An integrated and multi-purpose microscope for the characterization of atomically thin optoelectronic devices

Optoelectronic devices based on graphene and other two-dimensional (2D) materials, such as transition metal dichalcogenides (TMDs) are the focus of wide research interest. The characterization these emerging atomically thin materials and devices strongly relies on a set of measurements involving both optical and electronic instrumentation ranging from scanning photocurrent mapping to Raman and photoluminescence (PL) spectroscopy. Furthermore, proof-of-concept devices are usually fabricated from micro-meter size flakes, requiring microscopy techniques to characterize them. Current state-of-the-art commercial instruments offer the ability to characterize individual properties of these materials with no option for the in situ characterization of a wide enough range of complementary optical and electrical properties. Presently, the requirement to switch atomically-thin materials from one system to another often radically affects the properties of these uniquely sensitive materials through atmospheric contamination. Here, we present an integrated, multi-purpose instrument dedicated to the optical and electrical characterization of devices based on 2D materials which is able to perform low frequency electrical measurements, scanning photocurrent mapping, Raman, absorption and PL spectroscopy in one single set-up. We characterize this apparatus by performing multiple measurements on graphene, transition metal dichalcogenides (TMDs) and Si. The performance and resolution of each individual measurement technique is found to be equivalent to that of commercially available instruments. Contrary to nowadays commercial systems, a significant advantage of the developed instrument is that for the first time the integration of a wide range of complementary opto-electronic and spectroscopy characterization techniques is demonstrated in a single compact unit.

the in situ characterization of a wide enough range of complementary optical and electrical properties. Presently, the requirement to switch atomically-thin materials from one system to another often radically affects the properties of these uniquely sensitive materials through atmospheric contamination. Here, we present an integrated, multi-purpose instrument dedicated to the optical and electrical characterization of devices based on 2D materials which is able to perform low frequency electrical measurements, scanning photocurrent mapping, Raman, absorption and PL spectroscopy in one single set-up with full control over the polarization and wavelength of light. We characterize this apparatus by performing multiple measurements on graphene, transition metal dichalcogenides (TMDs) and Si. The performance and resolution is equivalent to commercially available instruments with the significant added value of being a compact, multi-purpose unit. Our design offers a versatile solution to face the challenges imposed by the advent of atomically-thin materials in optoelectronic devices.
Since the discovery of graphene, 1,2 the ability to isolate atomically thin materials has attracted growing interest within the scientific community owing to a unique set of unprecedented properties suddenly available on a single chip. 3,4 Two-dimensional semiconductors are currently enabling conceptually new devices, including light detectors and emitters, 5 transistors 6 and memories. 7 The characterization of such devices strongly relies on a set of techniques which combine electrical, optical and spectroscopic measurements. For example, inelastic light scattering (Raman) spectroscopy has been the main tool to study the properties of graphene and its derivatives, 8 while photodetectors based on 2D materials have been largely characterized by Scanning Photocurrent Microscopy (SPCM). To characterize both light-emitting and detecting devices, the knowledge of optical parameters such as the absorption and reflection coefficients of the material is of paramount importance. Commercial instruments are usually used to perform these characterizations in distinct self-standing systems which requires transport of the sample in different environments and retrofitting to accommodate the different holders designed for each specific tool.
In this work, we present an experimental apparatus developed to characterize optoelectronic devices based on graphene and other 2D materials. Centred around an upright metallurgical microscope, this system allows us to perform multiple measurements in one instrument, with no need to remove the device from its holder, reducing the risk of contamination or breakage. These measurements include: low frequency electrical transport, SPCM, absorption (transmittance and reflectance), micro-Raman, photoluminescence (PL) and electroluminescence (EL) spectroscopy and mapping, with or without polarized light. A unique aspect of the developed design is the ability to concurrently perform both the electrical and optical measurements. The system is equipped with multiple laser sources, spanning from UV to red light and two white light sources used for transmission and reflection illumination. Our design delivers the laser light in enclosed paths and an interlock system cuts the laser light to allow access to individual parts, making it extremely safe. High spatial resolution in spectroscopy and SPCM measurements is achieved by diffraction-limited focusing of Gaussian laser beams and a high performance microscope stage.
Electrical connections are secured by a custom-built PCB board, designed to reduce electrical noise and allow easy access to devices, without the need for long working distance microscope objectives. We characterize the instrument with a series of standard measurements: SPCM on graphenebased photodetectors as well as absorption, Raman and PL spectroscopy of a range of 2D materials and organic semiconductors. We prove its high functionality and demonstrate that the achieved resolution, both spatial and spectroscopic, and the overall performance, is equivalent to current commercial technologies, with the additional benefit of having a compact, multi-purpose, fully customizable, instrument with the potential for installation of additional measurement features.

