GB2579785A - Photovoltaic device - Google Patents

Abstract

A photovoltaic device, comprising: a photovoltaic cell 102; and a wavelength converter 106 on a light incident side of the photovoltaic cell, the wavelength converter comprising aggregates of nanoparticles that absorb light at a first wavelength and re-emit the light at a second, longer wavelength. The nanoparticles can comprise perovskite or ZnCuInS/ZnS core/shell quantum dots. The nanoparticles are embedded in a polymethyl methacrylate PMMA polymer matrix. The invention also extends to a photovoltaic module comprising: a plurality of photovoltaic cells; and a transparent cover 104 to protect the photovoltaic cells from environmental degradation, wherein the transparent cover is coated on an inner surface with wavelength converting nanoparticles. The nanoparticles are formed by spin-coating.

Description

PHOTOVOLTAIC DEVICE

FIELD

The present invention relates to photovoltaic devices, and in particular to a photovoltaic device with a wavelength converter for enhancing the efficiency of the photovoltaic device.

BACKGROUND

Solar energy is distributed over a wide spectral range including ultraviolet through visible and into infrared, with the peak spectral irradiance occurring with light of colour cyan. Conversely, the effectiveness of a photovoltaic (PV) cell varies with wavelength, so that a monocrystalline silicon PV cell is doubly effective at converting near-infrared light into electricity than cyan. This mismatch between solar irradiance and PV spectral responsivity means that a significant proportion of light energy is wasted, being converted into heat instead of electricity.

SUMMARY

An embodiment of the invention provides a photovoltaic device, comprising: a photovoltaic cell; and a wavelength converter on a light incident side of the photovoltaic cell, the wavelength converter comprising aggregates of nanoparticles that absorb light at a first wavelength and re-emit the light at a second, longer wavelength.

The use of nanoparticle aggregates is unconventional because typically it is considered desirable to prevent nanoparticle aggregation. However, effective results, i.e. improvements in power conversion efficiency, are obtained with nanoparticle aggregates. The wavelength converter may utilise down-shifting (whereby a photon of energy E2 is emitted after the absorption of a single photon of energy El> E2), down-conversion (whereby two or more photons of energy E2 are emitted following the absorption of a single photon of energy El> E2), or a combination of down-shifting and down-conversion. The relationship between wavelength, A, and energy, E, is described by the equation E=hc/A, where h is Planck's constant and c is the speed of light.

In embodiments, the nanoparticles comprise non-toxic nanoparticles. For example, the nanoparticles may comprise cadmium-free nanoparticles.

In one embodiment, the nanoparticles comprise quantum dots (QDs). However, other types of nanoparticles, such as quantum rods (QRs), may alternatively be used.

In one embodiment, the quantum dots comprise ZnCuInS/ZnS core/shell quantum dots. For aggregates of such quantum dots, light may be absorbed in a range of wavelengths with peaks at around 300 nm and around 550 nm. Light may be emitted in a range of between about 720 nm to 820 nm, as defined by the Full Width at Half Maximum, FWHM, of the corresponding emission curve, i.e., the width of the emission curve between the two cut-off points where the emission is one-half of a maximum emission.

In another embodiment, the quantum dots comprise perovskite quantum dots. For aggregates of such quantum dots, light may be absorbed in a range of wavelengths with peaks at around 280 nm and around 510 nm. Light may be emitted in a range of between about 500 nm to 530 nm, as defined by the FWHM of the corresponding emission curve.

Compared to non-aggregated quantum dots, aggregated quantum dots may exhibit a red-shift in the emission. For example, aggregates of ZnCuInS/ZnS core/shell quantum dots exhibit a red-shift of around 50 nm to 100 nm, for example around 80 nm, compared to non-aggregated ZnCuInS/ZnS core/shell quantum dots.

Invention nanoparticle aggregates are not limited to these compositions and may include modifications to, among other things, the shell and ligands.

The wavelength converter may be in the form of a layer of nanoparticles in front of the cover glass of the photovoltaic device, nanoparticles disposed in the cover glass, and/or a layer of nanoparticles behind the glass. In one embodiment, the quantum dots are embedded in a polymer matrix. For example, the polymer matrix comprises polymethyl methacrylate.

Generally speaking, a photovoltaic device can be a photovoltaic cell, a photovoltaic module having a plurality of cells, a string of photovoltaic modules (i.e., a one-dimensional array of modules), a sub-array, or a multi-dimensional array of photovoltaic modules.

