An Updated Review for Performance Enhancement of Solar Cells by Spectral Modification

Abstract

Photovoltaic technology has become one of the major renewable ways to generate electric power. However, the mismatch between the incident solar spectrum and photo-electric response efficiency of solar cells severely constrains their performance. Hence, spectral modification technologies, e.g., up-conversion (UC), down-conversion (DC), and luminescent down-shifting (LDS) technologies have been applied widely in the photovoltaic field to reform the incident spectrum to match the best response band possible. In this paper, we review the latest developments of the three technologies above in terms of material selection, optical characteristics, and photovoltaic performance. It is found that the three most popular materials for conversion are NaYF₄: Er³⁺, Yb³⁺, and Yb³⁺. The excitation bands for the three technologies are 800–1550 nm, 250–488 nm, and 250–488 nm, respectively, while the emission bands are 523–669 nm, 520–1031 nm, and 490–1010 nm, respectively. Furthermore, issues hindering the development of spectral modification technologies are pointed out, e.g., low absorption efficiency, poor quantum conversion efficiency, and hurdles in commercialization. Finally, suggestions and solutions to address the above-mentioned issues are provided.

1. Introduction

As global energy consumption escalates, nations increasingly prioritize the development of clean, renewable energy sources. Solar energy, among the foremost known renewable energies, holds immense promise. By the end of 2023, global photovoltaic capacity soared to 444 gigawatts. However, current photovoltaic cells achieve only about 27% efficiency at best. For instance, in the case of monocrystalline silicon photovoltaic cells, according to the Shockley–Queisser detailed balance theory [1], the theoretical maximum efficiency is 30%. Most of the unused energy is converted into heat, which is shown in Figure 1, significantly reducing the electrical efficiency of photovoltaic modules. In the case of monocrystalline silicon solar cells, to reduce energy losses, the first priority is to ensure that as many photons as possible with energies greater than the bandgap of silicon enter the cell. Secondly, these photons should ideally match the photovoltaic response efficiency of the silicon solar cell, as illustrated in Figure 2. However, the incident AM1.5 spectrum poorly overlaps with the photovoltaic response efficiency of monocrystalline silicon solar cells and fails to meet either of these criteria above. This phenomenon arises primarily because crystalline silicon cells exhibit poor response to a significant portion of short-wavelength photons. For instance, the PN junction depth of the crystalline silicon solar cell is about 400 nm [2], while photons at 460 nm only have a penetration depth of 281.58 nm [3,4,5], so they are not able to reach the depletion region of PN junction. Thus, few electron–hole pairs can be induced by these short-wavelength photons, causing the poor photovoltaic response of crystalline silicon solar cell. Therefore, an artificial “spectral modification” of the original spectrum is necessary.

In the field of photovoltaic, there are currently three main technologies for spectral modification: up-conversion (UC), down-conversion (DC), and luminescent down-shifting (LDS) [7]. The principle of UC technology involves combining two or more photons into one higher-energy photon. Conversely, DC technology, also known as quantum cutting [8], splits one high-energy photon into two or more lower-energy photons. LDS technology directly converts one high-energy photon into one lower-energy photon. Among the three technologies, UC and DC can successfully exceed the Shockley–Queisser limit, whereas the quantum efficiency of LDS technology only reaches a maximum of 100%, which is not able to supersede the S-Q limit [9].

The position of spectral modification layer relative to the cell for UC, DC, and LDS technologies is different. Photons with energies equal to or greater than the bandgap width can be absorbed by the cell, while photons with energies less than the bandgap width will pass through it [10]. To ensure that the photons about to be modified effectively enter the spectral modification layer without being absorbed by the cell, the sequential arrangement of the spectral modification layer and the cell varies depending on the technology used, as depicted in Figure 3 [11]. In UC technology, to minimize absorption of high-energy photons by the UC layer, it is common practice to place the UC layer beneath the photovoltaic cell. This allows the photovoltaic cell to first utilize high-energy photons to generate electron-hole pairs. Subsequently, low-energy photons pass through the cell, converted into high-energy photons by the UC layer, and then reflected back into the photovoltaic cell. In DC and LDS technologies, the approach is quite reversed. Since photovoltaic cells applied in the two technologies above cannot effectively utilize high-energy photons, the DC/LDS layer is typically placed above the cells. This allows high-energy photons to be converted into lower-energy photons before entering the photovoltaic cells, thereby generating electron–hole pairs at a high efficiency.

