After this paper was published ASAP on August 24, 2020, changes were made to the Acknowledgments. The corrected version was published August 25, 2020. Perovskite solar cells (PSCs) have sky-rocketed in recent years because of their high-performance, now at a certified 25.2%, as well as readily available fabrication methods and material components. They are now considered one of the main contenders to make a lasting impact in photovoltaics either as a stand-alone solution, for example on flexible substrates, or in multijunction architectures with established technologies such as silicon or CIGS. This tremendous potential has led to an unusually rapid development of perovskite materials by academic and industrial groups. This includes a coordinated effort for certified measurement protocols.(1) For example, efficiency certifications of PSCs now consider the need for a stabilized power output because of the idiosyncratic ion migration in PSCs. Similarly, there is an effort toward the normalization of stability measurements to eventually form an industrial protocol.(2) It is only logical to move toward standardization because it ensures comparability among research groups and industry. More binding measurement protocols also build credibility toward potential investors and are therefore a necessary precursor for industrialization. Another layer of normalizing the measurement of solar cells was recently introduced by several journals(3−5) in the form of a checklist for reporting performance parameters of solar cell. Here it is typical, among a series of requirements, to report on the discrepancy between the short-circuit current density (J_sc) from external quantum efficiency (EQE) (J_sc,EQE) and current density-voltage (JV) (J_sc,JV) measurements. The EQE is a basic measurement for solar cells. It measures the conversion of an incident photon to an electron by the photovoltaic device and is in general a function of the photon wavelength. The inset of the EQE(λ) can provide the spectral regions which contribute to the photocurrent generation of the solar cell (see Figure 1). Its spectral shape reveals, for example, how efficiently an absorber material transforms light into current and may reveal loss mechanisms within the device stack. Figure 1. JV curve and EQE spectrum of a typical perovskite solar cell. The definitions of J_sc,JV and J_sc,EQE are shown in the figure. Experimentally, the EQE can be measured with a light source, often a Xe lamp, with a monochromator. The light source is used to illuminate the sample through an appropriate shadow mask. The device is kept at short-circuit condition, and the current through the device is measured with a source meter for each wavelength. The raw data of the current needs to be normalized to the number of incident photons, conducted by calibration of the setup with a reference solar cell with a known spectral response for each wavelength. The EQE is the spectral response of the solar cell; therefore, it can be used to calculate the J_sc of the cell under illumination.(1)where S(λ) is photons per second. Typically, the J_sc is measured from the JV curve, which depicts the short-circuit current density as a function of applied voltage. The JV curve allows for extraction of the open-circuit voltage (V_oc), the fill factor (FF), and thus the power conversion efficiency (PCE). Hence, the J_sc,EQE provides a complementary method to compare with the J_sc,JV. Figure 1 shows the definition of J_sc,JV and J_sc,EQE. Optimally the J_sc,JV, as measured by the solar simulator, should match the J_sc,EQE. Comparing the two values is an important and easy check for the reliability of JV and EQE measurements, avoiding measurement errors due to wrong calibration. In recent years it emerged that for PSCs the reported J_sc,EQE is relatively consistently lower than the J_sc,JV. This is illustrated in Table 1, which shows certified PSCs and their respective J_sc,EQE and J_sc,JV. It is noteworthy that none of the certificates contain absolute values for the J_sc,EQE, although the spectral shape of the EQE is frequently reported. Thus, the reported J_sc,EQE values are not certified and were calculated in the respective laboratories. Often, the J_sc,EQE is close to, albeit less than, the J_sc,JV, within 10–20%, thus showing that the measurement from the JV curve is reasonably accurate. Except for one value, Table 1 shows that the J_sc,EQE is less than the J_sc,JV. J_sc,EQE shows the highest reported J_sc value from the integrated EQE spectrum. Frequently, this is not measured directly on the certified device. The certificates never report absolute values for J_sc,EQE. Δ = J_sc,JV – J_sc,EQE shows, with the exception of one data point, that there is a consistently lower J_sc,EQE. We exemplify this discrepancy further with our own PSCs in Figure 1. The EQE measurements were taken without white light bias. The J_sc,JV of this cell is 22.3 mA cm^–2, where the J_sc,EQE is 20.9 mA cm^–2 with a relative difference of 1.4 mA cm^–2, consistent with Table 1. The calibration for absolute EQE values is notoriously tricky. A typical EQE setup uses a Xe lamp in combination with a monochromator as light source. Therefore, the light intensity at a given wavelength is low compared to the standard 1 sun illumination used in the solar simulator. As a result, the charge density in the device is different, which can increase the mismatch of the J_sc,JV and the J_sc,EQE of the solar cell. This is pronounced in solar cells that show a nonlinear behavior of the photocurrent with light intensity. Such devices are usually recombination-limited.(16) Moreover, if such variations were random, we would expect to observe, at least as equally often, a slightly increased J_sc,EQE compared to J_sc,JV. However, the systematically lower J_sc,EQE speaks to the opposite, hinting that there is an underlying reason. The mechanism is hard to precisely pinpoint at this stage and goes beyond the scope of this Viewpoint. Nevertheless, we posit that it may be connected to the following observations/reasons or a combination thereof:

• The ion migration within the perovskite materials changes the interface dynamics during the relatively long measurement time required to record each wavelength for the J_sc,EQE compared to the relatively quick J_sc,JV.(17,18)
• Long-term sample degradation may be induced by longer measurement times.(19) A possible option in this case could be to reduce the measurement time and thus the effects of possible degradation of devices during measurements.
• The measurement conditions at the solar simulator are not the same as the conditions during the EQE measurement. It can be assumed that the JV sweep in the solar simulator corresponded to a prebiasing which is not the case in the EQE characterization. For example when a forward bias is applied on the perovskite solar cell, the electric field in the perovskite cell is influenced. The external bias leads to ion movement inside the perovskite layer and at the interfaces; as a result, a partial build-in potential is created, which influences the observed EQE.(20)
• A preconditioning measurement of the cell will affect the absolute scale and the shape of the EQE. This can cause an additional discrepancy between the J_sc values. Moreover, often a white light bias is applied during the EQE measurement. This results in a photovoltage which can also distort the J_sc,EQE measurement.(21) The light bias usually affects the preconditioning response time, which indicates how long it takes for the cell to stabilize under light bias and how stable the current signal is over the entire EQE measurement. Because some perovskite solar cells have a slow change in their performance after the initial preconditioning, it is uncertain how long we should wait for the cell to stabilize. On the other hand, a long waiting time can induce cell degradation.
• One more reason that could have a possible contribution to the difference in the J_sc values is the strong frequency dependence of the EQE for perovskite solar cells. Therefore, an EQE measurement which involves a chopper or perturbation frequency can result a disparity in these values. This might be due to the long response time scale of the perovskite.(22)

The ion migration within the perovskite materials changes the interface dynamics during the relatively long measurement time required to record each wavelength for the J_sc,EQE compared to the relatively quick J_sc,JV.(17,18) Long-term sample degradation may be induced by longer measurement times.(19) A possible option in this case could be to reduce the measurement time and thus the effects of possible degradation of devices during measurements. The measurement conditions at the solar simulator are not the same as the conditions during the EQE measurement. It can be assumed that the JV sweep in the solar simulator corresponded to a prebiasing which is not the case in the EQE characterization. For example when a forward bias is applied on the perovskite solar cell, the electric field in the perovskite cell is influenced. The external bias leads to ion movement inside the perovskite layer and at the interfaces; as a result, a partial build-in potential is created, which influences the observed EQE.(20) A preconditioning measurement of the cell will affect the absolute scale and the shape of the EQE. This can cause an additional discrepancy between the J_sc values. Moreover, often a white light bias is applied during the EQE measurement. This results in a photovoltage which can also distort the J_sc,EQE measurement.(21) The light bias usually affects the preconditioning response time, which indicates how long it takes for the cell to stabilize under light bias and how stable the current signal is over the entire EQE measurement. Because some perovskite solar cells have a slow change in their performance after the initial preconditioning, it is uncertain how long we should wait for the cell to stabilize. On the other hand, a long waiting time can induce cell degradation. One more reason that could have a possible contribution to the difference in the J_sc values is the strong frequency dependence of the EQE for perovskite solar cells. Therefore, an EQE measurement which involves a chopper or perturbation frequency can result a disparity in these values. This might be due to the long response time scale of the perovskite.(22) However, specialized studies would be required to elucidate and confirm the mechanism in more detail. The focus of this Viewpoint is to bring this issue to the attention of the community. Although this appears to be less relevant or reported for other PV technologies, it is noticeable for PSCs and thus should be studied in the future. Here, we propose that the requirement for matching J_sc,EQE and J_sc,JV should be discussed and analyzed within the scientific community and eventually revised if needed. Currently, there is little reflection about this topic despite the empirical evidence that has been observed by many research groups thus far. Therefore, in the interest of further standardization, it is important to also look more extensively at EQE measurements and to ensure that checklists for solar cells take this into account for PSCs. Views expressed in this Viewpoint are those of the authors and not necessarily the views of the ACS. The authors declare no competing financial interest. We acknowledge Stav Rahmany who fabricated the solar cells and measured their JV and EQE. We also acknowledge Dr. Weifei Fu for assistance with the data extraction in Table 1. This article references 22 other publications.