Another embodiment of the invention provides a photovoltaic module, comprising: a plurality of photovoltaic cells; and a transparent cover to protect the photovoltaic cells from environmental degradation, wherein the transparent cover is coated on an inner surface with wavelength converting nanoparticles.

Advantageously, the use of a coating means that no other changes to the optical, electrical or other properties of the photovoltaic module are necessary. The coating can therefore be straightforwardly applied to transparent covers that employ an anti-reflective coating on the outer surface. Furthermore, the coating is protected from environmental degradation on the inner surface (a surface facing the photovoltaic cells) of the transparent cover.

In one embodiment, the nanoparticles are formed as a spin-coated layer on the inner surface of the transparent cover. This provides a straightforward application technique.

Another embodiment of the invention provides a method of manufacturing a photovoltaic module having a plurality of photovoltaic cells and a transparent cover to protect the photovoltaic cells from environmental degradation, the method comprising: coating a surface of the transparent cover with wavelength converting nanoparticles; and arranging the transparent cover over the photovoltaic cells so that the surface which is coated with the wavelength converting nanoparticles faces the photovoltaic device.

In one embodiment, coating the surface comprises spin-coating a solution having a ZnCuInS/ZnS core/shell quantum dot concentration of about 25 mg/mL and a PMMA concentration of about 0.02 mg/mL onto the surface of the transparent cover.

BRIEF DESCRIPTION OF THE DRAWINGS

Certain examples are described in the following detailed description and in reference to the drawings, in which: Fig. 1 is a photovoltaic module according to an example; Fig. 2 shows the external quantum efficiency of a typical crystalline Si solar cell at 1 sun (1000 W/m2); Fig. 3 is a comparison of a bare c-Si photovoltaic cell (solid line) and a photovoltaic cell with a wavelength converting layer on top (dashed line); Fig. 4 shows maximum power density comparison for different configurations. The error bars represent the standard deviation; Fig. 5 shows V00 (lower part) and Jsc (upper part) of different devices; Fig. 6 shows maximum power density comparison of different configurations for samples cleaned by 02 plasma ashen Fig. 7 shows V. (lower part) and J. (upper part) of different plasma cleaned devices; Fig. 8 shows a photoluminescence contour map for the perovskite QD stock solution; Fig. 9 shows perovskite quantum dot emission at A=310nm of excitation wavelength; Fig. 10 shows a photoluminescence contour map for the ZnCuInS/ZnS quantum dot stock solution; Fig. 11 shows ZnCuInS/ZnS quantum dot emission at A=240nm of excitation wavelength; Fig. 12 shows aggregates of perovskite quantum dots; Fig. 13 shows a top view of a perovskite stock solution sample along with the data points used for the EDX analysis; Fig. 14 shows aggregates of ZnCuInS/ZnS quantum dots along with the data points used for the EDX analysis; Fig. 15 shows agglomeration signs and EDX data points of a sample; Fig. 16 shows transmission (dashed), reflection (solid) and absorption (dotted) of a perovskite stock solution; Fig. 17 shows transmission (dashed), reflection (solid) and absorption (dotted) of a ZnCuInS/ZnS stock solution; Fig. 18 shows photoluminescence (dotted), absorption (dotted and dashed), EQE of the solar cell (dashed) for perovskite quantum dots; Fig. 19 shows photoluminescence (dotted), absorption (dotted and dashed), EQE of the solar cell (dashed) for ZnCuInS/ZnS quantum dots; Fig. 20 shows topography AFM images of the quartz superstrate; Fig. 21 shows phase AFM images of the quartz superstrate; Fig. 22 shows a topography AFM images of the quartz + PMMA sample; Fig. 23 shows a phase AFM images of the quartz + PMMA sample; Fig. 24 shows topography AFM images of the quartz + PMMA+QDs sample; Fig. 25 shows phase AFM images of the quartz + PMMA+QDs sample; Fig. 26 shows topography AFM images of the drop casted sample; Fig. 27 shows phase AFM images of the drop casted sample; Fig. 28 shows quartz superstrate profile; Fig. 29 profile of a quartz superstrate coated with PMMA (Sample 30). Positive step height is visible (shaded area); Fig. 30 shows profilometry results of a superstrate coated with PMMA and QDs. Positive step height is visible (shaded area);

DETAILED DESCRIPTION

With reference to Fig. 1, a photovoltaic module 100 according to an example includes a plurality of interconnected photovoltaic cells 102 (only one of which is shown) and a transparent cover 104 which is coated on its inner surface with a layer 106 of luminescent down-shifting nanoparticles. The photovoltaic module 100 may also include other components that are not shown in Fig. 1, such as a weatherproof backing and a frame suitable for mounting.

