Thus, customers sometimes complain about the seemingly poor performance of perfectly operating devices, as the spectral and temperature conditions may deviate significantly from conditions under IEC. Especially for owners of private installations without much technical background, it can be hard to understand this difference, which sometimes lowers the customers’ acceptance. This challenge should be much less present in IPV applications, once an indoor characterization standard has been established for IPV.
With an optimized band gap and resulting choice of material, IPV efficiencies exceeding 50% can be achieved theoretically [10].
Besides the spectral conditions, the typical cell size and the available radiation intensity differ from outside applications. While outdoor intensities range around 100–1000 Wm-2, indoor intensities are typically between 0.1–10 Wm-2 [11–15]. As a result, the ratio of photoelectric current to electrical loss currents is reduced by orders of magnitude for many loss mechanisms.
The small size in the square centimeter or millimeter scale instead of square meters leads to a stronger influence of miniaturization effects. The loss effects due to shading from module integration and contacting, contact areas and surface areas cannot be neglected as in outdoor applications, but may reach the scale of the photovoltaic current. In real systems, the dominating loss effects depend on the specific material of choice, its thickness and its characteristics, such as doping and so forth. Thus, it is highly recommended that IPV designers first choose their material and then model realistic values for their spectral application conditions based on their choice.
The applicability of efficiency models depends, among other things, on the choice of material. For example, the well-known Shockley-Queisser limit refers to the electrochemical conversion limit of an ideal absorber material where, unlike in real systems, each photon contributes to the photovoltaic current [16, 17]. Indirect band gap semiconductors, such as silicon (Si), have other dominating loss mechanisms than direct band-gap materials such as GaAs, and all of those effects are temperature and intensity dependent. Optimization methods include doping and change of layer thicknesses of the n- and p-layer, respectively. A great introduction into semiconductor modeling in general has been given by Sze and Lee and by Hovel for solar cells in particular [18, 19]. For amorphous Si, these models have to be adjusted according to the material properties and effects [20, 21]. In organic materials, exciton processes have to be modeled [22]. In order to overcome this challenge for IPV, Freunek has recently published a handbook that shows material-dependent and application-specific models for photovoltaic efficiencies [23].
Gemmer, who investigated realistic indoor cell efficiencies for c-Si, a-Si and CIGS with analytical and numerical models, presented the first modeling study for photovoltaic performance under indoor conditions [20, 24]. A current limit to the use of available photovoltaic simulation programs in IPV is their optimization for outdoor applications in their efficiency models and spectra. In addition, in most programs the numerical models have been developed for silicon only and neglect low irradiance or diffuse illumination effects.
For Si, GaAs, and CdTe, Bahrami-Yekta and Tiedje investigated the indoor efficiency limits and the optimization of real devices in indoor conditions in detail [25]. With an absorption layer thickness about two orders of magnitude below the ones of standard outdoor cells, Si devices can achieve or even outperform their outdoor performance under artificial light. Table 3.1 summarizes their results for three materials. It is important to note that spectra for fluorescent tubes can vary significantly with manufacturer and lamp type.
Table 3.1 Indoor efficiencies modeled by Bahrami-Yekta and Tiedje for different photovoltaic materials (adapted from [25]).
| Material | Efficiency FL250 Lux [%] | Efficiency LED [%] |
| Si | 27.0 | 29.0 |
| GaAs | 37.1 | 40.3 |
| CdTe | 40.3 | 43.3 |
As Chapter 6 shows similar results to that, the optimization goals for indoor organic photovoltaics (OPV) are contradictory to outdoor applications. For IPV, the influence of the serial resistance in organic devices can be neglected whereas the parallel resistance needs to be maximized.
Customers, product engineers, and researchers require reproducible and retraceable characterization methods in order to compare research results and products. The efforts to establish a standard characterization process has just begun and is outlined in Chapter 5 of this book. Methods to overcome the current gaps in standards are described in Chapters 4 and 5.
The following sections outline general operating conditions, efficiencies and product aspects for the state-of-the-art at the time of writing.
3.2 Indoor Spectra and Efficiencies
Outdoor spectra are broadband spectra resulting from the thermal radiation of the sun. The human eye is sensitive only to a small fraction of it, which has also been characterized as the human visibility function V(λ) (see Chapter 4 and 5 for detailed introductions). The users of artificial light in indoor environments are humans, plants, and animals, so their radiation is optimized for their optical range of sensitivity, such as V(λ). Broadband radiation sources in indoor environments are halogen lamps and incandescent bulbs, as well as solar radiation filtered through a window. Depending on the place of installation, windows are manufactured with functional coatings of a spectral transmittance T(λ), such as heat or sun protection coatings. In general, the filtering function of these windows is optimized to V(λ). Thus, the indoor solar spectrum is different to the outdoor spectrum not only in its intensity, but also in its spectral distribution. Figure 3.1 depicts typical indoor spectra. The window spectra result is from the spectral irradiance of the so-called AM 1.5 solar spectrum [26] that is transmitted through a heat and sun protection coating of a window, respectively (see Chapter 4 for a detailed introduction).
As windows in most applications do not face the sky horizontally, the actual irradiance E will result from the incident power P received on an area A in an incident angle α following Eq. (3.1).
In some applications, a sensor might be directly installed on a window facing the outdoors. However, most devices will be installed within a certain distance to the window. As the intensity is inverse to the square of the distance from the emitter, the incident irradiance is additionally lowered following Eq. (3.2)