This paper primarily reviews the recent advancements in UC, DC, and LDS technologies, summarizing their materials selection, optical characteristics, and application effects. Moreover, we have shared our perspectives and improvement suggestions regarding the issues exposed by UC, DC, and LDS technologies, providing some insights for further research among peers.

2. Brief Introduction of Photovoltaic Cells and General Guidelines of Spectral Modification Materials

2.1. Classification of Solar Cells

In 1954, the world’s first solar cell was created at Bell Labs in the United States, marking the advent of photovoltaic technology. Since then, photovoltaic cells have evolved through three generations: silicon-based solar cells, thin-film solar cells, and the third-generation solar cells. Thin-film solar cells operate similarly to silicon-based solar cells, both based on the photovoltaic effect of semiconductor PN junctions. Compared to traditional silicon-based solar cells, thin-film solar cells are more flexible, easier to install, and perform better under weak light conditions. However, their efficiency is lower compared to silicon-based solar cells. The third-generation solar cells primarily include perovskite cells (PSCs), dye-sensitized cells (DSSCs) and quantum dot sensitized solar cells (QDSSCs). PSCs are promising new generation of photovoltaic technology known for their high efficiency and low cost [12], which utilize perovskite-type organic metal halide semiconductors as the photon-electron conversion layer. PSCs are suitable for various applications including building-integrated photovoltaics and portable electronics. However, their weak stability under harsh environmental conditions and toxicity have significantly hindered their development. DSSCs utilize a photosensitive dye absorbed onto a semiconductor material, typically titanium dioxide (TiO₂). When light hits the dye, it excites electrons, which are then transferred through the semiconductor to generate an electric current. DSSCs are flexible and transparent, suitable for applications where conventional solar cells are not feasible, like curved surfaces or portable electronics. They offer low production costs and easy fabrication processes [13] but are less efficient compared to silicon-based cells and can suffer from dye degradation and electrolyte leakage. QDSSCs use quantum dots like cadmium selenide (CdSe) or lead sulfide (PbS) semiconductor nanocrystals as the light-absorbing material. When photons are absorbed, electrons are excited across the quantum dot’s bandgap, creating electron–hole pairs. These pairs are then separated and collected to generate electricity. QDSSCs are notable for their tunable bandgaps, potentially high efficiency, and solution-processability [14]. They are envisioned for applications requiring high efficiency in small form factors, but concerns over toxicity of certain quantum dot materials and stability issues remain challenges. Hence, to attain superior photovoltaic efficiency, we should select appropriate spectral modification technologies and materials according to the various types of solar cells.

2.2. General Guidelines of Spectral Modification Materials

Based on the conclusion given by Suyver [15], radiative relaxation is dominant when the reduced energy gap to the next lowest level is larger than five high-energy phonons, while for smaller gaps, non-radiative multi-phonon emission becomes the dominant depopulation mechanism. Hence, to achieve higher quantum yield efficiency, materials with lower phonon energy should be selected as the matrix from among the three spectral modification technologies [16]. In spectral modification layers, one or more rare earth ions are typically doped. Due to the shielding effect of the 5s²5p⁶ shell, the 4f electrons of rare earth ions are less affected by crystal fields. As a result, in a low phonon energy matrix, luminescence processes are much more competitive with multi-phonon relaxation within these ions. Additionally, owing to the diverse arrangements of electrons’ 4f orbitals and f–f electronic transitions, rare earth ions can exhibit rich and sharp lines. With these excellent optical properties, rare earth elements have been extensively researched in the spectral modification field during recent years. As a result, in spectral modification technologies, choosing appropriate rare earth ions doping in low phonon energy matrix can effectively enhance spectral modification efficiency.