In this particular example, the photovoltaic cells 102 are crystalline silicon (c-Si) photovoltaic cells. However, other examples of photovoltaic cells 102 that could be employed include thin film photovoltaic cells such as cadmium telluride (CdTe) photovoltaic cells, copper indium gallium selenide (GIGS) photovoltaic cells, and amorphous silicon (a-Si) photovoltaic cells. Further examples of photovoltaic cells 102 that could be employed include perovskite photovoltaic cells, tandem photovoltaic cells, dye-sensitised photovoltaic cells, and organic photovoltaic cells.

The transparent cover 104 is the outermost layer of the photovoltaic module 100 and protects the remaining structure from the environment. In this particular example, the transparent cover 104 comprises a glass cover. However, it will be appreciated that any suitable material that allows incident sunlight to be transmitted to the underlying photovoltaic cells can be employed.

The luminescent down-shifting nanoparticles of the layer 106 may comprise quantum dots. Two examples are perovskite quantum dots and ZnCuInS/ZnS quantum dots. In either case, the quantum dots may be embedded in a polymer matrix such as polymethyl methacrylate, PMMA, which advantageously exhibits transparency in a broad range of wavelengths.

Experimental data that are considered to be indicative of real-world behaviour will now be described.

The first step of fabricating samples representative of the photovoltaic module 106 was the cleaning of superstrates. Two types of superstrates were used, namely quartz (due to its transparency in UV and low fluorescence) and glass (due to its relatively good optical properties and low cost). Regardless of superstrate, the cleaning procedure was the same. Initially the slides were placed in a beaker, filled with acetone and sonicated for 15 minutes. Next, the acetone was rinsed off with isopropyl alcohol (IPA) and the superstrates were blow dried with nitrogen gas. At later stages, two extra cleaning steps were explored. The first one was plasma ashing, where the superstrates were placed inside a plasma cleaner at 50W for two minutes. Spinning with IPA was another cleaning step that was investigated and due to its effectiveness was implemented at the cleaning procedure.

An exemplary method of forming a quantum dot thin film that is representative of the layer 106 comprises providing an ink formulation and depositing the ink solution by spin-coating.

In examples, the ink formulation was prepared by solution blending quantum dot and PMMA stock solutions. For the PMMA stock solution, 2mg of PMMA is mixed with 100mL of anisole under magnetic stirring for 8 hours. This PMMA stock solution is referred to herein as a varnish. Examples of quantum dot stock solutions are shown in the table below.

Quantum dots Solvent Concentration Perovskites Toluene 1mg/m L 0.15mg/mL ZnCuInS/ZnS Toluene 1mg/m L 0.15mg/mL 25mg/mL Examples of solution blending are provided in the table below.

Ink Formulation Method Ink 1 Tetrahydrofuran (THF):30% Toluene: 50% QD stock solution:1% Varnish:1 9% 1. THE and Toluene blended under gentle agitation; 2. QDs added and mixing continued until completely dissolved; and 3. Varnish added while stirring, until clear'.

Ink 2 QD stock solution ---Ink 3 Varnish: 180pL 1. 180pL of varnish added in a volumetric tube; QD stock solution: 20pL 2. QD stock solution sonicated for 30s; 3. 20pL of QD stock solution added to the volumetric tube; and 4. Ink gently shaken.

The ink formulation can then be spin-coated. Example spin-coating parameters are shown in the table below.

Acceleration time Os Spin time 60s Spin speed 800rpm Deceleration time Os In order to assess the performance of the thin films, an Oriel LCS-100 solar simulator fitted with an AM 1.5G filter was employed. A Keithley 2400 SMU, responsible for the data acquisition, completed the set up. The external quantum efficiency of a typical crystalline Si (c-Si) photovoltaic cell at 1 sun (1000 W/m2) can be seen in Fig. 2. From this figure it can be seen that luminescent downshifters that emit in the range of about 500-800 nm are suitable for photovoltaic cells such c-Si photovoltaic cells.

Current-voltage (I-V) values were recorded for a bare c-Si photovoltaic cell, to serve as a baseline in order to identify the increase in the efficiency of the cell. Subsequently, the samples were placed on top of the solar cell, with the coated surface placed in contact with the PV as shown in Fig. 1, and the characteristic I-V curves acquired. With the aid of the solar simulator the..15c, V00, FF and PCE were calculated and compared against the baseline values.