I Operating principles
Light-matter interaction is at the heart of optoelectronic devices. 9,10 These work by converting an electric signal into light or vice-versa. For this reason their design requires knowledge of both the optical and electrical characteristics of the materials used. In this section we will review the fundamental principles of light detection and emission, with focus on atomically-thin devices, and the associated measurement techniques, such as SPCM. We will also review the fundamentals of the spectroscopy techniques, absorption, Raman and PL, which allow the optical properties of the active materials to be studied. The two main categories in which optoelectronic devices are divided are light emitters (LED) and photodetectors (PDs).

I.A Light emission and detection in atomically thin devices
To characterize LEDs, it is necessary to be able to electrically drive the light emission mechanism whilst recording the spectrum and power of the emitted light. The electrical input can be DC, in the case of injection-devices such as p-n junction diodes, 10 or AC, as required by alternating current electroluminescent (ACEL) devices. 11 The spectral analysis of the emitted radiation is used to gain insight into the physics of the light-emitting material. This is usually achieved by efficiently collecting the emitted light and delivering it to a high-resolution spectrometer with minimal losses or aberrations in the optical path. High spatial resolution can be achieved by employing a microscope lens to collect the light. In this way the area of the LED can be accurately mapped, enabling the study of gradients, defects, inhomogeneities or the presence of localized emitters.
To characterize a PD, the basic experiment consists of shining light onto the device and recording its electrical response, such as measuring the current flow through it or a change in resistance. 10 Light can impinge on the whole surface of the device, known as flood illumination, or it can be delivered with a focused laser onto a specific area to allow for a spatially resolved photo-response.
Both techniques give insight on the physical nature of the observed photoresponse and on the efficiency of the device. In general, the magnitude of the generated photocurrent is given by: 10 where η i = (I ph /q)/φ abs is the internal quantum efficiency of the PD, defined as the number of charges, q, collected to produce the photocurrent I ph , divided by the number of absorbed photons φ abs , P opt is the incident optical power, ν is the energy of the photons, α is the absorption coefficient of the material (see section I.B) and δ its thickness. The exponent β depends on the photocurrent generation mechanism, in general it is equal to 1 for photovoltaic (PV) devices. A summary of the main quantities used to characterize LEDs and PDs is given in Table 1.
In graphene and TMDs several mechanisms are responsible for the observed photocurrent. 12 The most widely used technique to study such devices is SPCM, where a laser beam is scanned across the device and the electrical response recorded at each point, producing a two-dimensional map of the photoresponse. 13 SPCM allows the study of spatially resolved properties of optoelectronic devices, where the spatial resolution is ultimately defined by the laser spot size, d s .
Improvements in spatial resolution have been achieved by the use of near-field techniques, such as scattering scanning near-field optical nanoscopy (s-SNOM). 12 SPCM has been used, for example, to study the dynamics of hot carriers in graphene PDs, 14 to demonstrate the role of defects in observed photocurrent signals 14 and to characterize the photoactive regions in TMDs junctions. 15