3. Applications of Three Spectral Modifications to Photovoltaic Cells

3.1. Up-Conversion (UC) Technology

The bandgap width of perovskite solar cells (PSCs), thin-film solar cells, dye-sensitized solar cells (DSSCs), and other amorphous silicon (a-Si) solar cells is about 1.5 eV. Consequently, a significant number of photons with near-infrared wavelengths cannot be converted into electrons by these cells. Therefore, harnessing UC technology to exploit these unused photons holds substantial promise. Figure 4 illustrates several common types of up-conversion technology [17]. In ground-state absorption/excited-state absorption process (GSA/ESA), ion 1 consecutively absorbs two low-energy photons, followed by radiative relaxation, emitting a higher-energy photon. In the energy transfer up-conversion process (ETU), one ion serves as a sensitizer ion and another as an activator ion. Both ions simultaneously absorb a low-energy photon. Then, the sensitizer ion transfers energy to the activator ion. Subsequently, sensitizer ion returns to the ground state while the activator ion reaches a higher excited state, which later emits a higher-energy photon through radiative relaxation. In the co-operative energy transfer process (CET), similar to ETU, two ion 1s each absorb a low-energy photon, transferring absorbed energy to ion 2. This process excites an electron in ion 2 to a higher state, which then emits a higher-energy photon through radiative relaxation.

In 2003, François Auzel [18] first proposed the UC technology and provided detailed explanations of various energy transfer mechanisms such as ETU, ESA and photon avalanche. Additionally, his paper enumerates various doped elements used in UC technology, along with the classic experiments of that time. For subsequent scholars, Auzel’s work laid a solid foundation in mechanisms exploration and experiments of UC technology. With the accumulation of peers’ experiments in recent years, there has been an increasing variety of doped rare earth elements as well as materials and synthetic methods of the UC matrix. Consequently, the efficiency of UC has been steadily improved year by year.

Table 1. Recent developments of UC technology in solar cells.

Ref.	Materials	Solar Cell Type	Excitation Peak (nm)	Emission Peak (nm)	Emission Band (nm)	PCE Improvement (%)
[20]	NaYF4: Yb3+, Er3+	DSSCs	980	543 654	500–600 600–700	16
[21]	YF3: Yb3+/Er3+	c-Si	1550	653 669	625–700	13.26
[22]	β-NaYF4: Nd3+/Yb3+/Er3+	PSCs	980 808	523 541 659	520–570 640–675	7.13
[23]	TiO2: Yb, Er, F	DSSCs	980	525 545 658	515–560 640–690	37.8
[24]	NaYF4: Yb3+, Er3+	PSCs	-	-	-	20.6
[25]	GdBO3: Yb3+/Tb3+	QDSSSCs	980	545	450–650	-
[26]	ZnO-Al2O3-BaO-B2O3: Er3+	DSSCs	800	550	520–570	7.21
[27]	NaGdF4: Yb3+, Er3+	PSCs	980	415 525 540 660	510–560 630–690	10.6
			395	590 610 695	580–600 610–675 680–700
[28]	β-NaYF4: Yb, Er	PSCs	-	-	-	13.9
[29]	Li(Gd, Y)F4: Yb, Er	PSCs	980	520 540 650	515–530 540–550 645–620	25
[30]	TiO2: Ho3+, Yb3+, Li3+	PSCs	980	547 663	530–560 630–680	9.03
[31]	LiYF4: Yb, Er	PSCs	980	550 679	527–575 650–680	9.01
[32]	YLiF4: Yb, Er	PSCs	980	550 660	525–560 650–675	9.61
[33]	NaY(WO4)2: Er, Yb	Si	-	-	-	-

enumerates recent applications of UC technology in photovoltaic cells, including materials used in the UC layer, types of photovoltaic cells, absorption/emission bands, and power conversion efficiency (PCE). From the table, it can be seen that after doping with UC ions, the spectral modification material exhibits efficient emission from green to red light (523–669 nm) under excitation in the near-infrared range (800–1550 nm). This emission spectrum aligns well with the operational range of dye-sensitized and perovskite solar cells. Additionally, it is evident that Er³⁺ is highly favored among UC rare earth ions, attributable to its strong emission in the green to red light spectrum under excitation from 980 nm near-infrared light [19]. The emission peaks and corresponding energy level transitions of Er³⁺ are as follows: 536 nm: ⁴S_3/2 → ⁴I_15/2; 550 nm: ²H_11/2 → ⁴I_15/2; 655 nm: ⁴F_9/2 → ⁴I_15/2. When selecting a UC matrix material, most researchers opt for a sodium-based or lithium-based matrix, such as Na(Y,Gd)F₄ and Li(Y,Gd)F₄. Attributed to their low phonon energy, NaYF₄ and LiYF₄ can mitigate multiphonon relaxation phenomena and consequently enhance UC luminescence efficiency, as mentioned in Section 2.1.

