Photon upconversion (UC), the process of combining two or more low-energy photons to generate a higher-energy excited state, is of interest for applications ranging from imaging to photodynamic therapy to solar energy conversion. Of the available UC mechanisms, molecular upconversion through sensitized triplet–triplet annihilation (TTA-UC) is particularly appealing because it can be achieved with low intensity, noncoherent irradiation.(1) TTA-UC was first observed in the 1960s(2) but largely remained dormant until a renaissance in the early 2000s. This revival was facilitated, in part, by the increasing availability of new transition-metal-containing sensitizers and the recognition that they could be used to increase both the efficiency and the apparent anti-Stokes shift of TTA-UC.(3) Subsequently, there was an explosion of new molecules, mediums, motifs, and applications, as well as new insight into the TTA-UC mechanism.(4−7) Unfortunately, there was, and still is, growing diversity in the terminology for TTA-UC, including concepts as fundamental as the upconversion quantum yield (Φ_UC). Most commonly, but not always, Φ_UC describes the emission quantum yield (# of UC photons emitted/# of photons absorbed by the sensitizer) or the normalized yield to scale the maximum theoretical efficiency from 50% to 100% (2 × emission quantum yield due to the two photon nature of TTA-UC). This discussion is further complicated by more recent applications of TTA-UC in integrated UC solar cells(8,9) and synthetic organic chemistry(10,11) where no upconverted emission is observed but there is still an upconversion quantum yield (# of UC excited states generated/# of photons absorbed by the sensitizer). Sometimes the author’s chosen definition of Φ_UC is immediately introduced. In other instances, the reader must diligently search the Supporting Information, or the definition is not given at all. This disparate terminology complicates the interpretation and direct comparison between TTA-UC experiments across different research groups.

In this Viewpoint, we propose an updated labeling scheme for describing quantum yield in TTA-UC so as to better facilitate communication and comparisons across applications, materials, and intervening events and to introduce more consistency in the scientific literature.

There is a large body of published work leading to this discussion, but because of space limitations, we regret that we cannot cite all of these works here. In this Viewpoint, we propose an updated labeling scheme for describing quantum yield in TTA-UC so as to better facilitate communication and comparisons across applications, materials, and intervening events and to introduce more consistency in the scientific literature. The generally accepted mechanism for TTA-UC using a molecular sensitizer (S) and annihilator (A) pair is presented in Figure 1. Figure 1. Energy level diagram, mechanism, and the states/species in TTA-UC facilitated by a molecular sensitizer (S) and annihilator (A) pair. Briefly, upon light absorption, the sensitizer (S) is excited (EX in Figure 1) into its singlet excited state (¹S*) followed by intersystem crossing (ISC) into the triplet excited-state manifold (³S*). Then, sensitizer-to-annihilator (A) triplet energy transfer (TET) results in an annihilator triplet excited-state species (³A*). When two ³A* collide, they can annihilate (TTA), with one relaxing back to its ground state and the other being promoted to its higher-energy singlet excited state (¹A*). Once in the upconverted, ¹A* state, this molecule can undergo nonradiative decay, radiative relaxation (EM), sensitize a different species (sen) either through trivial or nontrivial mechanisms, or nonproductively back energy-transfer to the sensitizer molecule (ET). Singlet oxygen sensitization (OS) can also serve as quenching processes for the ³S* and ³A* states but is most likely to occur for the latter because of its orders of magnitude longer excited-state lifetime. Additionally, steps like singlet sensitization,(12,13) cooperative/cascading sensitization,(14,15) and/or singlet exciton sinks(16) can readily be incorporated into the scheme but are omitted here for clarity. Crucial to quantum yield determinations is the number of photons and species (# y) involved in each step of TTA-UC. These species are highlighted in orange in Figure 1 and summarized in Table 1. Using the terms in Table 1, the quantum yield of any single or multistep process can be defined (Φ = # y/# z) with several of the most common/important shown in Table 2. For example, the ³S* to ¹A TET yield (Φ_TET) is the resulting number of annihilators in the triplet excited states (# ³A*) divided by the initial number of sensitizer triplet excited states (# ³S*). Although not discussed in detail here, these can also be related to the relaxation rate constants (i.e., Φ_TET = rate of TET divided by the sum of all relaxation rates from ³S*).(17) Experimentally we do not directly count the numbers of these species, but they can be indirectly determined using a combination of steady-state and time-resolved spectroscopic techniques.(18,19) Of the quantum yields listed in Table 2, it is worth highlighting the differences between Φ_UC, Φ_UCg, and Φ_UCs. Here we have defined Φ_UC as the measured TTA-UC emission quantum yield. That is, the number of UC photons observed (# hν_obs) divided by the number of photons absorbed (# ¹S*, assuming higher-energy excited states, ¹S_n*, undergo internal conversion prior to ISC). Because of the two-to-one photon nature of TTA-UC, the maximum theoretical value of Φ_UC is 0.5 or 50%. This measured value is sometimes multiplied by two or normalized to 1.0 or 100% (i.e., Φ_UC or Φ_UC′). Curiously, in the singlet fission literature (SF; a one photon in, two excited states out process) there are no instances of dividing the SF yield by two. Presumably, this is because in both TTA and SF, bigger numbers are “better”. Nonetheless, given the traditional and IUPAC(20) definition of quantum yield (Φ = # of A/# of B), we strongly encourage the community to discontinue this normalization practice (i.e., 2 × Φ_UC), or if they do, to label the result as the normalized upconversion emission efficiency (η_UC).

