What is a solar cell？

A solar cell converts sunlight to electricity via PN junctions: commercial silicon ones, 180μm thick, use doped layers to generate 0.5V/cell, achieving 20% efficiency, powering gadgets or grids by photon - electron conversion.

Basic Working Principles

By 2023, the global cumulative installed capacity of solar power had exceeded 1.2 TW (1.2 billion kilowatts), accounting for 38% of total renewable energy generation, equivalent to reducing carbon dioxide emissions by 7 billion tons annually. Taking mainstream monocrystalline silicon cells as an example, laboratory efficiency has jumped from 6% at Bell Labs in 1954 to 26.81% in 2023 (Kaneka, Japan), while mass production efficiency has generally reached 23%-24%.

A single 182mm×182mm monocrystalline silicon cell's actual power generation capability is limited by "quantum efficiency": only photons with wavelengths of 400-1,100nm can be effectively absorbed, accounting for about 50% of the total solar spectrum; and on each square centimeter of area, approximately 10¹⁷ photons strike per second, but ultimately only about 25% are converted into free charge carriers (electron-hole pairs), with the rest lost due to reflection, thermal loss, or recombination.

Light Absorption

A single 182mm×182mm monocrystalline silicon solar cell placed in sunlight receives approximately 1.2×10²¹ photons per second – but only about 2.5×10¹⁶ of them can actually excite electron-hole pairs. The remaining 99.999975% of photons either pass through, are reflected, or collide and "annihilate each other." The main reason behind this is the material's bandgap (Eg), which acts like an "energy threshold": only photons with energy ≥ Eg can be absorbed by the semiconductor.

If we use Germanium (Eg=0.67eV), although it can absorb more long-wavelength light (1,100-1,850nm, accounting for 15% of the energy), the excess energy of high-energy photons (e.g., below 500nm) is wasted as heat (the portion of single-photon energy exceeding Eg turns into heat). Using Gallium Nitride (Eg=3.4eV) is too selective, only absorbing ultraviolet light below 365nm (accounting for 0.5% of energy), with most sunlight passing directly through.

1. What exactly is the Bandgap?

The photon energy (E=hc/λ, where h is Planck's constant, c is the speed of light, λ is the wavelength) must be ≥ Eg for an electron to jump from the valence band to the conduction band, generating free charge carriers. For example, silicon's Eg=1.1eV corresponds to the longest absorbable wavelength of 1100nm (red to near-infrared); after parameter tuning, perovskite materials have Eg=1.5eV, able to absorb up to 827nm (orange to near-infrared).

2. Why is Silicon's Bandgap the Most "Fortuitous"?

The energy distribution of the solar spectrum is "high in the middle, low at the ends": photons in the 400-1,100nm range account for 58% of the total energy, 1100-2500nm account for 38%, and ultraviolet (400nm) accounts for only 4%. Silicon's Eg=1.1eV just covers the band with the largest energy share. It neither wastes high-energy photon heat like Germanium nor misses too many mid-to-low energy photons like wide-bandgap materials. Lab data shows: using a material with Eg=1.1eV, the theoretical spectral matching efficiency (absorbable photon energy / total incident light energy) can reach 70%; if Eg=0.7eV (Germanium), the spectral matching efficiency drops to 65%.

3. How does the Bandgap Affect Actual Power Generation?

· Quantum Efficiency (QE): Refers to the proportion of absorbed photons that can generate electron-hole pairs. Silicon's QE > 90% at 900nm, but drops sharply to 0% above 1100nm; Germanium's QE is still 50% at 1500nm, but is <10% below 400nm. Overall, silicon's integrated external quantum efficiency (total generated current / number of incident photons) is about 28%, while Germanium's is about 25%.

· Thermal Loss: The portion of photon energy exceeding Eg turns entirely into heat. For example, a 3eV photon (ultraviolet) striking silicon uses only 1.1eV for electricity generation, the remaining 1.9eV becomes heat, causing the cell temperature to rise – which in turn lowers Voc (open-circuit voltage), creating a vicious cycle. Experiments show: for every 1% increase in the proportion of high-energy photons (λ < 400nm) in the incident light, the cell operating temperature increases by 0.2°C, and efficiency decreases by 0.09%.

· Material Thickness: The higher the bandgap, the thinner the absorption layer required. Gallium Nitride Eg=3.4eV, absorption coefficient >10⁶ cm⁻¹, requires only 0.1μm thickness to absorb light sufficiently; Silicon Eg=1.1eV, absorption coefficient ≈10⁵ cm⁻¹, needs to be 180μm thick. Thicker silicon wafers can absorb more long-wavelength light but increase resistive losses (current travels a longer distance in the silicon) and cost (silicon material accounts for 30% of cell cost). The industry's use of 180-200μm thin silicon wafers is a balance between thickness and efficiency.

4. How is the Bandgap "Adjusted" Industrially?

· Doping: Pure silicon Eg=1.1eV. After doping with Boron (P-type) or Phosphorus (N-type), Eg changes <0.01eV, almost unaffected for light absorption, but it forms the PN junction.

· Alloying: For example, Silicon-Germanium alloy (SiGe), where Eg decreases with increasing Germanium content: Eg=1.0eV at 10% Ge, Eg=0.8eV at 30% Ge. This material is used on the backside of Passivated Emitter and Rear Contact (PERC) cells to absorb more long-wavelength light, improving infrared response.

· New Materials: Perovskites can precisely control the bandgap between 1.4-1.7eV by adjusting composition (e.g., Formamidinium Lead Iodide (FAPbI₃) Eg=1.5eV, Methylammonium Lead Iodide (MAPbI₃) Eg=1.6eV), matching a wider solar spectrum. In 2023, Oxford PV's perovskite-silicon tandem cell used an upper perovskite layer (Eg=1.7eV) to absorb short wavelengths and a lower silicon layer (Eg=1.1eV) to absorb long wavelengths, achieving a total efficiency breakthrough of 32% – nearly 10% higher than the monocrystalline silicon limit.

5. The Bandgap is Not "The Wider the Better," Nor "The Narrower the Better"

Theoretically, the Shockley-Queisser (SQ) limit calculation shows: the theoretical efficiency of a single-junction cell is highest (33.7% → 35.5%) when Eg=1.34eV. But in reality, silicon (1.1eV) mass production efficiency has reached 23.5%, approaching its theoretical limit (29.4%); while Gallium Arsenide (GaAs) cells with Eg=1.34eV have a mass production efficiency of 47% (multi-junction), but the cost is 100 times that of silicon, limiting their use to satellites.