In recent experiments on UC materials in solar cells, Zhang et al. [23] fabricated a titanium (Ti)-mesh-supported “TiO₂ nanowire arrays (TNWAs)/Yb-Er-F tri-doped TiO₂ UC nanoparticles (YEF-TiO₂-UCNPs)” composite structure, which demonstrated exceptional performance. Based on the photoluminescence spectrum of the UC material measured in Figure 5, the energy transfer mechanism within the material is inferred as shown in Figure 6. Then, Zhang’s team tested their new photovoltaic assembly under the simulated solar irradiation of 100 mW cm⁻² (AM1.5G), observing that the short-circuit current density (J_sc) and a power conversion efficiency of the flexible DSSC with a Ti mesh-supported “TNWAs/YEF-TiO₂-UCNPs_60%P25_40%” composite structure reached 12.81 mA cm⁻² and 5.54%, increasing by 45.7% and 37.8% as compared to TNWAs/P25_100% (8.79 mA cm⁻², 4.02%), respectively, which provided researchers studying DSSCs with an excellent idea and demonstration. The result of testing is shown in Figure 7.

Beyond DSSCs, UC technology has also achieved remarkable success in PSCs. Chen et al. [24] incorporated 25 wt% NaYF₄:Yb³⁺, Er³⁺ nanoparticles into the Al₂O₃ scaffold layer, making it as the UC layer of PSCs. Results showed that the conversion efficiency increased from 10.81% to 13.04%, achieving an improvement percentage of 20.6%, which is illustrated in Figure 8. It is noteworthy that when the mass content of NaYF₄:Yb³⁺, Er³⁺ NPs increases, the performance of the device drops rapidly. Although more infrared light can be up-converted into visible light with the increased concentration of NaYF₄:Yb³⁺, Er³⁺ NPs, its larger diameters will reduce the pore size of the scaffold layer, leading to a decrease in the loading amount of PSCs. Deng et al. [29] synthesized the UC nanophosphors with Li(Gd,Y)F₄:Yb, Er and then integrated them into the hole transport layers (HTLs) and examined their optical and electrical performance in PSCs. At an optimal weight ratio of UC nanophosphors, the PSCs achieved an average power conversion efficiency of 18.34% under AM1.5G illumination, obtaining an increase over 25% compared with conventional HTL based PSCs (14.69%). Furthermore, nanophosphors synthesized in this experiment acted as efficient UC centers under the irradiation of a 980 nm near-infrared laser.

3.2. Down-Conversion (DC) Technology

For crystalline silicon solar cells, with a bandgap width around 1100 nm, they exhibit better utilization of near-infrared photons compared to other solar cells. However, during the process of converting ultraviolet/short wavelength photons into electrons, significant residual heat can cause the cells to heat up, thereby reducing electric efficiency. Figure 9 illustrates several solar cells’ bandgap variation when temperature changes. For instance, when cell temperature increases by 100 K, bandgap of a commercial monocrystalline silicon cell will decrease by 2.3%, leading to about a 10% decrease in electrical efficiency [34]. Therefore, converting the high-energy photons which generate excessive thermal energy during the photovoltaic process into low-energy photons near the bandgap width of crystalline silicon cells through DC technology can greatly enhance cells’ efficiency. Figure 10 illustrates several common types of DC technology, which is almost the reverse mechanism of UC technology. In ground-state absorption/excited-state absorption (GSA/ESA), ion 1 absorbs one high-energy photon and subsequently emits two lower-energy photons. In energy transfer down-conversion (ETD) ion 1 absorbs one high-energy photon, transferring part of energy to ion 2, and then both ion 1 and ion 2 emit a lower-energy photon. In co-operative energy transfer (CET), ion 1, after absorbing one high-energy photon, transfers energy to two ion 2s through energy transfer (ET).

Recent applications of DC technology in photovoltaic cells are illustrated in

Table 2. Recent developments of down-conversion technology in solar cells.