Given the traditional and IUPAC definition of quantum yield (Φ = # of A/# of B), we strongly encourage the community to discontinue this normalization practice (i.e., 2 × Φ_UC), or if they do, to label the result as the normalized upconversion emission efficiency (η_UC).

Given the traditional and IUPAC definition of quantum yield (Φ = # of A/# of B), we strongly encourage the community to discontinue this normalization practice (i.e., 2 × Φ_UC), or if they do, to label the result as the normalized upconversion emission efficiency (η_UC). In addition to Φ_UC we have defined the quantum yield of photons generated (Φ_UCg) with the difference being the output coupling yield (Φ_out).(21) The reason for this distinction is that Φ_UC is not necessarily intrinsic to the TTA-UC system but instead dependent on the sample architecture and the method of measurement (i.e., relative vs absolute). Output losses include waveguiding, scattering, inner filtering effects, etc. which are path length and sample holder dependent.(22) With the appropriate reference measurement, some of these losses are accounted for with an integrating sphere.(23) Alternatively, Φ_out can be estimated by dividing the emission quantum yield for direct excitation of A in the TTA-UC architecture, but without sensitizer, by the fluorescence quantum yield for dilute A in a comparable chemical environment. Arguably the most literal definition of upconversion quantum yield is Φ_UCs or the number of upconverted states generated (# ¹A*) per the number of photons absorbed (# ¹S*). Φ_UCs is the upconversion yield prior to ¹A* to S energy transfer, nonradiative decay, sensitization, and output coupling losses, and thus, regardless of the application (e.g., emission or sensitization) it can be compared across all TTA-UC systems. Also, it is worth noting that here we have defined Φ_TTA as the number of ¹A* generated per ³A* (Φ_TTA ≤ 0.5). TTA can also result in higher-energy triplet excited states (i.e., ³A* + ³A* = ³A_n* + ¹A), but this product yield is omitted from Φ_TTA because the net process is nonproductive (2 × ³A* to ³A*) and the resulting ³A* can return to the reaction pool to further undergo TTA.(19) TTA efficiency is sometimes partitioned into the efficiency of generating a contact pair and a spin-statistical factor for the fraction of contact pairs that generate a singlet (f).(24) Given the difficulty in measuring the latter, the debate regarding the rigorousness of the spin selection rules,(19,25−27) and reports suggesting interconversion between quintet and triplet correlated pairs,(28−30) we argue that Φ_TTA as defined here is a more general and encompassing definition. Φ_TTA can then be further partitioned into contact pair formation efficiency (Φ_cp) and a spin-fraction (f) components (Φ_TTA = Φ_cp·f). As mentioned above with UC, one could also report the normalized TTA efficiency (2 × Φ_TTA = η_TTA). The yields can then be related to each other using the equations presented in Table 3. They have been partitioned into equating Φ_UCs, Φ_UCg, etc. to (1) the individual steps or (2) experimentally measurable parameters. For the additive steps, we simply multiply by Φ_y and 1 – Φ_y for the productive and nonproductive processes, respectively. While conceptually convenient, it is far more likely to use experimentally measurable parameters (Φ_UC, Φ_f, Φ_ET, Φ_out, and Φ_LH) to calculate Φ_UCs, Φ_UCg, etc. as shown in the far-right column of Table 3. However, through a combination of both columns, important but difficult to directly measure parameters like Φ_TTA can be determined. One caveat with these calculations is that it only assumes the reaction pathways shown in Figure 1. Care must be taken to rule out and/or account for other quenching mechanism like those involving redox active species and solvent-dependent processes, for example.(31) Competitive quenching by oxygen is omitted for simplicity but could be included by multiplying by (1 – Φ_OS). For applications such as solar energy conversion, the efficiency metric of interest is based not on the number of absorbed photons (# ¹S*) but instead on the number of incident photons (# hν_in). To account for this, we can include Φ_LH (# ¹S*/# hν_in), also known as light-harvesting efficiency or LHE, as a multiplier to convert any of the above yields from the internal quantum efficiency (Φ_y) to the external quantum efficiency (Φ_yx) as exemplified for Φ_UCsenx in the bottom row of Table 3. The difference being the absorption cross section and the spectral overlap between the light source and the sensitizer absorption. The above discussion is focused on molecular