Current Output

A monocrystalline silicon solar cell labeled "conversion efficiency 23.5%" actually has its entire power generation capability hidden within its I-V curve when placed under standard test conditions (AM1.5G, 1000W/m², 25°C).

Taking a mainstream industry product as an example, its short-circuit current (Isc) is about 39 mA/cm², open-circuit voltage (Voc) is 0.67V, and fill factor (FF) is 80.5%. Multiplying these three gives the maximum power (Pmax = Isc × Voc × FF), which is then divided by the light intensity and cell area to obtain that 23.5% efficiency.

These three parameters are not isolated: low Isc may indicate poor light absorption by the material, low Voc might suggest short minority carrier lifetime, and low FF often hides resistive losses. PV engineers constantly adjust processes based on these three curves because every 0.1% increase in FF can earn an extra 500 hours of equivalent generation per year for a module.

Short-Circuit Current (Isc): The Maximum Instantaneous "Water Flow" the Cell Can Output

Short-circuit current is the current output when the cell's positive and negative terminals are directly connected, measured in mA/cm² (milliamps per square centimeter).

Baseline Value: Mass-produced monocrystalline silicon cell Isc is about 38-40 mA/cm² (corresponding to a 182mm×182mm module, single-cell Isc ≈ 10.5A); TOPCon cells, due to higher minority carrier lifetime, can reach 40.5 mA/cm²; perovskite-silicon tandem cells can even exceed 42 mA/cm² (absorbing more short-wavelength light).

Influencing Factors:

· Light Intensity: Isc is proportional to incident light intensity. For example, on a cloudy day with only 500 W/m² irradiance, Isc drops from 10.5A to 5.25A.

· Material Light Absorption: Anti-reflection coating on silicon wafers (e.g., SiNx) can reduce reflectivity from 30% to 1%, allowing more photons to excite electrons, directly increasing Isc by 5%.

· Temperature: For every 1°C temperature increase, Isc increases slightly by 0.05% (e.g., from 25°C to 75°C, Isc increases from 10.5A to 10.5 × (1 + 0.0005 × 50) = 10.526A), but this is offset by the decrease in Voc.

Industry Pain Point: Early cells struggled to achieve high Isc due to significant surface reflection and rear surface recombination losses. Now, using Laser SE (Selective Emitter) technology, which involves heavy doping locally in the emitter to reduce carrier recombination, Isc can be increased by an additional 0.5 mA/cm².

Open-Circuit Voltage (Voc): The Cell's "Voltage Holding" Capability When Idle

Open-circuit voltage is the potential difference between the positive and negative terminals when the cell is not connected to any load, measured in V (Volts).

Baseline Value: Monocrystalline silicon cell Voc is about 0.65-0.68V; PERC cells, through rear surface passivation, can achieve Voc up to 0.68V; Heterojunction (HJT) cells, due to fewer interface defects, can push Voc above 0.7V.

Influencing Factors:

· Material Bandgap: Silicon Eg=1.1eV, theoretical Voc limit ≈ 0.79V (using formula Voc = (kT/q) ln (Isc / (q × J₀)), where J₀ is the reverse saturation current density). The actual value is 0.1V below the limit, entirely due to recombination losses dragging it down.

· Temperature: For every 1°C temperature increase, Voc decreases by about 2mV (-0.3%/°C). For example, a cell with Voc=0.67V at 25°C will have Voc = 0.67 - 2mV × 50 = 0.57V at 75°C in summer, a direct 15% decrease.

· Minority Carrier Lifetime: The longer the minority carrier lifetime, the less recombination, and the higher the Voc. With high-quality silicon wafers having a minority carrier lifetime of 100μs, Voc ≈ 0.67V; if the lifetime drops to 10μs, Voc plummets below 0.6V.

Industry Breakthrough: Using an aluminum oxide (Al₂O₃) passivation layer to "clothe" the P-type silicon surface with an insulating layer reduces the surface state density (from 10¹² cm⁻² to 10¹⁰ cm⁻²), increasing Voc by 30mV, which directly adds 0.7% to the efficiency.

Fill Factor (FF): The Conversion Efficiency that Turns "Potential" into "Actual Profit"

The fill factor is the ratio of the cell's maximum power point (Pmax = I × V) to Isc × Voc, expressed as a percentage.

Baseline Value: Conventional Aluminum Back Surface Field (BSF) cells have FF ≈ 75%; PERC cells achieve 78%-79%; TOPCon and HJT can push to 81%-82%. Don't underestimate this 3% difference: for a 182mm module, increasing FF from 78% to 81% can earn an extra 200 kWh per year (based on 1000 hours of sunshine).

Influencing Factors:

· Series Resistance (Rs): The resistance encountered by current flowing through the silicon wafer, grid lines, and busbars. For every 0.1 Ω·cm² increase in Rs, FF decreases by 1%-2%. Now, using low-temperature silver paste (resistivity 2.5×10⁻⁶ Ω·cm) and ultra-fine grid lines (25μm wide), Rs can be pressed below 0.5 Ω·cm².

· Shunt Resistance (Rsh): Internal leakage current paths in the cell. If Rsh drops from 1000 Ω·cm² to 100 Ω·cm², FF will drop by 5%. Using high-quality silicon material (fewer impurities) and laser SE technology, Rsh can be increased to over 5000 Ω·cm².

· Contact Resistance: The resistance at the contact points between the metal grid lines and the silicon wafer. Optimizing the "firing process" (increasing temperature from 700°C to 800°C) can reduce contact resistance from 50 mΩ·cm² to 10 mΩ·cm², increasing FF by another 1%.

Industry Practice: A leading manufacturer used "busbarless technology" (eliminating 3-5 main busbars, keeping only fine grid lines) to reduce busbar shading (from 10% to 8%) and contact resistance, increasing FF from 79% to 81.5%, while also reducing cost per watt by 0.02 yuan.

Loss Tracing

The theoretical efficiency limit for a monocrystalline silicon solar cell is 29.4% (Shockley-Queisser limit), but the industry's mass production efficiency in 2023 had just reached 23.5% – a gap of 5.9 percentage points, equivalent to generating 59 watts less for every 100 watts of theoretical generation.

This 29% gap is not a single issue but the combined result of seven major "leakage vulnerabilities" such as light reflection, carrier recombination, resistive losses, and spectral mismatch.

1. Light Reflection: The "Surface Colander" Leaks 8% First

When sunlight first hits the cell surface, some photons are directly reflected. The natural reflectivity of monocrystalline silicon is as high as 30% (untreated), meaning 300W out of 1000 W/m² irradiance is directly reflected back into the atmosphere.