Ref.	Materials	Solar Cell Type	Excitation Peak (nm)	Emission Peak (nm)	Emission Band (nm)	Improvement
[39]	Yb2.96−xYxHo0.04Al5O12	-	454	550	525–575	-
[40]	SrWO4: Pr3+	-	250 450	645	-	-
[41]	CaS: Eu2+, Sm3+	-	467	625	575–675	-
[42]	ZnO: Er, Yb	Si	378	550 670 980	490–685 790–880	-
[43]	chlorophyll (Chl)	c-Si	-	670	650–800	Maximum power increases by 9.2%.
[44]	ZnO nano-rods array solar cell: PbS/CdS	ZnO nano-rods array solar cell	378	520 543 658	-	PCE: 8.6%.
[45]	Phosphate glasses: Pr3+-Yb3+	-	457	598 1031	590–620 980–1050	-
[46]	70TeO2-20ZnO-(10−x)Nb2O5: Er3+	-	488	556 672 818 980	525–610 650–720 800–850	Quantum yield efficiency: 91%.
[47]	CdO-P2O5: Pr3+-Yb3+	c-Si	443	604 648 977	575–650 950–1050	Quantum yield efficiency: 144%.
[48]	Ca2SiO4: Ce3+-Yb3+	Si	323	980	870–1050	Efficiency of energy transfer: 25%.
[49]	TeO2-GeO2-PbO: Eu3+	Si	405	613	610–630	Efficiency increase: 11.81%.
[50]	CaF2: Nd3+/Yb3+	c-Si	353	975	900–1125	Quantum yield efficiency: 91%.
[51]	TiO2 nano-rod array: Ce	PSCs	-	-	-	PCE: 10.1%.
[52]	C30H21EuF9NNaO9	PSCs	-	580 592 613 653	575–585 585–600 605–630	PCE: 16.13%.
[53]	Na9[EuW10O36]	PSCs	290	590 625 650 700	580–600 600–670 680–720	PCE: 14.36%
[54]	CH3MH3PbI3	PSCs	-	-	-	PCE: 25%

. In contrast to UC technology, DC materials are commonly excited by the light ranging from ultraviolet to blue (250–488 nm), emitting efficiently in the green-to-near-infrared spectrum (520–1031 nm). Therefore, DC technology aids in converting photons with poor response in crystalline silicon cells into photons with higher photovoltaic response. It is noteworthy that crystalline silicon cells exhibit the most optimal response near 1000 nm, aligning well with the emission wavelength of 980 nm from Yb³⁺ through its ²F_7/2–²F_5/2 radiative transition, making Yb³⁺ a preferred activator in most DC experiments. Furthermore, to enhance DC efficiency, researchers often incorporate one or more co-doping sensitizers, such as rare-earth elements like Pr³⁺ and Ce³⁺, whose excitation bands align well with the poor response band of cell and emission band align well with the excitation band of Yb³⁺.

Due to the current incomplete understanding of the microscopic energy transfer mechanisms involved in DC, experimental efficiencies vary widely. More than half of the experiments report quantum yield efficiency below 100%, significantly diverging from the theoretical ideal efficiency of 200% for DC technology. Fortunately, W. Romero-Romo et al. conducted a relatively successful experiment in recent years. Their team utilized melt-quenching method to prepare a CdO-P₂O₅: Pr³⁺/Yb³⁺ DC layer [47]. Based on the photoluminescence spectrum of the DC material measured in Figure 11, the energy transfer mechanism within the material is inferred as shown in Figure 12. Through calculating, the DC layer synthesized by W. Romero-Romo’s team may achieve a theoretical quantum efficiency enhancement to 144%. Additionally, their analysis using the Inokuti–Hirayama or Dexter models for energy transfer mechanisms indicated that non-radiative Pr³⁺ → Yb³⁺ energy transfer process arisen from Pr³⁺: ³P₀ and ¹D₂ levels are most likely dominated by electric dipole–dipole and quadrupole–quadrupole interactions, respectively.

In addition, Wang et al.’s [43] team innovatively followed preparation processes like Figure 13, fabricating a chlorophyll film as a down-conversion layer and then enhancing its stability by incorporating nano SiO₂ into the film-forming solution. Figure 14 shows the spectral analysis of Chl film. It can be seen from Figure 14a that adding Chl film with SiO₂ increased certain absorption of short-wavelength light (300–500 nm) for spectral modification without evident impacting on the absorption and utilization of long-wavelength light by PV modules and exhibited a strong emission in long-wavelength light (650–850 nm). Hence, the DC chlorophyll film synthesized by Wang’s team demonstrated a certain degree of down-conversion capability. From the improvement of power generation efficiency illustrated in Figure 15 and Figure 16, it can be seen that after encapsulated the DC chlorophyll film, efficiency of crystalline silicon increased by 10%. In addition, experimental results demonstrated that the efficiency of the encapsulated crystalline silicon solar cell only decreased by 1% after 30 days of outdoor exposure, which exhibited excellent stability. Undoubtedly, their experiment has opened a new avenue for the integration of inorganic and organic materials in solar applications.