A Fluoromax-4 by Horiba was employed in order to detect the luminescence of the thin films at the solid state. Initially, coarse settings as seen in the table below were employed. Later on, the scans were refined according to the results of the coarse scan and the quantum dot used in each film.

Experiment type 3D Acquisition (Excitation vs Emission vs Intensity) Rayleigh masking 1st order enable TRUE Rayleigh masking 2nd order enable TRUE EX1 Excitation 1 (Mono 1) Start 250nm End 1000nm Increment lOnm Front entrance slit 1 nm Bandpass Front exit slit 1 nm Bandpass Grating density 1200 (Blaze: 330) EM1 Emission 1 (Mono 2) Start 400nm End 1000nm Increment 2nm Front entrance slit 1 nm Bandpass Front exit slit 1 nm Bandpass Grating density 1200 (Blaze: 500) Detector S (SCD1) Units CPS Corrected C:/Program Files (x86)Uobin Yvon\Corr\Mcorr.SPC Detector R (SCD6) Units MicroAmps Corrected C:/Program Files (x86)1Jobin Yvon\Corr\Xcorr.SPC Detector algebra formula 1. S1/R1 2. S1c/R1c Detector accumulations 1 Accumulations mode None Detector cycles 1 High resolution images and composition analysis (EDX) were obtained using a field emission gun-SEM (FEG-SEM). The accelerating voltage used varied from 10 to 20 kV in order to obtain proper contrast.

Atomic force microscopy was used in order to ascertain the topology of the thin films. The equipment used was a Bruker Dimension 3100 equipped with a non-contact, high resonance, frequency probe in tapping mode. The parameters can be found in the table below.

Feedback Integral gain 0.5 Proportional gain 0.7 Amplitude setpoint 399.0 mV Drive frequency 313.609 kHz Drive amplitude 164.2 mV Scan Scan size 10.0 pm Aspect ratio 1.00 X offset 0.000nm Y offset 0.000nm Scan angle 0.00° Scan rate 0.498 Hz Tip velocity 9.96 pm/s Samples/line 2048 Lines 2048 Slow scan axis Enabled XY closed loop Off Other Z limt 7.748 pm Zrange 9.31 pm Microscope mode Tapping Tip bias control Ground Samples bias control Ground In order to optically characterise the photovoltaic module 106, UV-Vis tests were performed. The spectrophotometer used was the Jasco v-670 UV-Vis-NIR. Two sets of tests were conducted, transmission and reflection. Regarding the transmission tests, all the data were normalised with their respective superstrate. For the reflection measurements the baseline was a fluoropolymer-based spectralon (from Labsphere). Furthermore, an integration sphere was used, so as to increase the accuracy of the results. The parameters used for the transmission and reflection tests are presented in the table below.

Instrument name V-670 V-670 Model name V-670 V-670 Serial No. B065561154 B065561154 Accessory USE-753 ISN-723 Accessory S/N B065561154 8024461118 Cell length 10 mm Not Use Photometric mode VoT %R Measurement range 1500 -200 nm 1500 -200 nm Data interval 0.5 nm 0.5 nm UVNis bandwidth 2.0 nm 5.0nm NIR bandwidth 8.0 nm 20.0 nm Response Medium Medium Scan speed 400 nm/min 400 nm/min Change source at 340 nm 340 nm Change grating at 850 nm 850 nm Light source D2/WI D2/VVI Filter exchange Step Step Correction Baseline Baseline The results obtained from the reflection and transmission tests were employed to quantify the absorption of each sample according to the following equation: Transmission+Ref lection+Absorption=100% As can be seen from this equation, an assumption was made that the whole procedure was free of any interference.

Dektak XT by Bruker was employed in order to identify the thickness of the thin films. As the coatings produced were not uniform, this feature was used in order to identify the thickness of the produced thin film. A trace of 20mm was registered starting from an uncoated side of the superstrate, going over the thin film and finishing again at an uncoated part. This way a clear step was recorded each time. Glass superstrates were avoided due to their curvature and surface roughness. The settings used in the experiments can be seen in the table below. Attention was paid to use a small force of interaction between the sample and the stylus tip, due to the soft nature of the film. Lastly, it is pointed out that this type of profilometry produces a relative and not absolute measurements. Stylus profilometry settings are shown in the table below.

Type of scan Force of stylus [mg] Trace length [mm] Range [pm] Resolution [pm] Quick analysis 2 20 65.5 0.333 As noted above, the samples that produced the most remarkable results were the ones employing the ZnCuInS/ZnS core/shell quantum dots with a loading of 25mg/mL. These will now be described.