· Data Quantification: The industry uses anti-reflection coatings (SiNx) to solve this, pressing reflectivity below 1%. For example, a leading manufacturer's PERC cell, with the ARC refractive index controlled at 2.0 (silicon's refractive index is 3.5) and film thickness at 80nm (exactly 1/4 wavelength of incident light), reduces reflectivity from 30% to 0.8%, directly recovering a loss of 29.6 W/m² (based on 1000 W/m²).

· Hidden Detail: The ARC thickness isn't "the thinner the better." A deviation of ±5nm in thickness can cause reflectivity to rebound above 0.5%; film contamination (e.g., dust particles) can cause local reflectivity to soar to 5%, leading to uneven cell efficiency.

2. Carrier Recombination: The "Internal Borer" Steals 5%

Electron-hole pairs excited by photons recombine (electrons and holes recombine) before being collected by the electrodes, which is the most hidden loss. Recombination is of two types:

· Surface Recombination: The edges and rear surface of the silicon wafer have many defects, acting like "traps" that capture carriers. The surface recombination velocity (Sr) of an unpassivated silicon wafer is as high as 1×10⁵ cm/s, meaning 100,000 carriers recombine per second per square centimeter. Covering it with an aluminum oxide passivation layer (4nm thick) reduces Sr to 1×10³ cm/s, recovering 2% efficiency (corresponding to a 15mV increase in Voc).

· Bulk Recombination: Impurities (like oxygen, carbon) and dislocations in the silicon material form recombination centers within the bulk. High-quality monocrystalline silicon has a bulk recombination lifetime (τbulk) of 100μs, while multicrystalline silicon, due to more grain boundaries, has τbulk of only 1μs – this is the main reason why multicrystalline silicon efficiency is 2-3% lower than monocrystalline.

3. Resistive Losses: "Wires Too Thin" Drag Away 3%

When current flows internally through the cell, the resistance encountered turns energy into heat. These losses are of three types:

· Series Resistance (Rs): Includes the silicon wafer bulk resistance, grid line resistance, and busbar resistance. Monocrystalline silicon bulk resistivity is about 1 Ω·cm; the bulk resistance contribution for a 180μm thick wafer is 0.3 Ω·cm²; silver grid line resistance is the major contributor – traditional grid lines are 50μm wide, 3μm thick, each with a resistance of 0.1 Ω·cm²; 3 main busbars + 100 fine grid lines total Rs ≈ 0.8 Ω·cm². Using ultra-fine grid lines (25μm wide) + low-temperature silver paste, Rs can be pressed below 0.5 Ω·cm², increasing FF by 1.5%.

· Shunt Resistance (Rsh): Internal leakage current paths, such as material defects or edge leakage. If Rsh drops from 1000 Ω·cm² to 100 Ω·cm², FF will drop by 5%. High-purity silicon material (impurity concentration < 1×10¹⁵ cm⁻³) can increase Rsh to over 5000 Ω·cm².

4. Spectral Mismatch: "Picky Eating" Wastes 5%

Silicon's Eg=1.1eV can only absorb photons in the 400-1,100nm range, while infrared light above 1100nm in the solar spectrum (accounting for 38% of energy) is almost entirely transmitted through.

· Data Comparison: If a "perfect cell" could absorb all wavelengths of light, the theoretical efficiency would soar to 44% (upper limit of the SQ limit), but limited by Eg, silicon can only utilize 58% of the spectral energy.

· Improvement Attempt: Tandem cells are a solution. For example, perovskite (Eg=1.7eV) absorbs short wavelengths (400-600nm), silicon absorbs long wavelengths (600-1,100nm); complementing each other increases spectral utilization from 58% to 75%, with total efficiency breaking 32%.

5. Other "Small Leaks": Combined Drag Away 8%

· Auger Recombination: When a high-energy electron and hole recombine, they transfer energy to another electron instead of generating current. Auger recombination accounts for 3% in highly doped regions (like the emitter); using a low-doped emitter (doping concentration reduced from 1×10¹⁹ cm⁻³ to 1×10¹⁸ cm⁻³) can reduce this by 1%.

· Metal Contact Loss: If a Schottky barrier forms at the contact points between grid lines and the silicon wafer, it hinders current extraction. Using a "tunnel oxide layer + doped polysilicon" contact structure reduces the barrier height from 0.8eV to 0.3eV, cutting contact resistance by 50%.

Main Materials and Types

As of 2023, in the global solar cell market, silicon-based cells absolutely dominate with over 90% share – of which monocrystalline silicon accounts for 73%, and multicrystalline silicon for 17%; the remaining 10% are thin-film cells (mainly Cadmium Telluride CdTe, Copper Indium Gallium Selenide CIGS) and emerging perovskite cells.

The mass production conversion efficiency of monocrystalline silicon modules has jumped from 18% in 2010 to 23.5% in 2023 (LONGi's HPBC technology reached 26.5%), while the cost per watt-hour has dropped from 1.8 yuan/W to below 0.8 yuan/W.

Silicon-Based Cells

In 2023, global new PV installations were 110 million kilowatts, of which 92% used silicon-based cells – monocrystalline silicon accounted for 73%, multicrystalline silicon for 19%. This is backed by hard data: monocrystalline silicon module mass production efficiency increased from 19.8% in 2015 to 23.8% in 2023 (LONGi's HPBC technology lab efficiency 26.81%), and the unit cost of generation dropped from 0.5 yuan/kWh to below 0.25 yuan/kWh.

Silicon material's "compatibility" is unmatched: it can adapt to all scenarios from small residential systems (3-5kW) to hundred-megawatt-scale ground-mounted power plants, with modules generally promising a 25-year lifespan (first-year degradation ≤2%, thereafter ≤0.45% per year), more stable than thin-film cells' 15-20 years. But silicon-based also has worries: high-purity silicon material accounts for 40% of cell cost, and wafer thickness has been reduced to 150 micrometers (a human hair is about 70 micrometers in diameter).