sensitizers. However, recently, inorganic materials like quantum dots(32,33) and perovskite nanoparticles/thin films(34,35) have emerged as effective triplet sensitizers for TTA-UC. While a majority of the discussion holds true for inorganic sensitizers, there are a few distinct differences worth noting that are highlighted in Figure 2. Figure 2. Energy level diagram, mechanism, and the states/species in TTA-UC facilitated by an inorganic sensitizer (IS), triplet energy relay molecule (A_r), and molecular annihilator (A). Redundant steps from Figure 1 are deemphasized for clarity. The first change is that because of the small energetic difference and/or indistinguishable nature of the spin states, the inorganic sensitizers (ISs) lack distinct ¹IS* and ³IS* and that Φ_ISC is likely unnecessary.(36) As such, the excited state is simply labeled here as IS*. Second, because of their short lifetime, they typically employ a surface bound, molecular triplet energy transfer relay (A_r) to store a long-lived triplet state on the surface prior to TET to A in solution (Φ_TET). This adds an additional, sometimes reversible, TET to relay step, labeled here as TETr and Φ_TETr.(37) Back energy transfer can also occur from ¹A* to either the triplet sensitizer (Φ_ET) or the relay (Φ_ETr) with the former typically being more dominant because of the broad band absorbing nature of IS. The intense, broad absorption of IS also results in increased losses because of inner filtering (i.e., lower Φ_out) which in turn decreases Φ_UC. Nonetheless, the above expressions can readily be applied to IS systems by removing Φ_ISC and multiplying by Φ_TETr and 1 – Φ_ETr. Likewise, similar modeling can be envisioned for thermally activated delayed fluorescence (TADF) sensitizers where ISC becomes reversible (rISC).(38) As an aside, we have three additional comments, not directly related to quantum yield, regarding standardization in the TTA-UC literature. The first is in reporting of the “anti-Stokes shift.” By IUPAC standards, a Stokes shift is defined by the energy difference (in eV or cm^–1) between the peak maxima of absorption and emission of the same electronic transition.(20) Because TTA-UC absorption and emission processes are not derived from the same electronic transition, it is more appropriately described as an apparent anti-Stokes shift. Additionally, the magnitude of the apparent anti-Stokes shift should be reported as the difference between the lowest energy absorption peak of the sensitizer and emission peak of the annihilator,(39) and not the excitation wavelength. The former is based on the intrinsic properties of the molecules and is constant for a given sample. The latter is dictated by the excitation source, and even for the same TTA-UC sample, this value will change depending on the availability of excitation sources. Additionally, the latter can be deceptive in that you can make the shift artificially larger by using a higher-intensity laser and exciting into a low energy absorbance shoulder. Second is that the I_th value, the onset threshold intensity for the maximum efficiency regime,(40) should be reported in terms of both raw source intensity (mW/cm²) and excitation density (ex s^–1 cm^–2; photon flux × absorptance at λ_ex).(12) The former is useful because it allows for direct comparisons to solar flux available at that wavelength (vida infra) but is limited in that it is excitation wavelength-dependent. The latter (i.e., ex s^–1 cm^–2) enables a direct comparison between systems with different light-harvesting efficiencies, excitation wavelengths, and less than ideal overlap between laser sources and the absorption maximum of the sensitizer. Finally, TTA-UC activated at subsolar light flux does not simply mean exciting the sample with monochromatic light at intensities <100 mW/cm². For the AM1.5 solar spectrum, 100 mW/cm² is the integrated intensity over the entire solar spectrum ranging from 250 to 2500 nm. When comparing monochromatic excitation to solar intensities, it is pivotal to only integrate the AM1.5 solar spectrum over the sensitizer excitation range (i.e., the fwhm of the laser or the sensitizer absorption peak), which is typically <5 mW/cm². In line with this point, we strongly encourage authors to discontinue subjective and qualitative descriptions of incident power (e.g., extremely low power, ultralow intensity, etc.), because (1) “low power” means very different things in solar energy conversion and photodynamic therapy, for example, and (2) there are much more quantitative ways to define these terms. To summarize, here are the major take-home messages from this Viewpoint:

• We propose that the TTA-UC emission quantum yield (Φ_UC) is the number of UC photons observed per the number of photons absorbed (i.e., the traditional definition of light emission quantum yield).
• We strongly discourage the use of the 2 × Φ_UC normalization factor, but if used, it should be clearly indicated as the normalized TTA-UC emission efficiency (η_UC).
• Care should be taken in noting any output coupling losses because it is the difference between the observed (Φ_UC) and the intrinsic (Φ_UCg) TTA-UC emission quantum yield.
• Determining the yield of upconverted states (Φ_UCs), prior to relaxation, enables comparison across all TTA-UC systems and applications.
• The magnitude of the apparent anti-Stokes shift should be reported based on the difference between the emission peak of the annihilator and lowest energy absorption peak of the sensitizer, not the excitation wavelength.
• The I_th value should be reported in terms of both source intensity (mW/cm²) and excitation density (ex s^–1 cm^–2).
• When claiming TTA-UC occurs at subsolar flux, authors should only integrate the spectrum over the excitation range (<5 mW/cm²) and not the entire solar spectrum (100 mW/cm²).
• Authors should avoid qualitative descriptions of incident light power when more quantitative descriptions are readily available.

We propose that the TTA-UC emission quantum yield (Φ_UC) is the number of UC photons observed per the number of photons absorbed (i.e., the traditional definition of light emission quantum yield). We strongly discourage the use of the 2 × Φ_UC normalization factor, but if used, it should be clearly indicated as the normalized TTA-UC emission efficiency (η_UC). Care should be taken in noting any output coupling losses because it is the difference between the observed (Φ_UC) and the intrinsic (Φ_UCg) TTA-UC emission quantum yield. Determining the yield of upconverted states (Φ_UCs), prior to relaxation, enables comparison across all TTA-UC systems and applications. The magnitude of the apparent anti-Stokes shift should be reported based on the difference between the emission peak of the annihilator and lowest energy absorption peak of the sensitizer, not the excitation wavelength. The I_th value should be reported in terms of both source intensity (mW/cm²) and excitation density (ex s^–1 cm^–2). When claiming TTA-UC occurs at subsolar flux, authors should only integrate the spectrum over the excitation range (<5 mW/cm²) and not the entire solar spectrum (100 mW/cm²). Authors should avoid qualitative descriptions of incident light power when more quantitative descriptions are readily available. 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. Y.Z. and K.H. were supported by the National Science Foundation under Grant No. DMR-1752782. F.N.C. was supported by the U.S. Department of Energy, Office of Science, Office of Basic Energy Sciences, under Award No. DE-SC0011979. T.W.S. was supported by the Australian Research Council (Centre of Excellence in Exciton Science CE170100026). This article references 40 other publications.

