mid-IR photonics, but have been challenging to develop. Despite rapid technological advancements with mid-IR gain materials [
2], the lack of a high-performance saturable absorber, or a passive ultrafast optical switch, significantly limits the development of convenient mid-IR ultrafast lasers. One limitation comes from the difficulty of band gap engineering and growth techniques of conventional materials. For example, the traditional semiconductor saturable absorption mirror (SESAM), a nonlinear device used in passive mode-lock technology, is expensive and only suitable for a specified wavelength (below 2 μm) with a bandwidth of tens of nanometers [
3]. State-of-the-art mid-IR detectors normally use low band gap semiconductors such as HgCdTe alloys or quantum structures on III-V materials [
4], which are difficult to operate at room temperature. Thus, the mid-IR photonic system often suffers from high cost, bulky setup, and complex operation. A breakthrough in mid-IR science and technology will require revolutionary materials.
Graphene is a two-dimensional (2D) material that has one single layer of carbon atoms arranged in a honey-comb lattice [
5] and is a promising material to overcome the technical challenges for investigating the mid-infrared photonic functionalities [
6,
7]. Although the hexagonal lattice of carbon atoms has been found in the quasi-0D (fullerenes), quasi-1D (carbon nanotubes), and 3D (graphite) allotropes of carbon, graphene possesses unique electronic and optical properties which are different from other allotropes. Because of the linear energy dispersion relation, graphene exhibits a wavelength-independent saturable absorption, spanning from the visible light to mid-IR wavelengths. This cannot be satisfied by single-wall carbon nanotubes (SWNTs), where the working wavelength is determined by the diameter and the working bandwidth is only ~100 nm [
3]. For graphene, the carrier recombination is typically an order of magnitude faster than that of SWNTs [
8,
9]. Although there were initial doubts that too fast recombination rate may lead to insufficient saturable absorption, early experimental works nevertheless demonstrated good nonlinear absorption characteristics of graphene at the telecommunication wavelengths [
10–
12]. Graphene is a zero energy band gap material in which electrons behave like massless Dirac fermions with ultrahigh carrier mobility (up to 15000 cm
2/V/s). Graphene photodetectors have a potential ultra-high working bandwidth (up to 500 GHz), with the spectral region covering from the visible light to terahertz frequencies [
13]. Moreover, graphene, as a type of group IV material, shows great potential for the complementary metal-oxide semiconductor (CMOS) compatible process. This process offers the advantages of low-cost, high-volume, and reliable manufacturing.
Graphene has stimulated tremendous research interests in plasmonics–coherent electron oscillations that normally exist at the interface between metals and dielectrics. Graphene poorly demonstrates plasmonic effects at the visible light and near-IR wavelengths, but mid-IR plasmonics is feasible [
14]. Low carrier concentration opens opportunities for tunable carrier density by means of electrostatic doping or chemical doping, which is impossible to achieve in traditional noble materials, such as Au and Ag. Alternatively, nanostructures on graphene, such as nanoribbons or nanodisks, can help demonstrate localized mid-IR plasmonic devices [
15]. Graphene thus provides a great platform for low-loss and tunable mid-IR plasmonic applications. Furthermore, based on its unique chemical and physical properties [
16–
19], graphene and graphene-based nanomaterials have attracted strong interests from biologists. Specifically, the discovery of the adsorption of single-stranded DNA (ssDNA) onto graphene sheets, the ability of graphene to quench electron donors, the ability of graphene to protect biomolecules from enzymatic cleavage, as well as transportation capability in living cells and
in vivo systems have revealed the potential for graphene applications in biology and biochemistry. Thereby, a series of new methods have been developed for biofunctionalization with biomolecules using graphene.
In this review article, we discuss the latest status about the development of mid-IR graphene photonics and biochemical applications. Specifically, in Section 2, we discuss research progress on mid-IR ultrafast lasers by using graphene as a saturable absorber. In Section 3, we explain the working principles of different types of photodetectors and review previous efforts on the development of mid-IR graphene photodetectors. In Section 4, we discuss various biochemical applications including mid-IR plasmonic sensing, photodynamic therapy, and florescence imaging. In Section 5, we discuss the further perspective of mid-IR photonics with graphene and other 2D materials. In Section 6, we summarize this article.
Ultrafast lasers
In this section, we introduce the use of monolayer graphene as a saturable absorber in a passive mode-locked mid-IR laser. Specifically, we discuss recent experimental demonstrations of ultrashort pulse generation at mid-IR wavelengths in fiber lasers and solid-state lasers.