The characteristic I-V and P-V curves were extracted, as the first step to assess the performance of the plain c-Si solar cell 102 (Fig. 3 solid lines). From the graph, the V., Ise, Imp and Vmp were identified, allowing for the calculation of the PCE and FF. These figures of merit are necessary in order to quantify the performance of the PV.

Later the fabricated devices were placed on the c-Si cell and again, the generated photocurrent was recorded with the aid of the solar simulator.

As can be seen in Fig. 3 (dotted lines) a decrease in both Ise and P. was observed upon the application of the device while the V. increased. This increase can be explained because the layer 106 reduces the temperature of the PV by absorbing the high energy photons responsible for thermalisation.

From the graph above the maximum power density was calculated. In order to statistically verify the results, three samples were made and compared against the same number of all other configurations, as can be seen in Fig. 4.

In Fig. 5 two different trends can be observed. Starting from the upper part of the graph, where the J. is visible, one can notice that by adding the glass superstrate, there is a significant decrease. This can be explained as a reduction in photocurrent, meaning that less photons reach the active area of the solar cell possibly due to losses induced by the superstrate. The trend remains the same by the addition of PMMA. If one notices the response of the device when the QDs are added to the film, one may see that initially (samples 27a-c) there is no difference comparing to samples 26a-c. By adding the extra cleaning step, the wettability of the superstrate and the photocurrent both increase. This slight increase could be a result of the QDs down-shifting high energy photons. Carrier mobility is also increased by the decrease in temperature. Therefore, an increase in photocurrent generation might have its origins there. Photoluminescence tests need to be done in order to verify the origin of this increase in J..

Turning the attention to the lower part of the graph (V.), the opposite trend is observed. By adding the glass layer, the V. is increased. This finding suggests that the thermalisation effect on the PV is reduced.

These evidences suggest that, at this stage, the device seems to behave like a luminescent downshifting layer, due to the increase in J..

The table below shows a comparison of the figures of merit for the five different configurations discussed above.

Configuration Voc [V] Jsc [A/m2] FF [%] Maximum power density [W/m2] PCE [%] PV 6.48 ± 0.01 24.37 ± 0.02 60.73 ± 0.29 95.80 ± 0.7 9.58 ± 0.03 PV+Glass 6.72 ± 0.05 22.33 ± 0.14 62.00 ± 0.49 93.06 ± 0.65 9.31 ± 0.07 PV+Glass+PM MA 6.72 ± 0.05 22.32 ± 0.23 61.87 ± 0.62 92.75 ± 0.78 9.28 ± 0.078 PV+Glass+PM MA+QDs 6.76 ± 10-5 22.21 ± 0.11 62.50 ± 0.26 93.84 ± 0.83 9.38 ± 0.08 PV+ extra cleaned Glass +PMMA+QDs 6.76 ± 10-5 22.49 ± 0.21 62.23 ± 0.19 94.60 ± 0.87 9.46 ± 0.09 From the results above, it becomes clear that the cleaning process of the superstrates prior to deposition can play a role in the performance. In this occasion, apart from the standard cleaning procedure, IPA was spin coated on the superstrates and was let dry inside the spin coater. Noteworthy relative increase was also achieved when the superstrates were plasma cleaned. The results can be seen in Fig. 6.

Even though the error bars (standard deviation) might look substantial, if one notices the scale, one will see that the biggest error bar (green) is only 2.0W/m2. The table below shows the figures of merit for the devices cleaned by the 02 plasma asher.

Configuration Voc [V] J. [A/m2] FF [%] Maximum power PC E [%] density Lwinizi PV 6.52 ± 0.04 24.50 ± 0.14 59.90 ± 0.31 95.63 ± 0.66 9.56 ± 0.07 PV+ plasma cleaned Quartz 6.71 ± 0.06 22.95 ± 0.31 61.74 ± 0.92 95.04 ± 0.96 9.50 ± 0.10 PV+ plasma cleaned 6.72 ± 0.02 23.13 ± 0.10 61.35 ± 0.10 95.31 ± 0.35 9.53 ± 0.03 Quartz+PMMA PV+ plasma cleaned Quartz+PMMA+ QDs 6.76 ± 0.01 22.84 ± 0.05 61.90 ± 0.24 95.50 ± 0.21 9.55 ± 0.21 The trend observed for the J. and V. in this occasion is somewhat different than the one for the glass substrates (Fig. 7). While the V. increases, suggesting reduction in thermalisation losses, the J. for the film without quantum dots (samples 30a-c) is larger than the one with quantum dots (samples 29a-e).