I.B Absorption spectroscopy
Design and characterization of optoelectronic devices requires the optical properties of the photoactive materials, coatings and electrodes to be known. As shown in Equation (1) the absorption η P out /(I in · V ) Responsivity I ph /P opt Specific Detectivity D (S · ∆f ) 0.5 /NEP a I ph = Measured photocurrent, φ in = Incident photon flux, φ abs = Absorbed photon flux, P opt = Incident optical power, P out = Generated optical power, I in = Input electrical power, V = Applied voltage, S = Device area, ∆f = Operating bandwidth, NEP = Noise Equivalent Power; coefficient, α, of the material has an important role in the response of the device. Furthermore, transmittance and reflectance are key parameters for transparent electrodes in PDs and LEDs.
Light impinging on a slab of material of thickness δ is partially reflected at the first surface and partially transmitted in the material. This is then partially reflected at the second surface and partially transmitted outside. In general the transmittance of a free-standing slab of material, T , defined as the ratio of the transmitted power (I T ) to the incident one (I 0 ), is expressed, in terms of the reflectivity of the two surfaces (R 1 and R 2 ), as: 9 The absorption coefficient α(λ) is defined as the fraction of power absorbed per unit length into the material, such that the intensity at a distance x in the material is given by: I(x) = I 0 exp(−αx).
Solving eq. (2) for α allows the absorption coefficient to be extracted from the transmittance and reflectance of a material.
For a thin film deposited on a thick substrate, as with TMDs and other 2D crystals, an equivalent formula has been derived by Swanepoel 16 : where: Here n and s are the refractive indices of the medium and substrate respectively, R 1 is the reflectance of the air/medium interface, R 2 is the reflectance of the substrate/medium interface and R 3 the reflectance of the substrate/air interface. Each R i can be either measured directly or computed from the knowledge of n and s.
Within the non-interacting single particle theory, absorption is related to the optical transitions of electrons from occupied to empty states. The absorption coefficient is zero for photon energies below the bandgap (in absence of mid-gap trap states), it increases sharply for energies close to the bandgap and, in general, stays relatively constant, with the appearance of peaks related to resonant transitions. The band edge has a specific shape: 9,17 for indirect transitions (neglecting where E ph is the incident photon energy, E g is the bandgap energy, Ω is the associated phonon energy and m is equal to 2 for allowed and 3 for forbidden transitions; for direct transitions α ∝ (E ph − E g ) 0.5 . Thus, absorption spectroscopy is a viable method to determine the bandgap of a semiconductor and its nature, direct or indirect.
In the specific case of 2D materials it has been shown, for example, that the bandgap of few-layer graphene can be tuned with an applied electric field 18 and in TMDs the bandgap energy changes with the number of layers and induced strain. 19 TMDs also show a transition of the bandgap from indirect to direct when they are thinned to single-layers. 3 This large variety of physical phenomena makes absorption spectroscopy an important tool for studying novel, atomically-thin materials.

I.C Inelastic light scattering and luminescence spectroscopy
The inelastic scattering of light, or Raman effect, involves the scattering of a photon by an excitation in the examined material. 20 This is usually given by the transition of an electron from the highest occupied molecular orbital (HOMO), or valence band in crystalline solids, into an empty state, which could be a virtual state within the bandgap or a real state above the low- where the integration is carried over the collection angle of the microscope objective dΩ S , c 1 is a material dependent factor and T represents the transpose of the vector. Thus, by changing the excitation and collection axes and the polarization of the light it is possible to determine the crystallographic orientation of the material and identify the different vibrational modes. 21 The intensity of the Raman scattered light is dependent on the local temperature T and the ratio between Stokes (I S ) and anti-Stokes (I aS ) intensities can be used as a temperature probe. 22 This has a well-known expression: 23 where F is a parameter that depends on the optical constants of the material and C ex is a coefficient which accounts for the efficiencies of the optics and gratings at different wavelengths. Since the Raman signal is ∼ 10 6 times smaller than the Rayleigh process, a very narrow notch filter, centred at the excitation wavelength, is typically used to reject the unwanted signal.
Radiative transitions of electrons from an excited state to a lower level lead to the emission of light, a process known as luminescence. 9 These transitions can be initiated by the absorption of a photon, known as Photoluminescence (PL), fluorescence and phosphorescence, or by the application of an electric field through the material, known as Electroluminescence (EL). In both cases, detailed measurement of the spectral intensity of the emitted light can give insight into the properties of the material and the performance of the examined device. In a typical PL experiment the excitation source is a laser, with a wavelength appropriate for exciting resonant electronic transitions. The emitted light from the material is then dispersed for spectral analysis.
Usually a low-pass filter is used to reject the excitation wavelength. The detailed physical processes involved in PL are linked to the excitation and decay probabilities in the material. 9 The frequencydependent intensity of a luminescence process can be written, in general, as: where M 2 is the matrix element of the transition, g(hν) is the density of states (DOS) and Γ e represents the occupancy factors for the transition, which are related to the probability of the upper (lower) levels to be empty (occupied). White light for reflection microscopy is provided by a white LED along the optical path of the laser beams, while transmittance illumination is achieved using the microscope's built-in lamp and condenser lens. Both the white and laser light are directed to the objective using either a 50%/50% BS or the appropriate dichroic band-pass (BP) mirror. The reflected light collected by the objective is directed to a CCD imaging camera or, by using a flip mirror (FM), to the entrance slit of the spectrometer. The microscope is fitted with Olympus MPLanFL-N Semi-Apochromat infinity-corrected lenses with ×5, ×10, ×20 and ×50 magnification.