1. Monocrystalline Silicon: How Did It Go from "Expensive Item" to "King of Cost-Effectiveness"?

The rise of monocrystalline silicon is entirely driven by the dual wheels of "cost reduction + efficiency improvement." Early on (before 2010), the Czochralski (CZ) method was used to pull monocrystalline silicon ingots, with electricity consumption as high as 120 kWh/kg, wafer thickness 200 micrometers, costing 1.5 times that of multicrystalline silicon. But after 2015, two technological breakthroughs changed the situation:

· Ingot Pulling Process Upgrade: Continuous Czochralski (CCZ) technology increased pulling speed from 1.2 mm/min to 2.5 mm/min, reducing electricity consumption to 80 kWh/kg; replacing graphite with carbon-carbon composite materials for the hot zone reduced single-furnace pulling time from 24 hours to 18 hours. Now mainstream monocrystalline silicon ingots have a diameter of 210mm (weight about 800kg), which can be sliced into 2800 pieces of 182mm×182mm wafers (each 150μm thick), with the silicon material cost per wafer dropping from 18 yuan in 2015 to 3.5 yuan in 2023.

· Cell Technology Iteration: After the popularization of PERC (Passivated Emitter and Rear Contact) technology, monocrystalline silicon cell efficiency jumped from 19% to 22%; TOPCon (Tunnel Oxide Passivated Contact) pushed efficiency to 24.5% (data from Jinko Solar's mass production line); HJT (Heterojunction), although costly, has a bifaciality factor over 90% (PERC about 85%), providing 5-7% more generation gain in scenes with high reflectivity like deserts and snow-covered ground.

Now, leading companies (LONGi, Jinko, Trina) have pressed the non-silicon cost per watt of their monocrystalline silicon modules down to 0.15 yuan: silver paste usage dropped from 150 mg/cell in 2018 to 80 mg/cell (using 0BB busbarless technology), target material (TCO layer) cost dropped from 0.05 yuan/W to 0.02 yuan/W due to domestic substitution.

2. Multicrystalline Silicon: A Cheap but "Insufficient" Transitional Choice

Multicrystalline silicon cells are produced by the casting method, melting silicon material and cooling it into an ingot (with many grain boundaries, defect density 10 times that of monocrystalline silicon). Its advantage lies only in being "cheap": wafer thickness is 200-230μm (30-50μm thicker than monocrystalline), silicon material loss is 10% less; casting electricity consumption is 60 kWh/kg (monocrystalline 80 kWh/kg), resulting in a module cost 0.1-0.15 yuan/W lower than monocrystalline silicon.

But the disadvantages are also obvious: mass production efficiency is stuck at 19%-21% (monocrystalline 23%+), temperature coefficient is -0.4%/°C (monocrystalline -0.35%/°C). In the same 40°C environment, multicrystalline silicon module power is 3-4% lower than monocrystalline. In 2018, multicrystalline silicon's share was over 50%, now it's only 19%, mainly flowing to cost-sensitive markets – such as rural distributed projects in India and Southeast Asia, where users installing 3-5kW systems care more about initial investment than long-term efficiency.

3. The "Invisible Battlefield" of Silicon-Based Cells: Detailed Games in the Supply Chain

80% of the cost and efficiency of silicon-based cells are determined by the upstream supply chain. For example:

· Silicon Material Purity: Solar-grade silicon material requires 9N purity (99.9999999%), but monocrystalline wafer factories now compete to buy 11N electronic-grade silicon material (purity 99.999999999%) – because fewer impurities increase minority carrier lifetime from 10 microseconds to 20 microseconds, potentially adding 0.3% to cell efficiency. In 2023, electronic-grade silicon material price was 1.8 times that of solar-grade, but leading companies are willing to pay.

· Crucible Lifespan: Quartz crucibles used for ingot pulling previously lasted only 300 hours; now, through coating technology (silicon nitride coating), the lifespan is extended to 500 hours, reducing the crucible cost allocated per ingot from 15 yuan to 9 yuan.

· Module Encapsulation: Monocrystalline silicon modules using POE encapsulant (water vapor transmission rate 0.1 g/m²/day) are more resistant to aging than those using EVA encapsulant (0.5 g/m²/day), increasing power retention after 25 years from 85% to 90%, but at a higher cost of 0.05 yuan/W.

4. The Future of Silicon-Based: How Long Can It Stay Hot?

Theoretically, the efficiency limit of monocrystalline silicon cells is 29.4% (Shockley-Queisser limit). Now TOPCon has reached 24.5%, HJT lab efficiency is 26.81%, leaving 4-5 percentage points of improvement potential. However, as perovskite tandem technology (perovskite/silicon tandem efficiency 33.7%) approaches, silicon-based's "ballast" status may be challenged in the next 10 years. But in the short term (before 2030), silicon-based will remain mainstream.

Thin-Film Cells

Out of the global 110 million kW of PV installations in 2023, thin-film cells accounted for about 10% (11 million kW), mainly supported by two technologies: Cadmium Telluride (CdTe) and Copper Indium Gallium Selenide (CIGS). Although the share is not high, it has "special skills" that silicon-based can't match: for the same 1000W module, CdTe generates 10% more electricity than monocrystalline silicon at 45°C high temperature (0.5 kWh more per day), and flexible CIGS modules can be attached to curved roofs or car bodies.

But thin-film also has hard flaws: CdTe modules contain 0.8 grams of cadmium per kWh generated (requiring strict encapsulation to prevent leakage), and CIGS, due to the scarcity of Indium and Gallium, still has a mass production cost 1.5 times that of silicon modules (0.8-1.2 yuan/W vs 0.5-0.6 yuan/W).

1. Cadmium Telluride (CdTe): The "Power Generation Whiz" in High-Temperature Regions

Currently, 95% of the world's CdTe modules are produced by First Solar. In 2023, this company shipped 12GW, equivalent to 8 out of every 10 thin-film modules being from them.

· Low temperature coefficient is the killer feature. Silicon module power decreases by 0.4% per C temperature increase; CdTe decreases by only 0.25%. In the Middle East (summer surface temperature 50°C +), CdTe modules actually generate power 2 hours more per day than monocrystalline silicon, 12% higher annual generation (based on a 1000W module, earning 300 kWh more per year).

· Good low-light response. At 5 AM dawn or 7 PM dusk, silicon module efficiency drops to 50% of peak; CdTe maintains 65%. In regions with many cloudy/rainy days (e.g., UK, Kyushu Japan), the number of effective generation days per year is 15 days more than silicon.

· Extremely low cost. CdTe module mass production cost is 0.2-0.25 USD/W (approx. 1.4-1.7 yuan/W), nearly half the price of silicon modules. The secret lies in the process: using "vapor transport deposition" technology to directly deposit cadmium telluride thin film on glass, avoiding complex steps like slicing and texturing required for silicon wafers; equipment investment is only 1/3 of silicon lines.