It becomes clear once again, that cleaning the superstrate methodically can increase the overall performance of the device by increasing the wettability and therefore the coated area of the superstrate.

It is observed that the majority of the samples tested performed better than the uncoated superstrate, meaning that they were able to provide a relative increase that overcame the initial substantial losses induced by the superstrate. Two possible reasons have been identified in order to explain this relative increase. The first is that the quantum dots absorb the UV photons and by the down-shifting mechanism emit lower energy photons. The second reason could be that the quantum dots absorb the high energy photons but do not emit any lower energy photons. Instead, they increase the maximum power output by reducing the thermalisation losses of the PV. Photoluminescence tests can help identify which of the two possible explanations is more accurate. Summarising the results shown in the two tables above, it can be seen that the former explanation is plausible. In fact, the results shown in Fig. 5 indicate that the QD film is indeed down-shifting high energy photons, though the increase in Jsc is marginal.

When the plasma cleaning procedure was employed as shown in Fig. 7, the marginal increase was absent. That could mean that the film is not performing as expected and acts only as an absorption layer for high energy photons, contributing only to an increase in PCE by reducing the thermalisation effect. Lastly, it is pointed out that the cleaning procedure, apart from reducing the impurities in the surface of the superstrate, affected the wettability and therefore the uniformity of the film.

Fluorescence emission-excitation matrices were recorded, for the perovskite stock solution and a device consisting of a quartz superstrate PMMA and ZnCuInS/ZnS QDs in solid state. The results for the perovskite quantum dots can be seen in Fig. 8.

Emission was observed in the range between 510-540nm. According to the manufacturer's specification sheet, the expected emission of the quantum dots should be 530±15nm. Taking a closer look to the graph above one can observe that the emission is focalised at 518nm, for excitation wavelengths ranging from 260-360nm. The Stokes shift of these quantum dots ensure minimum reabsorption losses. While the quantum dots are still within specifications, the experimental data suggest that the QDs size distribution is shifted to smaller values (emission focalised at 518nm) while there are signs of agglomeration (red-shifted emission). More specifically the maximum emission was found to be at 310nm of excitation wavelength as is visible in Fig. 9.

Fig. 10 shows the contour map for the ZnCul nS/ZnS quantum dot stock solution. Emission is observed in the range between 700-850nm (focalised around 760nm). According to the manufacturer's specification sheet, the expected emission of the quantum dots should be 700±25nm. In this occasion the quantum dots present a considerable red-shift, which is attributed to the agglomeration. The excitation wavelengths that result in emission were limited, ranging from 240-250nm. The Stokes shift observed in these quantum dots is immense, thus no reabsorption was expected. One might also observe that the emission is abruptly cut at 850nm, the reason behind this, is that the correction of the detector is limited between 240850nm. A closer look at the emission of the ZnCuInS/ZnS quantum dots can be seen in Fig. 11. Maximum emission was pinpointed for excitation 240nm.

The results for the perovskite QDs depict the response of the stock solutions. That is because PMMA has showed its potential of encapsulating nanoparticles without inhibiting their quantum efficiencies.

The emission spectra of the QDs indicate the relative increase presented above is in fact because of the down-shifting mechanism. Of note though, is that the emission was rather faint for the perovskite quantum dots and the ZnCuInS/ZnS quantum dots were excited by a very low wavelength. In fact, the solar intensity at 240nm is significantly low. This can explain the fact that the relative increase was not able to overcome the initial loss from the glass or quartz superstrate. SEM and EDX analysis will be discussed below in order to identify the level of agglomeration, along with the composition of the QDs.

The recorded red-shift in the photoluminescence data suggested the existence of aggregates on the devices. SEM was employed in order to identify the scale of said agglomeration. Subsequently, EDX analysis was performed, in order to investigate the composition of the quantum dots.

Initially, samples deposited with quantum dot stock solutions were analysed. The agglomeration for the perovskite quantum dots can be seen in Fig. 12. The aggregates were found to be approximately 20 nm to 300 nm in size.

The table below shows results of EDX analysis of a perovskite stock solution sample. Area of spectra 1, 2 and 3 is given in Fig. 13.

Spectrum Label Spectrum 1 [Atomic °/0] Spectrum 2 [Atomic °A] Spectrum 3 [Atomic °A C 52.70 56.94 0 33.60 65.75 33.33 Si 11.28 34.25 9.70 Br 1.27 I 0.12 Cs 0.58 Pb 0.45 0.03 Total 100.00 100.00 100.00 Fig. 14 shows the agglomeration and the points used to perform the EDX measurements, on a quartz superstrate coated with ZnCuInS/ZnS stock solution. The aggregates were found to be approximately 170 nm in size.