中文翻译：

---

三重态-三重An灭光子上转换的量子产率

光子上转换（UC）是将两个或多个低能光子组合在一起以产生高能激发态的过程，它对于从成像到光动力疗法再到太阳能转换的各种应用都非常感兴趣。在可用的UC机制中，通过三重三重态-三重态三态trip灭（TTA-UC）进行分子上转换特别吸引人，因为它可以通过低强度，非相干照射来实现。（1）TTA-UC于1960年代首次观察到（2），但在很大程度上一直处于休眠状态，直到2000年代初复兴。这种复兴的实现在一定程度上是由于新型含过渡金属的敏化剂的可用性不断提高，并且认识到它们可用于提高TTA-UC的效率和明显的反斯托克斯位移。（3）随后，新分子，新介质爆炸了，_UC）。最常见的，但不总是，Φ _UC描述了发射量子产率（UC光子＃发射/光子＃吸收由敏化剂）或归一化产率从50％的最大理论效率扩展到100％（2×发射量子TTA-UC的两个光子性质决定了收率）。TTA-UC在集成UC太阳能电池中的最新应用（8,9）和合成有机化学（10,11）进一步使讨论变得更加复杂，在这些应用中，未观察到上转换发射，但仍存在上转换量子产率（UC数量）激发态的产生/敏化剂吸收的光子数量）有时候笔者的选择Φ的定义_UC立即介绍。在其他情况下，读者必须努力搜索支持信息，否则就根本没有给出定义。这种不同的术语使跨不同研究组的TTA-UC实验之间的解释和直接比较变得复杂。

在此观点下，我们提出了一种更新的标记方案，用于描述TTA-UC中的量子产率，以便更好地促进跨应用，材料和中间事件的交流和比较，并在科学文献中引入更多的一致性。

有大量已发表的著作引起了这一讨论，但是由于篇幅所限，我们感到遗憾的是我们不能在这里引用所有这些著作。 在此观点下，我们提出了一种更新的标记方案，用于描述TTA-UC中的量子产率，以便更好地促进跨应用，材料和中间事件的交流和比较，并在科学文献中引入更多的一致性。使用分子敏化剂（S）和an灭剂（A）对的TTA-UC的公认机理如图1所示。图1.分子敏化剂在TTA-UC中的能级图，机理和状态/物种（S）和歼灭者（A）对。简而言之，光吸收后，敏化剂（S）被激发（图1中的EX）进入其单重激发态（¹ S *），然后系统间交叉（ISC）进入三重激发态歧管（³S *）。然后，敏化剂到an灭者（A）的三重态能量转移（TET）导致了hil灭者的三重态激发态物质（³ A *）。当两个³ A *发生碰撞时，它们会an灭（TTA），其中一个放松回到其基态，另一个放松到其更高能量的单重态激发态（¹ A *）。一旦处于上转换的¹ A *状态，该分子便会经历非辐射衰变，辐射弛豫（EM），通过琐碎或非琐碎的机理敏化不同的物种（sen）或非生产性地将能量转移回敏化剂分子（ET） 。单线态氧敏化（OS）还可作为³ S *和^3的淬灭过程A *指出，但后者最有可能发生，因为它的激发态寿命更长，数量级更长。另外，诸如单重态敏化，（12，13）协同/级联敏化，（14，15）和/或单重态激子阱（16）之类的步骤可以容易地并入方案中，但是为了清楚起见在此省略。确定量子产率的关键是TTA-UC每个步骤中涉及的光子和物种数（＃y）。这些物质在图1中以橙色突出显示，并在表1中进行了概述。使用表1中的术语，可以定义任何单步或多步过程的量子产率（Φ=＃y /＃z），其中几种最常见/重要信息如表2所示。例如，³ S *至¹ A TET的产量（_ΦTET）是三重激发态（＃³ A *）的除能器数除以敏化剂三重激发态（＃³ S *）的初始数。尽管这里没有详细讨论，但这些也可能与松弛率常数相关（即_ΦTET = _TET的比率除以³ S *中所有松弛率的总和）。（17）实验上，我们不直接计算这些物种的数目，但它们可以通过稳态和时间分辨光谱技术的组合间接确定。（18,19）在表2列出的量子产率中，值得强调_ΦUC，Φ之间的差异。_UCG，Φ _UCS。在这里，我们已经定义Φ _UC为测量的TTA-UC发射量子产率。也就是说，UC的光子数观察到（＃ħ ν _OBS由光子数除以）吸收（＃¹ S *，假设高能量的激发态，¹小号_Ñ *之前，经历ISC内转换）。因为TTA-UC的两到一个光子性质，Φ的最大理论值_UC为0.5或50％。此测量值有时乘以二或标准化为1.0或100％（即，Φ _UC或Φ _UC''）。奇怪的是，在单线态裂变文学中（SF；一个光子进入，两个激发态出过程），没有出现将SF产量除以2的情况。大概是因为在TTA和SF中，更大的数字“更好”。尽管如此，考虑到传统的和IUPAC量子产率的（20）的定义（Φ=＃A的/ B的＃），我们强烈鼓励社会停止该标准化实践（即，2×Φ _UC），或者如果他们这样做，来标记结果作为归一化的向上变换发光效率（η _UC）。