But it also has a fatal weakness: controversy over cadmium's toxicity. Although the risk of cadmium leakage is extremely low after module encapsulation (First Solar tests show that after breakage in a dry environment for 24 hours, leakage is <0.1 μg/m³, far below the safety threshold of 10 μg/m³), Europe still restricts its use in residential areas. Currently, 90% of CdTe projects are concentrated in Texas, USA (world's largest single CdTe plant 1.5GW) and Rajasthan, India (desert high-temperature scenarios).

2. Copper Indium Gallium Selenide (CIGS): "Efficiency Experiment" on Flexible Substrates

CIGS is the "technical faction" among thin films, its biggest feature being the ability to make flexible modules – Hanergy's flexible CIGS module launched in 2020 is only 2mm thick, can bend to a 30-degree arc, and be attached to warehouse roofs or RV bodies. But its problems are also obvious: expensive and difficult to mass-produce.

· High efficiency potential but difficult to realize. In the lab, small-area CIGS cell efficiency has reached 23.4% (HZB Institute, Germany), close to monocrystalline silicon levels; but mass-produced module efficiency is only 18%-19% (data from Hanergy's latest line), because it's hard to control coating uniformity over large areas – on a 1 square meter glass, a film thickness deviation over 5% causes a 1% efficiency drop.

· Material scarcity bottlenecks. Indium (In) and Gallium (Ga) in CIGS are rare metals: global indium reserves are only 19,000 tons (enough for 200GW CIGS modules), gallium is even scarcer (2023 global production only 400 tons).

· High cost but scenario-specific necessity. Hanergy's flexible CIGS modules sell for 1.2 yuan/W, 0.6 yuan more expensive than silicon, but when used in BIPV (Building-Integrated Photovoltaics), they can save on decoration costs – used directly as curtain walls, no need for additional mounting structures. German passive house projects love them because they are light (15 kg/㎡ vs silicon modules 25 kg/㎡), and can be installed on old building roofs with insufficient load-bearing capacity.

3. Other Thin Films: Each Has Its Own "Niche"

Besides CdTe and CIGS, two other thin-film cells operate in segmented markets:

· Amorphous Silicon (a-Si): The earliest commercialized thin-film technology, now basically only used in low-power devices like calculators and electronic watches. Low efficiency (mass production 8%-10%), but good weather resistance (operates from -40°C to 85°C); outdoor electronic devices using it can have a lifespan exceeding 10 years.

· Gallium Arsenide (GaAs): The "top stream" of space photovoltaics. NASA's Perseverance Mars rover uses triple-junction GaAs cells, achieving 28% efficiency under Mars sunlight (43% of Earth's), 10% higher than silicon cells. But the cost is outrageously high (500 USD/W).

Perovskite Solar Cells

In 2009, Japanese scientists first made solar cells using perovskite material, with an efficiency of only 3.8%; in 2023, in the lab of EPFL (Switzerland), single-junction perovskite cell efficiency had surged to 26.1% (certified by ISFH), just one step away from monocrystalline silicon's theoretical limit of 29.4%.

Perovskite/silicon tandem cell efficiency reached 33.7% in 2023, leaving silicon-based tandems (29.8%) behind. This "7-fold efficiency increase in ten years" is 20 times faster than silicon cells (10% increase in 30 years). But its Achilles' heel is equally glaring: current mass-produced modules decay 30% in efficiency after 1000 hours at 60°C, 85% humidity (silicon modules degrade ≤20% in 25 years); lead content is 0.5% (2 grams of lead per square meter), potentially contaminating soil if encapsulation fails.

1. Why Can Efficiency Increase So Fast? A "Combination Punch" of Materials and Processes

· Material inherently has "light absorption buff". Perovskite's chemical formula is ABX₃ (e.g., Methylammonium Lead Iodide ( MAPbI₃), with a crystal structure like "nano-scale LEGO," efficiently absorbing visible light – a 1-micrometer thick perovskite layer has over 90% absorption (silicon requires 180 micrometers). For the same power generation, perovskite modules can be made thinner (0.3mm vs silicon 0.5mm), 60% lighter (12 kg/㎡ vs 30 kg/㎡), directly halving the roof load pressure.

· Solution-based preparation saves cost. Silicon cells require ingot pulling, slicing, texturing, 10 process steps; perovskite uses "spin coating" or "slot-die coating," applying precursor solution onto a substrate and heating for crystallization, consuming only 1/10th the energy of silicon (0.1 kWh/W vs silicon 1 kWh/W). GCL's 100MW production line takes only 8 minutes from material input to output (silicon line takes 24 hours).

· Interface engineering boosts efficiency. In 2019, a team at Oxford University discovered that adding a layer of "self-assembled monolayer" (SAM) between the perovskite layer and the electron transport layer could reduce carrier recombination rate from 30% to 5%, directly increasing efficiency by 3%.

2. Where Does the Poor Stability Lie? Water and Heat are Two Major "Natural Enemies"

Perovskite's "short life" is mainly because the material itself fears water and heat:

· Decomposes upon contact with water. Perovskite's crystal structure hydrolyzes in water, generating lead iodide (yellow precipitate) and methylamine (toxic gas). The mainstream encapsulation solution now uses glass-glass laminate with encapsulant (similar to laminated glass), requiring water vapor transmission rate ≤10⁻⁶ g/m²/day (silicon modules are 10⁻⁴). But tests show that after encapsulant aging (after 5 years), water vapor penetration increases 5-fold, causing the perovskite layer to corrode from the edges, with efficiency decaying 0.2% per month (silicon decays 0.04% per year).

· High-temperature degradation. Perovskite's bandgap (1.5eV) shrinks with increasing temperature, and electron mobility decreases. At 60°C, module efficiency decreases 0.05% per hour; at 85°C, it decreases 0.1% per hour. Accelerated aging tests by the Chinese Academy of Sciences showed that unencapsulated perovskite cells retained only 50% of initial efficiency after 100 hours at 85°C, 85% humidity (silicon can last 1000 hours).

· Light-induced fatigue. Under continuous illumination, the perovskite layer generates "halogen vacancies," like "small pits" in the cell, hindering electron flow. Microquanta's modules degraded 15% after one year of outdoor exposure (silicon degrades 2%), mainly for this reason.

3. What Are the Bottlenecks in Mass Production? Yield and Lead Pollution are Roadblocks

The biggest hurdle for perovskite mass production now is not efficiency, but yield and environmental concerns:

· Low yield increases cost. GCL's 100MW production line currently has a yield just over 80% (silicon lines are 95%+). The problem lies in "large-area uniform film formation" – lab small areas (1cm²) can achieve perfect crystallization, but for large 1m×0.6m modules, the film thickness deviation between edges and center can reach 10%, causing local efficiency to be 2% lower. To improve yield, more precise coating equipment is needed (unit price increases from 5 million to 20 million yuan), pushing equipment investment per GW to 300 million yuan (silicon line is 150 million yuan).