The table below presents EDX analysis of a ZnCul nS/ZnS stock solution sample. Area of spectra 10, 6 is given in Fig. 14.

Spectrum Label Spectrum 10 Atomic [%] Spectrum 6 Atomic [%] C 22.67 0 21.30 23.83 Si 48.21 76.17 S 1.20 Cu 142 Zn 3.15 In 2.06 Total 100.00 100.00 SEM/EDX analysis was also performed on one of the best performing samples, but the images were quire blurry due to the effect of the PMMA. Agglomeration was still present as can be seen in Fig. 15 while the EDX analysis confirmed, once again, the composition of the QDs.

The table bellows shows EDX analysis of the sample shown in Fig. 15.

Spectrum Label Spectrum 4 Spectrum 5 Atomic [%] Atomic [%] C 71.70 54.95 0 17.79 32.68 Si 4.17 12.37 S 2.58 Cu 0.98 Zn 1.93 In 0.84 Total 100.00 100.00 From the results presented above, it is evident that the emission characteristics (red-shift due to agglomeration) of the devices are in accordance with the SEM data. In addition to that, the EDX analysis revealed the composition of the perovskite quantum dots and confirmed the composition of the ZnCuInS/ZnS. Nevertheless, it needs to be noted that the nature of the QD ligands is still not known.

UV-Vis spectroscopy was employed in an effort to identify the optical response of the films in terms of reflection, transmission and absorption, at different wavelengths. Figs. 16 and 17 respectively show the optical response of the perovskite stock solution and the ZnCuInS/ZnS.

On the perovskite quantum dots two absorption peaks are visible at 280nm and 510nm. For the ZnCuInS/ZnS quantum dots while the peaks are not as pronounced, they are spotted at 300nm and at 550nm.

Combining the data derived from the photoluminescence and an EQE plot acquired from the manufacturer Figs. 18 and 19 were created.

Comparing the absorption of the perovskite QDs against the EQE of the solar cell (Fig. 18), it is obvious that even though the nanoparticles have a peak in a region where the EQE is nonexistent (280nm) the second peak lies within the PL curve. This could possibly result to reabsorption losses. Also, the fact that the emission peak is quite narrow needs to be pointed out. It needs to be noted that he photoluminescence data were normalised by the maximum photoluminescence intensity and are presented in Figs. 18 and 19 as a percentage of said intensity.

Focusing on the ZnCul nS/ZnS QDs (Fig. 19), even though there are no signs of reabsorption, since the absorption peaks do not interfere with the emission peak, the absorption of the QDs even though present, is characterised as weak. In addition, the peak is situated in an area where the PV is already sensitive (600nm).

Atomic force microscopy tests were conducted in order to acquire details about the surface of the samples. Scans were performed on the quartz superstrate, the superstrate coated with the PMMA varnish and lastly to the complete device. Lastly, a scan was performed on a sample of drop casted stock solution of ZnCuInS/ZnS quantum dots. The data acquired were of two types, topography and phase.

As can be seen in Fig. 20, the superstrate itself is quite uniform in terms of topography, since the maximum peak to valley difference is lOnm. One can spot that this difference is due to a particle depicted in white in Fig. 20. Considering Fig. 21, one can see that the particle is of the same phase as the rest of the superstrate. This means that this particle is actually debris of the superstrate.

Figs. 22 and 23 respectively show the topography and the phase of the sample coated with the polymer. On the topography front, the sample appears to be uniform, apart from some white spots on the middle and upper left side. Crosschecking with the phase analysis, it becomes clear that these spots are in fact undissolved PMMA powder.

Subsequently the sample that consisted of PMMA and ZnCuInS/ZnS quantum dots was tested and the results can be seen on Figs. 24 and 25. The powder particles of undissolved PMMA are present and they manifest themselves both in the topography (Fig. 24) and in the phase (Fig. 25).

Figs. 26 and 27 show the topography and phase of the drop casted sample. One can notice that in terms of topography (Fig. 26), there is some significant difference (-300nm) between the peaks and valleys of the sample. Moving to the phase data (Fig. 27) two phases are visible, one of these phases appears highly crystalline. These crystalline phases are the ones that appear to be responsible for the fluctuations on the surface in terms of topography. These results, along with the agglomeration evidence from the SEM, suggest that these crystalline phases are in fact quantum dots.