鉴于量子产率的传统和IUPAC的定义（B的Φ=＃A的/＃），我们强烈鼓励社会停止这种标准化的做法（即2×Φ _UC），或者如果他们这样做，来标记结果作为归一化的向上变换发光效率（η _UC）。

鉴于量子产率的传统和IUPAC的定义（B的Φ=＃A的/＃），我们强烈鼓励社会停止这种标准化的做法（即2×Φ _UC），或者如果他们这样做，来标记结果作为归一化的向上变换发光效率（η _UC）。除了Φ _UC我们已经定义生成（Φ光子的量子产率_UCG，不同之处的输出耦合的产率（Φ）_出）。（21）这样做的原因的区别是，Φ _UCTTA-UC系统不一定是固有的，而是取决于样品结构和测量方法（即相对与绝对）。输出损失包括波导，散射，内部过滤效果等，这些路径长度和从属样本保持器（22）使用合适的参考测量，其中的一些损失占有积分球（23）可替代地，Φ_出可以通过在可比的化学环境中用TTA-UC结构直接激发A的发射量子产率除以没有敏化剂的A的荧光量子产率来估算。可以说，上转换量子产率的最直接的定义是_ΦUCs或生成的上转换态的数量（＃¹A *）每吸收的光子数（＃¹ S *）。Φ _UCS是之前的上变频产量^1个A *至S的能量转移，无辐射衰变，敏化，和输出耦合损耗，并且因此，不管应用程序（例如，发射或敏化）的，可以在所有TTA-UC系统相比。另外，值得注意的是，此处我们将_ΦTTA定义为每³ A *产生的¹ A *数量（_ΦTTA≤0.5）。TTA还可导致更高能量的三重态激发态（即³ A * + ³ A * = ³ A _n * + ¹A），但由于_TTA的净过程是非生产性的（2× ³ A *至³ A *），并且所得³ A *可以返回到反应池中进一步进行TTA ，因此从_ΦTTA中省略了该产物的收率。（19）TTA效率有时可分为生成接触对的效率和产生单重态（f）的接触对的比例的自旋统计因子。（24）鉴于难以测量后者，关于电池组的严格性的争论自旋选择规则，（19,25-27）和表明五重峰和三重态相关货币进行相互转换的报告，（28-30），我们认为，Φ _TTA此处定义是更一般的并包围定义。Φ _TTA然后可以将其进一步划分为接触对形成效率（_Φcp）和自旋分数（f）分量（_ΦTTA = _Φcp · f）。如上文关于UC所述，还可以报告归一化的TTA效率（2× _ΦTTA = _ηTTA）。然后，产率可与彼此使用表3中提供的等式它们已被划分成等同Φ _UCS，Φ _UCG等，以（1）的各个步骤或（2）实验性的可测量参数。用于添加剂的步骤，我们简单地乘以Φ _ÿ和1 - Φ _ÿ分别用于生产和非生产过程。虽然概念上方便，这是更为有可能使用实验可测量参数（Φ _UC，Φ _˚F，Φ _ET，Φ_出来，和Φ _LH）来计算Φ _UCS，Φ _UCG等所示的最右侧的列表3.但是，通过结合使用两列，很重要但很难直接测量诸如_{ΦTTA之类的}参数可以确定。这些计算的一个警告是，它仅假设如图1所示的反应路径。必须小心排除和/或考虑其他淬灭机理，例如涉及氧化还原活性物质和溶剂依赖性过程的淬灭机理。（31由氧）竞争性淬火为了简化省略，但可以通过由（1乘以包括- Φ _OS）。对于应用如太阳能转换中，感兴趣的效率度量不是基于所吸收的光子的数目（＃¹ S *），而是入射的光子数（＃ħ ν_中）。为了解决这个问题，我们可以包括Φ _LH（＃¹ S * /＃ħ ν_在），也被称为捕光效率或LHE，作为乘数转换任何上述产率从内部量子效率（Φ _Ý）到的外部量子效率（Φ _YX）例举Φ _UCsenx在表3的底行中。