The exploration of graphene-based mid-IR pulsed lasers started in 2012, with several groups reporting developments around 2 μm. Typical saturable absorbers were in the form of either graphene-polymer composite films or chemical-vapor deposition (CVD) grown graphene followed by transferring to metal or dielectric mirrors. For fiber lasers, liquid exfoliated graphene-based composite films were first used to demonstrate a
Q-switched laser [
20] and a mode-locked thulium fiber laser [
21], respectively. The use of composite films ensures higher absorption and larger modulation depth, which is desirable for pulsed operation in fiber lasers. On the other hand, single-layer CVD graphene-based saturable absorber mirror has the advantages of high transmittance (~ 97%) and controllable and scalable fabrication. The relatively smaller modulation depth (typically below 1%) of single-layer graphene makes it a good alternative to the widely used SESAMs. The first CVD graphene enabled mode-locked solid-state crystal laser was based on a Tm-doped calcium lithium niobium gallium garnet (Tm:CLNGG) gain media, where the generation of 60 mW, 729 fs pulses at a wavelength of 2018 nm was achieved [
22]. Graphene was transferred onto a dielectric mirror and the device was operated in the reflectance mode as a cavity end mirror [
22]. By transferring graphene to a transparent dielectric substrate, a Tm:Lu
2O
3 crystal laser was mode-locked in such a way that shorter (~410 fs) and more powerful (~270 mW) pulses were obtained [
23]. Thus far, the longest wavelength graphene enabled mode-locked solid-state laser is based on a Cr: ZnSe laser, operating at 2500 nm, in which graphene transferred onto a CaF
2 substrate acted as an absorber in the transmission mode [
24]. The laser features a further reduced pulse duration of 226 fs, with 80 mW output power. Graphene-based saturable absorber devices can take other forms such as those directly deposited on SiC substrates.
Q-switched operation of a Tm:YAG laser was demonstrated using such a device [
25], but the use of such devices is not widely applicable to fiber lasers.
It should be mentioned that in the 2 μm wavelength region, more conventional methods such as those based on carbon nanotubes and SESAMs also work quite effectively. For example, SWNTs were used to mode-lock a solid-state lasers at a wavelength of 2.35 μm [
26]. The real comparative advantage of graphene lies in its scalability further into the mid-IR range. Wei and coworkers demonstrated the first graphene-based pulsed laser operation at a wavelength close to 3 mm, where graphene was deposited onto a fiber ferrule mirror using the optically driven deposition method [
27]. The laser emitted 2.9 μs pulses with a pulse energy up to 1.67 μJ. By optimizing the fiber laser cavity and using a multi-layer graphene covered gold mirror, as shown in Fig. 1, the same group of researchers reported mode-locked operation at a wavelength of 2.8 μm with a pulse duration ~42 ps [
28]. This is the longest laser wavelength at which graphene-based passive mode-locking has been demonstrated to date. The performance of mid-IR graphene mode-locked lasers is shown in Table 1.
Photodetectors
In this section, we discuss different working mechanisms of graphene photodetectors, including the photovoltaic effect, photoconductive effect, photothermoelectric effect, and bolometric effect, as shown in Fig. 2. Specifically, we introduce experimental demonstrations of various types of mid-IR photodetectors, based on graphene and hybrid architectures, which are realized by combining graphene with other structures or materials.
For the photovoltaic effect, electrons present in the valence band, being excited by incident photons, jump to the conduction band and become free, as shown in Fig. 2(a). Photoexcited electron-hole pairs are then separated by the internal electrical field at the graphene p-n junction. The built-in electrical field in graphene is formed by the electrostatic doping, or by using contact materials with different work functions, which can dope graphene beneath n-type or p-type. Mueller et al. utilized interdigitated metal fingers of different metals to enlarge the built-in electrical field as well as increase the light-detection region in graphene. They demonstrated a 10-GHz graphene photodetector at the wavelength of 1.55 μm. The potential spectral band of the photodetector can be expanded to the wavelength of 6 μm [
29]. In our previous work, we integrated graphene on a silicon suspended membrane waveguide [
30] and silicon-on-sapphire (SOS) waveguide [
31] to increase the graphene absorption by using the evanescent field coupling, as shown in Figs. 3 and 4. The built-in electrical field at the graphene/silicon heterostructure enables a high-responsivity photodetection under the 2.75 μm wavelength light excitation.