· Lead pollution controversy is hard to resolve. Perovskite contains lead (Pb), about 2 grams per square meter module (silicon modules contain no heavy metals). Although leakage risk is low after encapsulation (IEC standard requires lead leakage <1 μg/L after 1000 hours immersion), public acceptance is low. The EU's "Batteries and Waste Batteries Regulation" requires lead content in PV modules to be ≤0.1% by 2030, which current technology cannot meet. Some teams are trying to replace lead with tin, but tin-based perovskite efficiency is only 60% of lead-based (15% vs 25%), and it oxidizes more easily.

4. Can It Replace Silicon in the Future? Short-term Look at Tandems, Long-term Compete on Scenarios

Perovskite is unlikely to completely replace silicon but will break through in tandem cells and specific scenarios:

· Perovskite/silicon tandem is the "optimal solution". Silicon cells absorb long-wavelength light (infrared), perovskite absorbs short-wavelength light (UV + visible), combining them can increase spectral utilization from 29% (monocrystalline silicon) to 45% (tandem). Current tandem module efficiency is 33.7%, possibly reaching 35% in the next 3 years, potentially reducing LCOE to 0.15 yuan/kWh (silicon is 0.25 yuan/kWh). LONGi, CATL are all laying out tandem production lines, expecting mass production by 2028.

· Opportunities in BIPV and flexible scenarios. Perovskite can be made into flexible modules (0.3mm thick), attached to curved roofs or cars, 60% lighter than silicon modules, reducing installation cost by 30%. Hanergy's flexible perovskite modules have been tested at Hangzhou Asian Games venues, generating 5% more efficiently than traditional BIPV.

· Extreme environments as "testing grounds". In space (high radiation), deserts (high temperature), polar regions (low temperature), perovskite's lightweight and low-light response might be more reliable than silicon. NASA has funded perovskite space cell R&D, aiming for use in lunar bases by 2030.

Applications

In 2023, global new PV installations reached 295GW (IRENA data), equivalent to adding a 100MW power plant every 38 minutes; China's cumulative installed capacity exceeded 5.3TW (CPIA), enough to support 15% of the annual electricity consumption of China's 1.4 billion people.

Taking North China as an example, a 1MW ground-mounted power plant generates 1.5 million kWh annually on average. At a benchmark electricity price of 0.38 yuan/kWh, annual income exceeds 570,000 yuan; environmentally: every kWh of PV electricity generated equates to saving 320 grams of standard coal and reducing 810 grams of CO₂ emissions.

Centralized PV Power Plants

In the Gobi desert in Northwest China, a 100MW centralized PV power plant is operating at full capacity – 320,000 monocrystalline silicon modules spread out like a blue sea, feeding 280,000 kWh of clean electricity into the grid every hour.

Such plants require a single investment of 300-400 million yuan (for 100MW scale), but after completion, they generate 150-180 million kWh annually on average. Based on Northwest China's desulfurized coal benchmark electricity price of 0.28-0.32 yuan/kWh, annual revenue is 42-57.6 million yuan, allowing cost recovery in 7-9 years. More directly: the electricity generated by a 1MW plant annually is enough for 1300 households for one year; over its 25-year lifespan, a 100MW plant can reduce 10 million tons of CO₂ emissions (equivalent to planting 550 million fir trees).

How is a Power Plant Actually Built? Steps from Site Selection to Grid Connection

Centralized plants cover hundreds of acres; the first step is choosing the right location. Areas with annual sunshine hours below 1300 are basically not considered (e.g., Sichuan Basin), prioritizing "sun-rich" regions like the Northwest (Xinjiang, Gansu) and North China (Inner Mongolia) – Hami, Xinjiang, has an average annual sunshine of 3300 hours; a 100MW plant there can generate 330 million kWh per year, nearly double that of East China.

After selecting the site, it's "building the structure": what modules to use? The mainstream now is bifacial TOPCon modules (conversion efficiency 22.5%), generating 5-8% more than traditional PERC (the backside can capture ground-reflected light). Choose fixed or tracking mounts? Fixed mounts cost less (15% of initial investment) but generate less; single-axis trackers cost 10% more (16.5% of investment) but can increase utilization hours by 20% (e.g., from 1600 to 1920 hours), making them more cost-effective overall.

How Strong is the Actual Generation Capacity? Real Data from Cells to System

New modules degrade 2% in the first year (industry standard), then 0.45% annually thereafter, retaining 84% efficiency after 25 years. A Gansu plant tracked data for 5 years: actual generation efficiency was 1.2% lower than rated (due to dust shading, temperature increase), but through monthly cleaning (cost 0.02 yuan/W), the loss was reduced to 0.8%.

From cell generation to grid injection, there are inverter losses 1.5%, transformer losses 0.8%, line losses 1.2%, totaling 3.5% loss. That means a 100MW plant theoretically can generate 165 million kWh (100MW × 1650 hours), but actually delivers 159 million kWh to the grid; the missing 6 million kWh is "leakage."

How is the Money Calculated? A Practical Breakdown of Cost, Revenue, and Risk

Modules account for 45% (100MW requires about 2000 tons of silicon material, 60% of module cost), mounting structures 15% (trackers are 5 million yuan more expensive than fixed), inverters 12% (string inverters are 20% more expensive than central inverters but more efficient), others (cables, construction, grid connection) 28%. After silicon material prices dropped in 2023, module cost fell from 1.8 yuan/W to 1.4 yuan/W, reducing the initial investment for a 100MW plant from 360 million yuan to 320 million yuan, directly saving 4 years of O&M costs ( 5 million yuan annually).

A Northwest plant receives a guaranteed purchase price of 0.3 yuan/kWh, with surplus electricity participating in market trading (premium 0.05-0.1 yuan/kWh). Detailed calculation: annual generation 165 million kWh, 70% sold at guaranteed price (0.3 yuan) earns 34.65 million yuan, 30% sold on market (0.35 yuan) earns 17.325 million yuan, total 51.975 million yuan. Deducting O&M cost ( 5 million yuan), loan interest ( 12 million yuan), annual net profit is 34.975 million yuan, with payback in 8.5 years.

Where are the risks? Mainly curtailment. In 2022, a Gansu plant faced a 15% curtailment rate due to grid absorption issues (24.75 million kWh less generation), directly losing 7 million yuan. Now new plants must be equipped with 10% energy storage (2 hours), which adds 30 million yuan investment but can suppress the curtailment rate to below 5%, making it more stable long-term.