II Instrumentation II.A Optics
The spectroscopy section is composed by a focusing achromatic lens (T horLabs AC254-150-A) with x/y and focus adjustment. The spectrometer is a Princeton Instruments Acton SP2500, equipped with three dispersion gratings (1200 g/mm with 500 nm and 750 nm blaze, and 1800 g/mm with 500 nm blaze) and a Princeton Instruments PIXIS400-eXcelon back-illuminated, peltier cooled, CCD camera. In our multi-purpose instrument, each grating is used for different measurements as they differ by number of grooves and blaze wavelength (at which the efficiency is at maximum and the same for both S and P polarizations). In white light spectroscopy applications, corrections to account for the optics and gratings efficiencies, at different wavelengths, can be applied using Princeton Instruments IntelliCal system.
The microscope stage is a Prior Scientific OptiScan ES111 with ProScan III controller. The minimum step size is 10 nm. Focus control is also achieved trough the same controller.
The whole system is built on a vibration-isolated 120 × 90 cm 2 optical table. The laser light delivery system is enclosed within ThorLabs stackable tube lens system, allowing a light-tight connection. The microscope stage, foot and objective turret is covered with a light-tight enclosure, formed by a metal frame and a conductive fabric curtain connected to the ground of the electrical circuit to ensure shielding from electro-magnetic noise. The front of the curtain can be lifted to allow access to the stage and it is fitted with laser interlocks. The light-tight delivery system and enclosure, together with the magnetic interlocks, make the system a Class 1 laser product. The same light-tight tubes are used also to deliver light to the spectrometer, enabling the use of our instrument without the need of a darkened room.
The custom-built epi-illumination system allows flexible and quick configuration of the microscope for different measurements. As shown in fig. 3, the optical path can be configured for scattering, transmission/reflection spectroscopy and laser light illumination (as in SPCM) simply by replacing or removing the appropriate filters and BS. Each BS unit is fitted with magnetic interlocks for safety purposes. Figure 3a shows the configuration for reflectance measurements: light from the white LED travels freely (no laser BS fitted) after being collimated and reaches the objective though a 50%/50% BS; reflected light is collected by the same objective and partially transmitted by the BS to a FM which directs it into the spectrometer. A similar arrangement is used for transmittance spectroscopy, as shown in fig. 3b. In this case the BS is replaced with a mirror and the white light from the incandescent bulb is collimated by the microscope condenser and, after travelling in the sample, collected by the objective and directed to the spectrometer. Design of optical systems involving lasers rely strongly on the knowledge of the shape and size of the beams. In our setup, solid-state diode lasers are used and all the optical components are chosen so that to minimize deviations from the TEM 00 laser mode, which has a Gaussian intensity distribution: where r is the distance from the optical axis and r b is the radius of the beam, measured at the point at which the beam intensity falls by 50% (FWHM). As the beam is transmitted through a circular aperture of radius r a , such as the pupil aperture of the objective lens, the transferred power is: where T = r b /r a is defined as the Gaussian beam truncation ratio. The focused beam profile is Gaussian for T < 0.5 and converges to the Airy pattern for T → ∞. The focused spot size diameter (d s ) can be expressed as: 24,25 where NA is the numerical aperture of the lens, λ is the wavelength of the laser and K is the kfactor for truncated Gaussian beams, which is a function of T only. Urey 24 computed approximate expressions for K for the two cases T < 0.5 (Gaussian) and T > 0.4 (Truncated Gaussian). In the same work an expression to determine the depth of focus, ∆z (i.e. the distance along the optical axis at which the irradiance drops by 50%), was determined: where f # is the ratio between the clear aperture diameter and the focal length of the lens (also called f -number) and K 2 is the focus k-factor, again a function of T only. In our apparatus we found that T = 1.03 for the UV (375 nm) laser and T = 0.52 for the visible lasers when using the ×50 objective. In table 2 we show the results obtained for our system using equations 10, 11 and the results in Ref. Urey 24 . The calculated values show that our system is able to focus the laser light well in the diffraction-limit of the objective lens, allowing ultra-high spatial resolution. The very narrow depth of focus allows substrate contributions to be minimized.