中文翻译：

---

钙钛矿太阳能电池中的电流密度不匹配

在2020年8月24日尽快发布本文之后，对声明进行了更改。校正后的版本已于2020年8月25日发布。钙钛矿太阳能电池（PSC）近年来由于其高性能（已认证的25.2％）以及易于使用的制造方法和材料组件而飞速发展。现在，它们被认为是对光伏发电产生持久影响的主要竞争者之一，无论是作为独立解决方案（例如在柔性基板上）还是在具有成熟技术（例如硅或CIGS）的多结架构中。这种巨大的潜力已导致学术界和工业界异常快速地发展钙钛矿材料。这包括对经过认证的测量协议的协调工作。（1）例如，PSC的效率认证现在考虑到稳定的功率输出的需要，因为PSC中的异质离子迁移。类似地，人们也在努力进行稳定性测量的标准化，以最终形成一个工业协议。（2）朝着标准化迈进是合乎逻辑的，因为它可以确保研究小组与行业之间的可比性。更具约束力的衡量协议还建立了对潜在投资者的信誉，因此是工业化的必要先兆。最近，几本期刊（3-5）以检查清单的形式介绍了另一项标准化太阳能电池测量的层，用于报告太阳能电池的性能参数。在这一系列要求中，典型的是报告短路电流密度之间的差异（Ĵ _SC）从外部量子效率（EQE）（Ĵ _SC，EQE）和电流密度-电压（JV）（Ĵ _SC，JV）的测量。EQE是太阳能电池的基本度量。它测量光伏器件将入射光子转换为电子的过程，通常是光子波长的函数。EQE（λ）的插图可以提供有助于太阳能电池光电流产生的光谱区域（见图1）。它的光谱形状揭示了例如吸收材料将光转换为电流的效率，并且可能揭示了器件堆叠内部的损耗机制。图1.典型钙钛矿太阳能电池的合资曲线和EQE光谱。J _sc，JV的定义和Ĵ _SC，EQE示于该图中。实验上，可以使用带有单色仪的光源（通常是Xe灯）测量EQE。光源用于通过适当的荫罩照亮样品。该设备保持短路状态，并且通过源表针对每个波长测量通过该设备的电流。电流的原始数据需要归一化为入射光子数，这是通过使用参考太阳能电池对每个波长具有已知光谱响应的装置进行校准来进行的。EQE是太阳能电池的光谱响应。因此，它可用于计算光照下的电池的J _sc。（1）其中S（λ）是每秒的光子。通常，J _sc是从JV曲线测量的，该曲线描述了短路电流密度与施加电压的关系。JV曲线允许提取开路电压（V _oc），填充系数（FF）以及功率转换效率（PCE）。因此，J _sc，EQE提供了一种与J _sc，JV比较的补充方法。图1显示了J _sc，JV和J _{sc，EQE的定义}。最佳地_，由太阳模拟器测量的J _sc，JV应该与J _sc，EQE匹配。比较这两个值对于检查JV和EQE测量的可靠性是一项重要且简便的检查，避免了由于错误的校准而导致的测量错误。近年来发现，对于PSC，报告的J _sc，EQE相对一致地低于J _sc，JV。在表1中对此进行了说明，该表显示了经过认证的PSC及其各自的J _sc，EQE和J _sc，JV。值得注意的是，尽管经常报告EQE的频谱形状，但所有证书都不包含J _sc，EQE的绝对值。因此，报告的J _sc，EQE值未经认证，是在各自的实验室中计算得出的。通常，尽管J _sc，EQE小于J _sc，JV，但仍在10％至20％之内，因此表明从JV曲线测得的结果是相当准确的。除一个值外，表1显示J _sc，EQE小于J _sc，JV。J _sc，EQE显示了从综合EQE光谱中报告的最高J _sc值。通常，不能在经过认证的设备上直接进行测量。证书从不报告J _sc，EQE的绝对值。Δ= J _sc，JV –J _sc，EQE表示，除一个数据点外，存在一个始终较低的J _sc，EQE。我们在图1中用我们自己的PSC进一步说明了这种差异。EQE测量是在没有白光偏差的情况下进行的。该电池的J _sc，JV为22.3 mA cm ^–2，其中J _sc，EQE为20.9 mA cm ^–2，相对差为1.4 mA cm ^–2，与表1一致。众所周知，绝对EQE值的校准非常棘手。典型的EQE设置使用Xe灯和单色仪作为光源。因此，与太阳模拟器中使用的标准1太阳照明相比，给定波长的光强度较低。结果，器件中的电荷密度不同，这会增加太阳能电池的J _sc，JV和J _sc，EQE的失配。在太阳能电池中，这表现出光电流随光强度呈非线性行为，这是很明显的。此类设备通常受到重组限制。（16）此外，如果此类变化是随机的，我们希望至少同样频繁地观察到J _sc，EQE略有增加。与J _sc，JV相比。但是，系统地较低的J _sc，EQE则相反，暗示存在根本原因。该机制在此阶段很难精确定位，并且超出了此观点的范围。尽管如此，我们认为它可能与以下观察/原因或其组合有关：