差异是吸收截面和光源与敏化剂吸收之间的光谱重叠。上面的讨论集中在分子敏化剂上。然而，近来，无机材料如量子点（32,33）和钙钛矿纳米颗粒/薄膜（34,35）成为了TTA-UC的有效三重态敏化剂。虽然大多数讨论都适用于无机敏化剂，但在图2中突出显示了一些值得注意的区别。图2.无机敏化剂促进的TTA-UC的能级图，机理和状态/物种（IS），三重态能量中继分子（A _r）和分子an灭器（A）。为了清楚起见，不再强调图1中的冗余步骤。第一个变化是，由于自旋态的小高能差和/或不可区分的性质的，所述无机增感剂（ISS）缺乏鲜明¹ IS *和³ IS *和Φ _ISC很可能是不必要的。（36）这样，此处将激发态简单标记为IS *。其次，由于寿命短，它们通常使用表面结合的分子三重态能量转移继电器（A _r）在溶液中TET转化为A（_ΦTET）之前在表面上存储长寿命的三重态。这为中继步骤增加了一个附加的，有时是可逆的TET，此处标记为TETr和_ΦTETr（37）返回的能量转移，也可能发生从¹ A *要么三重态敏化剂（Φ _ET）或中继（Φ _ETR）与前者典型地是因为IS的宽频带吸收性质更占优势。的激烈，宽吸收也在增加，因为内滤波（即，降低Φ损失结果_出），其进而降低Φ _UC。尽管如此，上面的表达式可以容易地应用到通过去除Φ系统_ISC和由Φ乘以_的TetR和1 - Φ _ETR。同样，对于ISC可逆（rISC）的热激活延迟荧光（TADF）敏化剂，也可以设想类似的模型。[38]另外，关于TTA- UC文学。首先是报告“反斯托克斯转移”。根据IUPAC标准，斯托克斯位移是由相同电子跃迁的吸收和发射峰值最大值之间的能量差（eV或cm ^–1）定义的。（20）因为TTA-UC吸收和发射过程并非源自相同的电子转换，更恰当地描述为表面上的反斯托克斯移位。另外，表观抗斯托克斯频移的大小应报告为敏化剂的最低能量吸收峰与hil灭器的发射峰之间的差（39），而不是激发波长。前者基于分子的固有特性，并且对于给定的样品而言是恒定的。后者由激发源决定，即使对于相同的TTA-UC样品，该值也会根据激发源的可用性而变化。此外，后者可能具有欺骗性，因为您可以通过使用更高强度的激光并激发到低能量吸收率的肩部来人为地增大位移。二是，我_个值，即最大效率方案的起始阈值强度，（40）应以原始光源强度（mW / cm ²）和激发密度（ex s ^–1 cm ^–2；光子通量× _λex的吸光度）报告）。（12）前者很有用，因为它可以直接比较该波长下的可用太阳通量（下称vida），但由于它与激发波长有关而受到限制。后者（即ex s ^–1 cm ^–2）可以直接比较具有不同集光效率，激发波长，激光源与敏化剂吸收最大值之间的重叠度不理想的系统。最后，在亚太阳光通量下激活的TTA-UC不仅意味着用强度小于100 mW / cm ^2的单色光激发样品。对于AM1.5太阳光谱，100 mW / cm ²是在250至2500 nm整个太阳光谱范围内的积分强度。在将单色激发与太阳强度进行比较时，至关重要的是仅对通常小于5 mW / cm ²的敏化剂激发范围（即激光器的fwhm或敏化剂吸收峰）的AM1.5太阳光谱进行积分。。根据这一点，我们强烈鼓励作者中止对入射功率（例如极低功率，超低强度等）的主观和定性描述，因为（1）“低功率”在太阳能转换中的含义大不相同。例如，光动力疗法，以及（2）有更多定量方法来定义这些术语。总结一下，这是该观点的主要内容：