II.B Electronics
Optoelectronic devices based on graphene and 2D materials are usually fabricated from thin flakes deposited on a Si substrate capped with a SiO 2 layer, where electrical contacts are defined via lithography and metal evaporation. We designed a chip carrier for our microscope which allows versatile and easy connection of such devices. As shown in fig. 4, the carrier is composed of a PCB board (34 × 29 mm 2 ) with two standard 11-way pin strips, where 20 of these pins are connected to gold-coated pads and one to a central 15 × 15 mm 2 gold-coated pad (one is not connected).
The central pad is used to contact, using silver paint, the doped silicon substrate which provides Full automation of the system is achieved with a custom-made LabView-based software which is able to communicate to all electronic instruments via GPIB/USB bus and modulate the lasers via digital-to-analog interface (DAC) (National Instruments NI-DAQ); Native software is used to control the spectrometer and CCD camera, interfaced with the same LabView software.
III Performance of multi-purpose microscope system III.A Optoelectronic devices As a further system characterization, we demonstrate the polarization capabilities of our apparatus by measuring the polarization dependence of the observed photocurrent. We focus the laser at the graphene/metal interface and rotate the λ/2 waveplate, in order to change the polarization of the incident light, while recording the photocurrent. It is well known that, in this type of device, hot-carriers dynamics in graphene, combined with the direct absorption of the metal contacts, lead to a polarization-dependent photocurrent. 27 In fig. 5c we show a plot of the photocurrent as a func-tion of polarization angle θ. We observe that I ph is maximum when the polarization is orthogonal (⊥) to the metal contact and minimum when it is parallel ( ) to it, as expected. 27 The angle θ is defined as the angle between the polarization of the laser and the vertical (y) axis of the microscope stage. Since our device sits at an angle φ 57°with respect to y (see fig. 5a), the polarization dependence should have a "phase shift" of the same amount, i.e. I ph ∝ sin(2(θ − φ)), as verified in fig. 5c (solid blue line).

III.B Absorption spectroscopy
In section I.B we showed that, in order to extrapolate the absorption coefficient of a material, we need to perform transmittance and reflectance measurements. These can be done in our system by using the optical configurations shown in fig. 3a-b. As a characterization of the performance of our setup we present a series of measurements of transmittance and absorption coefficient for different materials. Figure 6a shows the visible transmittance of single-layer and multi-layer CVD-grown graphene.
As expected, 28 single layer graphene shows a flat transmittance ∼ 97.7% across the whole visible spectrum, while nickel-grown multi-layer CVD graphene 29 shows a transmittance of ∼ 75 − 80%.
The contribution of the transparent substrate, quartz in this case, has been subtracted from the data by performing a calibration measurements on a bare substrate.
We now show two examples of absorption coefficient measurements, performed by measuring transmittance and reflectance and applying eq. (3). The first example is a thin layer of organic crystalline semiconductor, Rubrene. The optical and electrical properties of this organic semiconductor have been intensively studied. 30 Therefore, it offers a good standard test the performance of the developed multi-purpose microscope system. Figure 6b shows the measured α(λ) of a Rubrene crystal on glass, using unpolarized light, perpendicular to the a-b facets and parallel to the c facet of the crystal (see Ref. Irkhin et al. 30 for details). The measured spectrum is in very good agreement with literature, both in the intensity and position of the peaks. 30 As a second demonstration of the spectroscopy capabilities of our system, we measured the absorption coefficient of ultra-thin HfS 2 . This is a well known TMD which was well characterized in the past as a bulk crystal 31 and recently as a thin flake on a substrate. Figure 6c, inset, shows the absorption coefficient of a ∼ 25 nm thick flake of freshly-exfoliated HfS 2 . The values agree well with literature. 17,31 Owing to the proportionality between α and E (see section I.B), we are able to extrapolate the bandgap of the material by plotting α 2 as a function of E. The intercept of the extrapolated linear part mark the direct bandgap of the material E g = 2.75 ± 0.01 eV, in very good agreement with established values. 31