3.3. Luminescent Down-Shifting (LDS) Technology

Similar to DC technology, which mainly aims to enhance the photovoltaic response of crystalline silicon solar cells to high-energy photons, LDS technology theoretically converts incoming high-energy photons into equivalent lower-energy photons near the bandgap width. The mechanism of LDS is simpler compared to up-conversion and down-conversion processes: An ion absorbs one high-energy photon, pumped to a high excited state, then undergoes non-radiative relaxation to a lower excited state, and subsequently emits a lower-energy photon, known as Stokes shift. However, the absorption spectrum of one single ion is limited, which restricts the full utilization of the incident solar spectrum. Therefore, energy transfer mechanisms between ions can be employed to broaden the spectral bands. For instance, by co-doping one rare-earth ion with another, as illustrated in the diagram: a rare-earth ion Re³⁺ absorbs a high-energy photon, pumped to the second excited state, and then transfers part of its energy to ion 1 via fluorescence resonance energy transfer (FRET), causing the ground-state electron in ion 1 pumped to the first excited state. Finally, ion 1 undergoes radiative transition, emitting a lower-energy photon. The FRET mechanism is shown in Figure 17.

Table 3. Recent developments of down-shift technology in solar cells.

Ref.	Materials	Solar Cell Type	Excitation Peak (nm)	Emission Peak (nm)	Emission Band (nm)	Quantum Yield Efficiency
[55]	Cs3Bi2Br9	Si	380	490	450–510	-
[56]	SrGa4O7: Cr3+, Yb3+	-	435	760	650–1100	31.4
[57]	MAPbBr3+ PMMA hybrid film	Si	350	525	500–550	85
[58]	TeO2–ZnO–Na2O: Ce3+-Yb3+/Tb3+- Yb3+/Pr3+-Nd3+-Yb3+	Si	430 355 488	520/1010 550/1010 648/980	-	-
[59]	YAG:Ce	c-Si	250 340 460	550	500–700	92
[60]	Eu-Zn+ PMMA	c-Si	350	610	600–625	63
[61]	YAG:Ce3++ EVA	c-Si	340	550	500–700	-
[62]	ZnO	Si	345	525	450–625	41.6
[63]	ZnO+ PMMA	Si	335	500	450–600	-
[64]	PDMS+ perovskite quantum dots	Si	365	530	500–560	97

illustrates recent applications of LDS technology in photovoltaic cells. The excitation band of LDS materials typically ranges from ultraviolet to blue light (250–488 nm), while their emission band spans from green to near-infrared light (490–1010 nm). Consequently, through LDS technology, a significant number of short-wavelength photons, which crystalline silicon cells cannot efficiently respond to, can be converted to wavelengths where the cells exhibit higher photovoltaic response. Moreover, unlike DC and DC technologies, most LDS materials demonstrate high quantum efficiency (greater than 60%), thus ensuring relatively stable conversion effect. In recent years, some teams have made significant strides in the study of LDS materials, achieving nearly 100% quantum yield efficiency in LDS material.

In recent years, LDS technology has made significant advancements in crystalline silicon solar cells. Meng et al. [64] modified PDMS composite films by removing the Cd_xZn_1−xSe_yS_1−y/ZnS quantum dots template and combined it with perovskite quantum dots to synthesize a novel LDS layer which not only provides anti-reflective properties but also enables down shifting of light frequencies. The team applied this synthesized LDS layer to amorphous silicon solar cells, then observed an obvious improvement in Isc, which is illustrated in Figure 18, leading to an absolute value of 1.05% increment in PCE from 8.95% to 10% for the amorphous silicon solar cells.

Moreover, Xu et al. [57] fabricated MAPbBr₃/PMMA hybrid film as an LDS layer, making a PCE enhancement of about 3% on Si heterojunction solar cell (SHJ) and 0.7% on passivated emitter and rear cell (PERC). It is delightful that the LDS layer has a PLQY of over 85% (excited by 350 nm) in an air atmosphere for more than 30 days, which is illustrated in Figure 19, demonstrated the material’s high efficiency and stability. Figure 20 shows the photoluminescence spectra excited by 350 nm, indicating that MAPbBr₃/PMMA hybrid film has a strong ability to convert ultraviolet light into green light. Hence, the experiment conducted by Xu et al. provides a novel perspective for peers’ research on DS materials.