• 我们建议，TTA-UC发射量子产率（Φ _UC）是每光子数观察到UC的光子数被吸收（即，光发射量子产率的传统定义）。
• 我们强烈鼓励使用2×Φ的_UC归一化因子，但如果使用的话，应当清楚地表示为归一化的TTA-UC发射效率（η _UC）。
• 应注意在记录任何输出耦合损耗，因为它是观察到的（Φ之间的差_UC和本征（Φ）_UCG）TTA-UC发射量子产率。
• 在弛豫之前确定上转换态（_ΦUCs）的产量，就可以在所有TTA-UC系统和应用中进行比较。
• 表观抗斯托克斯位移的大小应根据the灭器的发射峰与敏化剂的最低能量吸收峰之间的差而不是激发波长来报告。
• 应根据源强度（mW / cm ²）和激发密度（ex s ^–1 cm ^–2）报告第I_个值。
• 当声称TTA-UC发生在次太阳通量时，作者应仅对激发范围（<5 mW / cm ²）上的光谱进行积分，而不应对整个太阳光谱（100 mW / cm ²）进行积分。
• 当更容易获得定量描述时，作者应避免对入射光功率进行定性描述。

我们建议，TTA-UC发射量子产率（Φ _UC）是每光子数观察到UC的光子数被吸收（即，光发射量子产率的传统定义）。我们强烈鼓励使用2×Φ的_UC归一化因子，但如果使用的话，应当清楚地表示为归一化的TTA-UC发射效率（η _UC）。应注意在记录任何输出耦合损耗，因为它是观察到的（Φ之间的差_UC和本征（Φ）_UCG）TTA-UC发射量子产率。确定上转换态（_ΦUCs），然后再放宽，即可在所有TTA-UC系统和应用之间进行比较。表观抗斯托克斯位移的大小应根据the灭器的发射峰与敏化剂的最低能量吸收峰之间的差而不是激发波长来报告。应根据源强度（mW / cm ²）和激发密度（ex s ^–1 cm ^–2）报告第I_个值。当声称TTA-UC发生在次太阳通量时，作者应仅积分激发范围（<5 mW / cm ²）上的光谱，而不应积分整个太阳光谱（100 mW / cm ^2））。当更容易获得定量描述时，作者应避免对入射光功率进行定性描述。本观点表达的观点仅为作者的观点，不一定是ACS的观点。作者宣称没有竞争性的经济利益。YZ和KH在国家科学基金会的资助下，获得了DMR-1752782的资助。FNC在美国能源部科学办公室，基础能源科学办公室的资助下获得了DE-SC0011979奖。TWS得到了澳大利亚研究理事会（Exciton Science卓越中心CE170100026）的支持。本文引用了其他40种出版物。