For the photoconductive effect, as shown in Fig. 2(b), photo-excited carriers alter the conductance of graphene by ∆
σ∝
m·∆
n, where
m is the carrier mobility and D
n is the photo-excited change of carrier density, which scales linearly with the carrier lifetime. Compared to the photovoltaic detector, graphene photoconductors normally have a higher responsivity, but lower dynamic response, since the responsivity is improved by increasing the photo-excited carrier lifetime. Liu et al. proposed and demonstrated a mid-IR photodetector with a double-layer graphene sandwiching a thin tunnel barrier structure [
32]. Under light illumination, the top layer graphene acts as a gate and introduces an electrical field to the bottom graphene layer. This structure prolongs the recombination time of photo-excited carriers and achieves a high responsivity in the mid-IR spectral region. Zhang et al. introduced electron trapping centers and a band gap to the graphene by fabricating the quantum dot-like structure, which helps increase the carrier lifetime and enhance the photoresponse [
33]. In another piece of work, the metallic antenna was used to enhance the graphene absorption and improve the responsivity of photodetectors at the mid-IR resonant frequency [
34]. The antenna rods also acted as nanoelectrodes, providing enhancement of photocarrier collection.
For the photothermoelectric effect, the incident light generates a temperature gradient at the graphene interface junction where doping levels of two sides are different, and finally results in a thermoelectric voltage (
S1-
S2)·Δ
T, where
S1 and
S2 are Seebeck coefficients in different doping regions and D
T is the electronic temperature difference between the different doping sides of graphene, as shown in Fig. 2(c). The device is operated under the zero-bias condition, which is promising for the low dark-current. The sign of the photothermoelectric current is opposite to the photovoltaic current in the graphene bipolar p-n junction and the unipolar p
+-p or n-n
- junction. Hsu et al. demonstrated a graphene-based thermal imaging system by integrating graphene photothermoelectric detectors with silicon nitride membranes at the wavelength of 10.6 μm [
35]. Badioli et al. observed an enhanced photothermoelectric photoresponse under the mid-IR light excitation [
36] since mid-IR light can provide an efficient way to excite bulk or surface phonons, which can in turn locally heat up graphene charge carriers.
For the bolometric effect, the light illumination produces heat and results in a change of the resistance of the graphene device, as shown in Fig. 2(d). Since the resistance change should be converted to a detectable electrical signal, it often requires an external bias and is operated in homogenous highly doped graphene without the need of the p-n junction structure. This type of thermal detector provides a high sensitivity and does not require cryogenic cooling to mitigate noise. Unlike the photoconductive effect, in which photo-excited carriers increase the carrier density and enhance the device conductance, the photo-induced temperature increase will reduce the carrier mobility and reduce the conductance. In our former studies, we have demonstrated a highly p-type doped graphene photodetector integrated on the silicon nitride waveguide to enhance the absorption at 1.55 μm wavelengths [
37]. The spectral bandwidth of the proposed device can cover the transparent window of silicon nitride material (up to 5.5 μm). Yan et al. demonstrated the graphene bolometer with a due-gate configuration which creates a tunable bandgap and electron-temperature-dependent conductivity in the bilayer graphene [
38]. The bolometer exhibits a low noise-equivalent power, and a high intrinsic speed. Freitag et al. also report the mid-IR bolometric photocurrent measurement in graphene nanoribbon arrays on SiO
2 substrate [
39], where a strong coupling exits between light and various dipole-carrying excitations. The study directly proved the importance of the substrate’s phonons in the photocurrent generation process. The performance of mid-IR graphene photodetectors is shown in Table 2.
Although graphene has an ultra-wide spectral bandwidth, the atomic layer thickness results in only ~2.3% absorption of the incident light which limits the responsivity of graphene photodetectors. One way to enhance the light absorption is to integrate graphene onto waveguide devices, in which the elongated optical interaction length can dramatically increase the absorption toward ~100% [
40]. In our previous study, we proposed and demonstrated to use different structures, such as silicon slot waveguides [
41] and silicon nitride microring resonators [
42], to further enhance graphene to light interaction, as shown in Figs. 5 and 6. As waveguides are transparent in the mid-IR spectral region [
43], ultracompact, waveguide integrated, mid-IR graphene photodetectors are a promising technology for the future.
Biochemical applications
With excellent optical and electronic properties, remarkable biocompatibility and ease of functionalization, graphene and its derivatives have received increasing attention in biochemical applications. In this section, we discuss recent experimental progress on mid-IR plasmonic sensing, photodynamic therapy, and fluorescence imaging applications.
Graphene plasmonics was first observed in the nanoribbon structure at the terahertz frequency in 2011 [
44]. Since then, research efforts have been made to shift the frequency of localized surface plasmon resonance (LSPR) to the mid-IR spectral region by using electrostatic or chemical doping and fabricating nanostructures, such as nanoribbons or nanodisks, on the graphene layer [
45–
47]. With the atomic layer thickness and the limited carrier density, graphene LSPR devices possess two excellent characteristics, namely ultrahigh light intensity confinement and frequency tunability, which make graphene serve as an excellent platform for ultrasensitive and tunable mid-IR sensors.