4. Primary Concerns and Challenges

Over the past two decades, extensive experimentation has substantiated the feasibility of spectral modification technologies in photovoltaic cells. However, significant challenges persist in this field: low absorption efficiency of spectral modification layers, suboptimal quantum efficiency during spectral modification processes, and hurdles in commercialization.

Low absorption efficiency: Regardless of whether it is UC, DC, or LDS materials, low absorption remains a common issue across these three technologies. We have identified some representative applications of three technologies on solar cells in recent years and displayed the absorption rates of each spectral modification layer in Figure 21. In the graph, (a–c) [24,47,65] represent the transmittance of spectral modification layers prepared using UC, DC, and LDS technologies, respectively. However, the absorption rates and peak widths for the required bands: long-wave, medium-short-wave, and medium-short-wave are suboptimal. Therefore, methods to enhance absorption rates and broaden the absorption spectra of spectral modification materials need to be explored.

Suboptimal quantum efficiency: the Formula (1) for quantum conversion efficiency is as follows [66]:

$${\eta }_{QY}={\displaystyle \frac{n_{emission}}{n_{absorbed}}}$$

(1)

where ${\eta }_{QY}$ represents the quantum yield efficiency of UC/DC/LDS materials, $n_{absorbed}$ represents the number of absorbed photons and $n_{emission}$ is the number of emitted photons. In the domains of UC/DC/DS technologies, the primary issue exposed is the suboptimal value of $n_{emission}$. There are two primary factors that contribute to this issue. Firstly, the generation efficiency of charge carriers is too low. Secondly, many charge carriers are dissipated before they are transported to external circuit.

Hurdles in commercialization: At present, nearly all spectral modification technologies remain confined to the laboratory stage due to the following reasons. Firstly, diverse synthesis techniques, selection of matrix materials and doped elements have their own merits. Thus, an optimal commercial solution has yet to be established. Secondly, the characterization of the spectral modification layer is excessively complex, hindering its commercial application. Thirdly, the cost is prohibitively high, and with the current state of research, the improvement in photovoltaic efficiency is insufficient to offset the increased cost.

5. Recommendations and Future Outlook

To address these issues mentioned in Section 4, we undertake the following analysis, offering recommendations for peers in spectral modification field.

Concerning the issue of low absorption efficiency, one approach is to coat the spectral modification layer with anti-reflection materials [67]. For instance, Liu et al. [68] coated silicon solar cells with a layer of SiO₂, containing plasmonic indium-tin-oxide nanoparticles (ITO-NPs) of 3 wt%, making the cell’s efficiency increase by 17.90%.

To address the low generation efficiency of charge carriers, experimental approaches include adjusting rare-earth elements, improving synthesis methods, tuning synthesis temperatures, and modifying the crystal structure of the spectral modification layer to improve its energy transfer mechanisms. However, this issue remains particularly challenging across all spectral modification technologies because the energy transfer mechanisms for UC, DC, and LDS technologies are not conclusively or uniformly understood; they are inferred based on reverse analysis of photoluminescence spectra, leading to inherent inaccuracies. Therefore, it is difficult to make targeted adjustments to the synthesis process of the spectral modification layer.

As for the losses of carriers after generation, Liu et al. [69]. provide us with a promising approach: Their team firstly constructed a Ti-mesh supported TiO₂ nanowire arrays (NWAs)/up-conversion luminescence Er³⁺(2.0 mol%)-Yb³⁺(1.0 mol%) co-doped TiO₂ nanoparticles (UC-EY-TiO₂ NPs) composite structured photoanodes for fully flexible dye sensitized solar cells. Then they coated the TiO₂ NWAs/UC-EY-TiO₂ NPs with an Nb₂O₅ thin layer to further suppress electron recombination losses. Combined with Pt/ITO-PEN counter electrode, the new assembly exhibited an enhanced photovoltaic conversion efficiency of 8.10%, a 68% improvement compared to TiO+ NWAs/undoped TiO₂ NPs based DSSCs (4.82%). Moreover, for the purpose of investigating the inhibitory effect of Nb₂O₅ on the electron recombination losses, Liu’s team performed the electrochemical impedance spectra (EIS) on the new assembly, of which the Nyquist curves and the corresponding Bode curves are shown in Figure 22. After being coated with Nb₂O₅ layer, the diameter of the right semicircle (Figure 22a) increased and the peak frequency (Figure 22b) of the middle frequency region (0.1–100 Hz) decreased, which demonstrated that electron recombination resistance of the photoanode/electrolyte interface was promoted. In addition, Liu’s team quantitatively acquired the electron lifetime according to the Formula (2) as follows [70]:

$${\tau }_e={\displaystyle \frac{1}{\omega }}={\displaystyle \frac{1}{2\pi f_p}}$$