• 与相对较快的J _sc，JV相比，钙钛矿材料内的离子迁移在记录J _sc，EQE的每个波长所需的相对较长的测量时间内改变了界面动力学。（17,18）
• 较长的测量时间可能导致长期样品降解。（19）在这种情况下，一种可能的选择是减少测量时间，从而减少测量期间设备可能降解的影响。
• 太阳模拟器的测量条件与EQE测量期间的条件不同。可以假定，太阳模拟器中的合资扫描对应于预偏置，而EQE表征中的情况并非如此。例如，当在钙钛矿太阳能电池上施加正向偏压时，钙钛矿电池中的电场受到影响。外部偏压导致钙钛矿层内部和界面处的离子运动；结果，产生了部分内建电位，这会影响观察到的EQE。（20）
• 单元的预处理测量将影响EQE的绝对规模和形状。这可能会导致J _sc值之间的其他差异。此外，在EQE测量期间通常会施加白光偏压。这导致产生光电压，该光电压也会使J _sc，EQE失真（21）光偏置通常会影响预处理响应时间，这表明电池在光偏置下稳定需要多长时间，以及整个EQE测量中电流信号的稳定程度。由于某些钙钛矿型太阳能电池在初始预处理后性能会缓慢变化，因此不确定我们应该等待多长时间才能稳定下来。另一方面，长时间的等待会导致细胞降解。
• 可能对J _sc值差异产生影响的另一个原因是钙钛矿太阳能电池EQE的强烈频率依赖性。因此，涉及到斩波或扰动频率的EQE测量会导致这些值的差异。这可能是由于钙钛矿的响应时间长所致。（22）

与相对较快的J _sc，JV相比，钙钛矿材料内的离子迁移在记录J _sc，EQE的每个波长所需的相对较长的测量时间内改变了界面动力学。。（17,18）较长的测量时间可能导致长期样品降解。（19）在这种情况下，可能的选择是减少测量时间，从而减少测量期间设备可能退化的影响。太阳模拟器的测量条件与EQE测量期间的条件不同。可以假定，太阳模拟器中的合资扫描对应于预偏置，而在EQE表征中情况并非如此。例如，当在钙钛矿太阳能电池上施加正向偏压时，钙钛矿电池中的电场受到影响。外部偏压导致钙钛矿层内部和界面处的离子运动；结果，产生了部分内建电位，这会影响观察到的EQE。（20）单元的预处理测量将影响EQE的绝对规模和形状。这可能会导致J _sc值。此外，在EQE测量期间通常会施加白光偏压。这导致产生的光电压也会使J _sc，EQE测量失真。（21）光偏置通常会影响预处理响应时间，这表明电池在光偏置下稳定需要多长时间以及电流信号的稳定性如何。在整个EQE测量中。由于某些钙钛矿型太阳能电池在初始预处理后性能会缓慢变化，因此不确定我们应该等待多长时间才能稳定下来。另一方面，长时间的等待会导致细胞降解。一个可能对J _sc的差异可能有贡献的原因值是钙钛矿型太阳能电池EQE的强烈频率依赖性。因此，涉及到斩波或扰动频率的EQE测量会导致这些值的差异。这可能是由于钙钛矿的响应时间长所致。（22）但是，将需要专门研究来阐明和更详细地确定该机理。该观点的重点是使这个问题引起社区的注意。尽管这似乎与其他光伏技术的相关性或报道较少，但对PSC而言却很明显，因此应在以后进行研究。在这里，我们建议匹配J _sc，EQE和J _{sc，JV的要求}应该在科学界内部进行讨论和分析，并在需要时进行最终修订。当前，尽管到目前为止许多研究小组已经观察到了经验证据，但对该主题的思考很少。因此，出于进一步标准化的目的，重要的是还要更广泛地研究EQE测量，并确保太阳能电池的检查清单将PSC纳入考虑范围。本观点中表达的观点是作者的观点，不一定是ACS的观点。作者宣称没有竞争性的经济利益。我们感谢制造太阳能电池并测量其合资企业和EQE的Stav Rahmany。我们还感谢傅维非博士在表1中进行数据提取的帮助。本文引用了其他22种出版物。