Based on the graphene nanoribbon structure, in 2014, Li et al. first used the near-field enhancement of light-matter interactions by graphene plasmons to probe the chemical bond information in the analyte [
48]. They studied the interaction between graphene plasmons and vibrational modes in surface-adsorbed thin polymer films of poly(methylmethacrylate) and polyvinylpyrrolidone. The detection sensitivity was increased by a factor of ~5 compared to that of light attenuated by the carbonyl vibration alone. In 2015, Rodrigo et al. reported a high-sensitivity mid-IR plasmonic biosensor [
49], which was used for chemically specific label-free detection of protein monolayers. Due to the dynamic tunable capability of the plasmonic resonant frequency, complex refractive index variations can be extracted from the plasmonic resonance shift and spectral dips with the amide I and II bands in the extinction spectra. This reveals the presence of the protein with chemical specificity. With great light intensity enhancement in the atomic layer thickness, the spectral shifts and absorption signals of graphene plasmonic sensors were demonstrated 6 times and 3 times larger than those obtained using metallic plasmonic sensors, respectively.
Since graphene can respond to photons in a broad spectrum [
50], many graphene-based methods have been developed, such as lasers for imaging, drug delivery, and photodynamic therapy. Since the report from Dai Group that nano-graphene oxide (NGO) works for cellular imaging and drug delivery [
51], a number of developments have been reported on this topic. Such NGO materials can easily be conjunct with other organic molecules to form complex and uptaken by cells. Therefore, they are natural carriers of molecules that can provide therapeutic effects such as oxidative stress to cells when they are activated by lasers. In this regard, NGO can be combined with other drugs as the carrier for drug delivery [
52,
53] and finally photodynamic therapy (PDT) [
54,
55]. Considering the direct absorption of NGO to lasers and the thermal effect, NGO itself can directly work as the PDT drug. The targeting process of NGO to diseased tissue or tumor can be achieved by immunomolecules or simply by metabolism [
56,
57]. After an uptake of NGO by the targeted cells or tissue, a higher thermal effect can be generated by IR laser irradiation which can then kill those cells or tissue [
58–
62]. The low toxicity of NGO is thus promising for clinic applications. NGO has a broad spectral bandwidth, various tissues with different optical properties and laser irradiation wavelengths are suitable for PDT [
63,
64].
Interestingly, during the research of PDT with graphene or NGO particles, it has been noticed that the graphene and NGO particles can emit visible-range fluorescence if the density of IR photons is high enough [
65]. The multi-photon excitation was efficient and the fluorescence was stable due to the unique optical property of graphene. Therefore, graphene and NGOs are also widely used for in
vivo multi-photon microscopy with ultrafast lasers [
66], especially deep-tissue imaging [
67]. Raman microscopy can also be achieved [
68,
69]. Importantly, these methods above can be achieved simultaneously by only graphene or NGOs [
70].
Future prospective
The unique optical nonlinearities of graphene have shown great promises for enabling unprecedented ultrafast photonic capabilities in the mid-IR spectral region. In addition to enabling ultrafast bulk lasers, refined growth and fabrication methods will open up new routes for realizing the new-generation of integrated lasers. Indeed, the success in using graphene to mode-lock a near-IR vertical-external-cavity surface-emitting laser (VECSEL) has demonstrated the potential for an all-integrated graphene mode-locked laser [
71]. On the other hand, highly tunable Dirac fermions offer another important direction for further exploitation. Voltage-controlled devices have shown flexible parameter control capabilities in graphene-based optical devices [
72]. Combining such a feature with other advanced photonic methods such as plasmonic nanostructures and micro-cavities can provide great flexibility in the design and optimization of photonic functional devices [
73]. Extensive efforts are being made in this line of research and the impact is expected to be far reaching. Finally, it should be pointed out that graphene is not the only strategic material for the mid-IR range. Emerging materials such as 2D black phosphorus, which exhibits a small bandgap of ~0.3 eV, have also stimulated a great deal of research efforts in the area of mid-IR photonics. In addition to saturable absorbers, photodetectors and other photonic devices have been theoretically exploited and experimentally demonstrated [
72–
75].
Summary
In this article, we have reviewed recent experimental progress on mid-IR graphene photonics and biochemical applications. With the distinctive band structure and broadband spectral region, graphene has tremendous potential applications in the mid-IR photonics and biotechnology which have been hardly accessible using traditional materials. In the future, graphene photonics and its biochemical applications are anticipated to bring us exciting scientific and technological advances.
Higher Education Press and Springer-Verlag Berlin Heidelberg