(2)

where ${\tau }_e$ is lifetime of electron and $f_p$ represents the the corresponding frequency of the largest phase peak in the region (0.1–100 Hz), as shown in Figure 22b. The calculation results show that when coated with Nb₂O₅, lifetime of electron increased from 51.46 ms to 91.58 ms, indicating that the electron recombination losses could be effectively restrained.

To achieve commercialization of spectral modification technologies, it is essential to first refine experimental protocols, enhance the absorption efficiency of solar spectra and efficiency of quantum yield, and stabilize the crystal structures of UC, DC, and LDS layer materials.

In recent years, the rise of perovskite solar cells has further propelled the development of the spectral modification technologies, particularly UC technology. However, due to the poor stability and insufficient long-term reliability of perovskite cells, it is essential to consider enhancing their stability when applying spectral modification technologies. In addition to reducing the ultraviolet light incident on the cells through DC or LDS technologies, appropriate surface treatments [28] or encapsulation of the spectral modification layer with a stable shell [55] can also be employed to further extend the lifespan and reinforce the stability of perovskite cells.

6. Conclusions

Based on existing research in the photovoltaic field, this paper reviews the applications of UC, DC, and LDS technologies in solar cells over the past five years. It analyzes these technologies from various aspects including the selection of matrix materials and doped elements, types of cells, energy transfer mechanisms, optical performance, and practical application effects. Additionally, we have offered our insights and suggestions on several major issues related to the UC, DC and LDS technologies: low absorption efficiency of spectral modification layers, suboptimal quantum efficiency during modification processes, and hurdles in commercialization. Conclusions are summarized as follows:

1.

Up-conversion technology is commonly used in solar cells like thin-film solar cells, perovskite cells (PSCs), dye-sensitized cells (DSSCs) and quantum-dot-sensitized solar cells (QDSSCs), with the method of doping Er³⁺ in NaYF₄ matrix as an up-conversion layer being quite popular. Its emission band typically ranges from 523 nm to 669 nm, corresponding to green to red light, while the excitation band ranges from 800 nm to 1550 nm, corresponding to near-infrared light.

2.

Down-conversion technology is commonly applied in silicon-based solar cells, typically doping Yb³⁺ in its matrix material. Its emission band typically ranges from 520 nm to 1031 nm, corresponding to green to near-infrared light, while the excitation band ranges from 250 nm to 488 nm, corresponding to ultraviolet-to-blue light. The quantum yield efficiency of down-conversion is usually lower than 100%.

3.

Luminescent down-shifting technology is typically used in silicon-based cells. Similar to down-conversion, it usually involves doping Yb³⁺ in the luminescent down-shifting matrix material. Its emission band generally spans from 490 nm to 1010 nm, corresponding to green to near-infrared light, while the excitation band ranges from 250 nm to 488 nm, corresponding to ultraviolet-to-blue light. The quantum yield efficiency of luminescent down-shifting is usually higher than 60%.

4.

Based on the current state of research, there are three common issues prevalent in the fields of spectral modification: low absorption efficiency of spectral modification layers, suboptimal quantum efficiency during spectral modification processes, and hurdles in commercialization.

5.

In response to the main issues present in spectral modification, this review proposes the following solutions: Concerning the issue of low absorption efficiency, one method that can be employed is coating the spectral modification layer with anti-reflection materials to enhance the absorption rate of incident photons. To address the suboptimal quantum efficiency during spectral modification process, we can take experimental approaches including adjusting rare-earth elements, improving synthesis methods, tuning synthesis temperatures, and modifying the crystal structure of the spectral modification layer to improve its energy transfer mechanisms. Moreover, to ensure more charge carriers are transported to external circuits, it is an excellent method for coating a layer to suppress electron recombination losses. To achieve commercialization of spectral modification technologies, it is essential to first refine experimental protocols, enhance the absorption efficiency of solar spectra and efficiency of quantum yield, and stabilize the crystal structures of UC, DC, and DS layer materials.