III.C Raman spectroscopy
To demonstrate the ability of our set-up to also perform inelastic light spectroscopy we characterized it by studying two well-known materials: Si and graphene. Figure 7a shows Characterization of graphene and other 2D material relies strongly on Raman spectroscopy. 8 The resonant Raman spectrum, acquired with our setup, of exfoliated graphene is shown in fig. 7b.
We observe the G peak at ∼ 1580 cm −1 , originating from the E 2g mode and the double-resonant 2D band at ∼ 2680 cm −1 , originating from the A 1g mode. The 2D band is sensitive to the number of layers: it is given by a single Lorentzian peak for single-layer graphene, the convolution of four Lorentzian peaks for bi-layer graphene and the convolution of six peaks for tri-layer graphene. 32 In fig. 7b we show the ability of our system to resolve such multi-peak structures (green lines) and thus discriminate the number of layers in graphene flakes. Figure 7c shows the same spectrum of single-layer graphene acquired at different incident powers. The inset shows the height of the G peak as a function of incident power, adhering to the expected linear relationship. This serves to confirm that no artefacts are introduced in the spectra by the experimental apparatus. It is worth noting that we are able to resolve clearly the G and 2D bands with an incident power density as low as 50 kW/cm 2 with an acquisition time of only 2 s for the whole spectral range.
In our setup it is possible to perform polarized Raman spectroscopy using the configuration shown in fig. 3d. This technique has recently been applied to the investigation of graphene 33 and TMDs, 34 where strong polarization dependence is given by valley anisotropy, highlighting the role of Raman spectroscopy in the study of electron-phonon coupling. To demonstrate the polarization capabilities of our setup, we characterized the dependence of the Raman modes of Si upon incident polarization. Figure 8a shows the Raman spectrum of Si with the [100]-surface perpendicular to both the incident and the scattered light (zxxz and zxyz configurations), acquired with λ exc = 514 nm at 2.1 MW/cm 2 incident power for 2 s. We observe the 1TO mode at 520 cm −1 and a twophonon mode at ∼ 900 − 1000 cm −1 . Using eq. (5) and the expression for the Raman tensor of the 1TO mode, we find that the Raman intensity of this mode is given by: 21 where ϑ is the polarization angle between the incident and scattered light. As the angle between the incident polarization and the [100] axis ϕ is varied, we observe a decrease of the 1TO peak, while the two-phonon mode changes only slightly, as shown in fig. 8a. Equation (12)

III.D Luminescence spectroscopy
In this section we characterize the ability of our system to measure luminescence phenomena: photoluminescence (PL) and electroluminescence (EL). As an example of a PL measurement we use single-layer WS 2 , a semiconducting TMD with good potential in optoelectronic applications. 35,36 The room-temperature PL spectrum is shown in fig. 9a. This is acquired using the configuration shown in fig. 3d, with λ exc = 514 nm, P = 5.1 kW/cm 2 and 10 s acquisition time; the power is kept low to avoid causing damage to the sample. A strong peak is observed, centred at 2.02 eV and the Raman modes can be also observed in the same spectrum (highlighted in the green box). As an example of EL device characterization, we use a ZnS 2 -based ACEL device with both graphene and functionalized graphene transparent electrodes. 38 Figure 9b shows the spectral emission of such device with graphene and FeCl 3 -intercalated graphene electrodes. The spectra can be decomposed into multiple Gaussian peaks, corresponding to the intraband transitions 39 in ZnS 2 .
The use of our instrument allows the accurate comparison of the two electrode materials and the performance of the different devices (see Torres Alonso et al. 38 for details).

IV Conclusion
We have presented the design of a new multi-purpose instrument for the characterization of optoelectronic devices based on 2D materials, capable of performing multiple electrical and optical measurements simultaneously. We demonstrated the performance of the setup with a series of techniques, involving a multitude of materials, showing results at the level of the state-of-the-art commercial equipment. The ability to perform low-frequency electrical transport measurements, SPCM, Raman, absorption and PL spectroscopy, combined with full automation, high sensitivity and low noise, make our instrument ideal to face the challenges imposed by the advent of atomically-thin materials in optoelectronic devices.
Customization of each section allows multiple routes for future upgrades and expansion, these will include: a vacuum chamber microscope stage with temperature control, multi-wavelength Raman spectroscopy, auto-focusing and high-frequency (RF) measurements. The cost and size is very contained, making it suitable for small laboratories and the safety features introduced make it very easy to operate, with minimal training required.  . Four main components can be identified: upright microscope with custom-built EPI illumination system and motorized XYZ stage; Custom-built laser-enclosure; Spectrometer and CCD camera; Sample-holder PCB board with BNC break-out box. The electrically screened, light-tight, enclosure of the sample is shown in transparency and sectioned, this is directly grounded. Tube lenses are used to cover the light path of the laser beams, magnetic interlocks are located in the screen enclosure and in the BS cubes, effectively making the system a Class 1 laser product. Labels and abbreviations as in Figure 1.