Moving on to the profilometry measurements, at the beginning, in order to cross check the data derived from the AFM a bare quartz superstrate was examined. The results can be seen in Fig. 28. One can notice that the superstrate is quite flat, apart from some single point spikes that are attributed to dust or surface impurities as quartz debris (as shown in the AFM results).

The next logical step was to investigate a sample coated with PMMA. The results are visible in Fig. 29. The data showed that the produced film had a thickness of 87nm ± 24nm. The presence of single point spikes is visible here as well. An important notice is that the intensity of the peaks is the same for the uncoated quartz and the sample coated with PMMA.

Lastly, the best performing sample was tested. Fig. 30 shows that the thickness achieved in this occasion was 91nm ± 150nm. As it was mentioned before, the glass superstrate due to its roughness and curvature was producing very inconsistent results and therefore was not considered.

Reflecting upon the results, one can argue that the errors of the measurements are too elevated. Indeed, that is the case, though if one takes a look at the graphs, it is visible that these single point spikes play a major role. The intensity of some is similar to the one recorded for the bare superstrate, so one can reason that the spikes originate from the superstrate. Bearing in mind the results recorded by the AFM, one needs to take into account the effect of undissolved PMMA powder to the height profile. Similarly, the agglomerates of QDs as shown in the SEM images occasionally reached 1pm. One cannot exclude the possibility of noise in the tip of the stylus that could heavily affect the results. In addition, the step is what is of importance here. Lastly, since profilometry was the last test that was conducted, impurities and contamination of the samples might have also affected the results.

In summary then, and as is evident in Figs. 4 and 6, thin film with quantum dot aggregations improves the PCE of photovoltaic devices (the lesser power output than the bare PV has its origins in the small, but significant absorption in the visible range of the quartz substrates used to obtain the experimental data).

The Summary and Abstract sections may set forth one or more but not all exemplary embodiments of the present invention as contemplated by the inventor(s), and thus, are not intended to limit the present invention and the appended claims in any way.

The breadth and scope of the present invention should not be limited by any of the above-described exemplary embodiments, but should be defined only in accordance with the following claims and their equivalents.

Claims (15)

1. CLAIMS1. A photovoltaic device, comprising: a photovoltaic cell; and a wavelength converter on a light incident side of the photovoltaic cell, the wavelength converter comprising aggregates of nanoparticles that absorb light at a first wavelength and re-emit the light at a second, longer wavelength.
2. 2. A photovoltaic device according to claim 1, wherein the nanoparticles comprise nontoxic nanoparticles.
3. 3. A photovoltaic device according to claim 1 or 2, wherein the nanoparticles comprise quantum dots.
4. 4. A photovoltaic device according to claim 3, wherein the quantum dots comprise ZnCuInS/ZnS core/shell quantum dots.
5. 5. A photovoltaic device according to claim 3, wherein the quantum dots comprise perovskite quantum dots.
6. 6. A photovoltaic device according to any one of claims 3 to 5, wherein the quantum dots are embedded in a polymer matrix.
7. 7. A photovoltaic device according to claim 6, wherein the polymer matrix comprises polymethyl methacrylate.
8. 8. A photovoltaic module, comprising: a plurality of photovoltaic cells; and a transparent cover to protect the photovoltaic cells from environmental degradation, wherein the transparent cover is coated on an inner surface with wavelength converting nanoparticles.
9. 9. A photovoltaic module according to claim 8, wherein the wavelength converting nanoparticles are formed as a spin-coated layer on the inner surface of the transparent cover.
10. 10. A method of manufacturing a photovoltaic module having a plurality of photovoltaic cells and a transparent cover to protect the photovoltaic cells from environmental degradation, the method comprising: coating a surface of the transparent cover with wavelength converting nanoparticles; and arranging the transparent cover over the photovoltaic cells so that the surface which is coated with the wavelength converting nanoparticles faces the photovoltaic device.
11. 11. A method according to claim 10, wherein the nanoparticles comprise quantum dots.
12. 12. A method according to claim 11, wherein the quantum dots comprise ZnCuInS/ZnS core/shell quantum dots.
13. 13. A method according to claim 12, wherein the quantum dots are embedded in a polymer matrix.
14. 14. A method according to claim 13, wherein the polymer matrix comprises polymethyl methacrylate, PMMA.
15. 15. A method according to claim 14, wherein coating the surface comprises spin-coating a solution having a ZnCuInS/ZnS core/shell quantum dot concentration of about 25 mg/mL and a PMMA concentration of about 0.02 mg/mL onto the surface of the transparent cover.