How to Use Land Cost-Effectively? More Than Just "Occupying Land for Generation"

Gobi desert land rent is cheap ( 500 yuan/acre/year), but a 100MW plant occupies 3000 acres, costing 1.5 million yuan annually in rent. Is there a cheaper way? "Agrivoltaics" – increasing module spacing to 15 meters and planting shrubs underneath (sand fixation, can also sell carbon credits). A plant in Inner Mongolia did this, reducing land rent to 800,000 yuan/year (50% government subsidy), earning 300,000 yuan annually from carbon credits, netting a profit of 200,000 yuan.

Even better is "pastoral-photovoltaics": raising the module height to 4 meters and raising sheep underneath. A 100MW plant in Qinghai found that sheep grazing reduced shading (increasing generation by 3%), sheep manure fertilized the land reducing fertilizer need (saving 200,000 yuan/year), and mutton could be sold ( 500,000 yuan/year), tripling the comprehensive land revenue.

Land type also matters: it must be "unused land" or "general farmland" (cannot be basic farmland). In 2023, a company was required to dismantle a plant built on basic farmland, losing 200 million yuan, a profound lesson.

How Important is O&M? Details Determine Profit Margins

A 100MW plant employs 10 O&M personnel (monthly salary 6000 yuan), annual wage cost 720,000 yuan. But it's not about headcount; it's about technology: using drones for inspection (scanning 500 acres per hour, identifying hot spots) is 10 times faster than manual; cleaning robots (cleaning 10MW of modules daily) save 70% water compared to manual cleaning (manual cleaning 1MW requires 2 tons of water, robots use only 0.6 tons).

A real case: a Hebei plant lost 2% efficiency due to weed shading, costing 3.3 million yuan annually. After installing smart weeding machines (automatically identifying weeds, precise herbicide spraying) costing 200,000 yuan/year, it earned 2.8 million yuan more annually, netting 2.6 million yuan.

Then there's equipment aging: inverter lifespan is 10 years, replacement cost at end-of-life is 2 million yuan/MW (200 million yuan for 100MW). Now choosing "upgradable" inverters (with reserved interfaces) allows upgrading to more efficient models after 10 years for 50 million yuan, saving 150 million yuan compared to full replacement.

Distributed Photovoltaics

In the first year, they generated 9800 kWh; her household used 6000 kWh (saving 3480 yuan in electricity bills), and sold the remaining 3800 kWh to the grid (earning 1520 yuan), recouping the cost in two years. Now in the third year, she earns 5000 yuan passively annually, and after the 25-year contract ends, the modules can be sold to recyclers for 800 yuan/kW.

In 2023, China added 150GW of residential PV (40% of global total), with average annual earnings per household of 5000-7000 yuan; commercial & industrial is even more impressive: a factory in Guangdong installed a 1MW rooftop plant, saving 900,000 yuan annually on electricity bills, earning 300,000 yuan more than installing storage.

How Much Money Can You Really Make Installing PV at Home? A Clear Calculation

8-12kW systems are most common (suitable for 100-150㎡ roofs), costing 40,000-60,000 yuan to install (modules 60%, inverter 20%, mounting 10%, other 10%). Taking Jiaxing, Zhejiang as an example, average annual sunshine is 1100 hours; an 8kW system generates 8800 kWh/year (8kW × 1100 hours).

Assuming average monthly household consumption is 300 kWh (annual 3600 kWh) at a price of 0.58 yuan/kWh, self-consumption saves 3600 × 0.58 = 2088 yuan; surplus electricity is 8800 - 3600 = 5200 kWh, sold at the local desulfurized coal price of 0.4 yuan/kWh, earning 5200 × 0.4 = 2080 yuan; total annual revenue is 4168 yuan.

How to calculate payback? Initial cost 50,000 yuan, minus local subsidies (e.g., some areas in Zhejiang subsidize 0.1 yuan/kWh, annual subsidy 880 yuan), actual annual net revenue is 4168 + 880 = 5048 yuan. Payback period = 50,000 / 5048 ≈ 9.9 years (approx. 10 years). But module lifespan is 25 years, so the latter 15 years are pure profit: 5048 × 15 ≈ 75,700 yuan.

Do Businesses Save More on Electricity Bills Than They Earn? A Factory Owner's Real Ledger

A 1MW system (approx. 6000㎡) generates 1.2 million kWh/year (based on 1000 utilization hours). If the self-consumption rate is 70% (industrial electricity price 0.8 yuan/kWh), it directly saves 1.2M × 70% × 0.8 = 672,000 yuan; the surplus 30% (360,000 kWh) is fed into the grid at 0.4 yuan/kWh, earning 144,000 yuan; total annual revenue 816,000 yuan.

Faster payback: initial investment 4 million yuan (modules 45%, inverter 15%, mounting 10%, other 30%), annual O&M cost 50,000 yuan, loan interest 600,000 yuan (assuming 50% loan, 6% interest), annual net profit 816,000 - 50,000 - 600,000 = 166,000 yuan? Wait – businesses often use the "Energy Performance Contracting" model: a third party invests, the business only provides the roof, enjoying a discount on electricity bills (e.g., 10% off). A textile factory in Zhejiang signed a contract: the plant belongs to the investor, the factory pays 0.72 yuan/kWh for electricity (original 0.8 yuan), saving 1.2M × 70% × (0.8 - 0.72) = 67,200 yuan annually. Over a 20-year contract, savings total 1.344 million yuan, more cost-effective than self-investment.

Renting it to a PV company can also bring 3-5 yuan/㎡/year in rent (6000㎡ roof earns 18,000-30,000 yuan/year). A logistics park in Jiangsu did this, installing a 2MW plant on the roof, earning 240,000 yuan annually in rent, double the income from renting warehouse space ( 120,000 yuan/year).

Will Installation Cause Leaks? Real Risks of Roof Aging and Solutions

The reason is that mounting system installation includes waterproofing reinforcement – waterproof membranes cover screw holes, sealed with sealant, adding two extra layers of protection compared to standard roofing. A residential community in Zhejiang had no leaks for 5 years after PV installation; module shading reduced roof temperature by 8°C in summer (reducing AC use, saving owners money).

In 2022, a self-built rural house hired unqualified workers who damaged the original waterproofing during mounting installation. The roof leaked after 3 years, costing 20,000 yuan in repairs. The lesson: must hire qualified installers and specify "5-year waterproofing warranty" in the contract.

Another issue is roof load-bearing. Concrete roofs require a load capacity of 15 kg/㎡. An 8kW system weighs about 2000 kg (modules + mounting). Spread over a 60㎡ roof, that's 33 kg/㎡ – exceeds the limit? Don't panic – actual weight of modules + mounting is about 12 kg/㎡ (8kW uses ~20 panels? Wait, 8kW / 400W per panel ≈ 20 panels; each panel ~20kg? Let's recalc: 20 panels × 20kg = 400kg; mounting ~200kg; total ~600kg; 600kg / 60㎡ = 10 kg/㎡). Ordinary concrete roofs can handle this. For old houses with prefabricated slab roofs, reinforcement might be needed, costing 50 yuan/㎡ (60㎡ costs 3000 yuan), but it avoids collapse risk.

Who Manages Broken Modules? O&M Details Determine 25-Year Revenue

Solar panels aren't "install and forget"; O&M quality can cause a 30% difference in revenue. For example, dust shading: in Northwest China, dust accumulation causes a 5% drop in annual generation; regular cleaning (quarterly, cost 0.03 yuan/W) can recover 4%. A household plant in Zhejiang, where the owner cleaned the modules monthly with a pressure washer, generated 8% more annually than the neighbor (earning an extra 300 yuan/year).

Inverters are more delicate: string inverter lifespan is 10-15 years, requiring replacement. Replacing an inverter for an 8kW system costs 8000 yuan (16% of initial investment), but after replacement, efficiency increases from 97% to 98.5%, generating an extra 600 kWh/year (earning an extra 300 yuan/year), paying back in 5 years.

Then there's insurance: spending 500 yuan/year on "PV property insurance" covers panel breakage, lightning damage. In 2023, a household plant in Jiangsu was hit by hail, breaking 20 panels; the insurance company paid 12,000 yuan (at cost 600 yuan/panel); without insurance, it would have been out-of-pocket.

BIPV (Building-Integrated Photovoltaics)

The T3 terminal roof of Beijing Daxing International Airport, covering 120,000 square meters, has no traditional color steel sheets – replaced by CdTe thin-film BIPV modules, seamlessly integrated like dark blue glass curtain walls. This system generates 12 million kWh annually, enough for the terminal's public area lighting for 2 years, and saves 2 million yuan/year in traditional roof maintenance.

BIPV modules are 800 yuan/㎡ more expensive than conventional PV panels (former 1800 yuan/㎡, latter 1000 yuan/㎡), but over a 25-year lifespan, the traditional solution requires replacing color steel sheets 3 times (200 yuan/㎡ each time), total maintenance cost 1200 yuan/㎡, while BIPV saves 400 yuan/㎡.

How is BIPV Different from Conventional PV? Not "Stuck On," but "Grown Together"

Tests on a Shanghai office building's BIPV curtain wall (area 5000㎡) showed:

· Waterproofing: Module joints sealed with EPDM rubber strips showed no leakage during heavy rain (50mm/h), whereas the traditional color steel roof leaked in 3 spots simultaneously.

· Insulation: The temperature on the back of the modules was 10°C lower than behind color steel sheets (reducing summer indoor AC load by 15%);

· Load-bearing: Module + backsheet thickness is 12cm, more wind-resistant than color steel roofs (5cm + insulation layer) (can withstand Category 12 typhoons).

Is Building Such a Roof/Wall Expensive? Calculating the 25-Year Total Cost

BIPV's initial cost is indeed higher: for an industrial building, the traditional solution is "color steel roof (500 yuan/㎡) + retrofitted PV (1000 yuan/㎡)", total cost 1500 yuan/㎡; BIPV directly replaces both, costing 1800 yuan/㎡, 20% more expensive. But calculating the full lifecycle:

· Color steel sheet lifespan is 10 years, requiring 2 replacements over 25 years, maintenance cost 500 yuan/㎡ × 2 = 1000 yuan/㎡.

· BIPV lifespan is 25 years, no roof replacement needed, and it generates extra electricity revenue: 120,000 kWh / 1000㎡ × 25 years × 0.4 yuan/kWh = 120,000 yuan / 1000㎡.

Total: Traditional solution total cost 1500 + 1000 = 2500 yuan/㎡; BIPV total cost 1,800 - 120 = 1680 yuan/㎡ (subtracting extra generation revenue), actually saving 820 yuan/㎡. A factory in Zhejiang calculated this and switched from the planned "color steel + PV" to BIPV, saving 3 million yuan in full lifecycle costs.

How is the Actual Generation Performance? Stable Income Even on Cloudy Days and in Winter

BIPV generation is affected by building orientation and weather, but data is more stable than imagined. A BIPV roof on a Shenzhen hotel (area 8000㎡, module efficiency 16%):

· Summer (avg. 4 hours sunshine/day): Annual generation 1.92 million kWh (8000㎡ × 16% × 1000W/㎡ × 4h × 365 days ≈ 1.869 GWh ≈ 1,869,000 kWh, actual ≈ 1.92 million kWh after losses);

· Winter (avg. 2.5 hours sunshine/day): Annual generation 1.2 million kWh;

· Cloudy days (1 hour sunshine): Still generates 768,000 kWh/year (20% of total generation).

Comparison with retrofitted PV: For the same area, due to spacing left for maintenance access, the actual light-receiving area is 15% less, resulting in 288,000 kWh less annual generation (losing 115,200 yuan). BIPV doesn't have this problem; modules are tightly packed, utilizing even the corners.

What About Leaks, Aging, Breakdowns? Real Users' Pitfalls and Solutions

A Beijing mall installed BIPV curtain walls in 2018 with no leaks for 5 years – the secret is "three-layer waterproofing": weather-resistant sealant on module frames, aluminum alloy pressure strips on joints, plus waterproof membrane overlay. In contrast, a villa owner in 2020 hired a small factory that didn't treat the edge gaps; the roof leaked the next rainy season, costing 50,000 yuan in repairs (re-sealing + replacing damaged modules).

Aging data exists: BIPV module degradation rate is ≤2% first year, then 0.4% annually, retaining 88% efficiency after 25 years (national standard requires ≥80%). A BIPV system on a Suzhou office building operated for 6 years, with annual generation dropping only 3% (due to dust shading, restored after regular cleaning).

As for repairs? Choosing brand manufacturers offers better guarantees. A leading BIPV company offers a 12-year module warranty, 10-year inverter warranty, free replacement for non-human damage. A school in Zhejiang had 15 BIPV roof modules shattered by hail (2cm diameter) last year; the manufacturer replaced them within 48 hours, with no impact on generation.