Open Access
Issue
EPJ Photovolt.
Volume 15, 2024
Article Number 34
Number of page(s) 15
DOI https://doi.org/10.1051/epjpv/2024030
Published online 21 October 2024

© M.T.S.K. Sen et al., Published by EDP Sciences, 2024

Licence Creative CommonsThis is an Open Access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

1 Introduction

Currently, the conversion efficiencies of conventional homojunction crystalline silicon (c-Si) solar cells are mainly limited by the recombination of charge carriers occurring at the metal/silicon interface. This limitation is minimized by including a stack of passivating and carrier-selective layers in between the Si absorber and metal electrodes which effectively suppresses the recombination at the c-Si surface while simultaneously being conductive to either electrons or holes generated in the c-Si absorber. Nowadays, the highest conversion efficiency of Si solar cells comprising amorphous silicon heterojunction (SHJ) or doped polycrystalline Si contacts are manufactured by using this type of so-called passivating contact scheme [1,2]. Nevertheless, these highly doped Si-based passivating contacts are a source of parasitic absorption, which consequently reduces the total amount of photogenerated carriers inside the Si absorber, resulting in a lower current. For instance, doped poly-Si, typically used in passivating contact structures, suffers from high parasitic absorption when heavily doped [3]. In the case of amorphous silicon, its direct band gap of approximately 1.7 eV in combination with its heavily doped layers hinder the short-circuit current density (Jsc) of SHJ solar cells [4,5].

In order to minimize the Jsc losses sustained by these layers, alternative transparent selective contacts, such as metal oxides have been explored due to their wide bandgap, capability to extract charge carriers, passivation quality on c-Si, and the relatively simple physical vapor deposition (PVD) techniques that have been used to fabricate them [6,7]. For instance, evaporated MoOx has been investigated for its transparency in the blue wavelength region and its ability to act as a hole-selective contact for c-Si solar cells [811]. Recently, a conversion efficiency of 23.83% has been achieved by replacing the p-type hydrogenated amorphous silicon (a-Si:H(p+)) layer with a MoOx hole-selective contact in a SHJ solar cell structure [12]. Nevertheless, a MoOx/a-Si:H(i)/c-Si contact has been shown to degrade considerably at standard SHJ annealing conditions [13,14], causing the appearance of an S-shaped current–voltage (IV) curve, and accordingly a loss in fill factor (FF). Additionally, while the doped a-Si:H layer is omitted in that structure, the intrinsic a-Si:H layer is still used, meaning that the high transparency of the MoOx layer is not fully exploited due to the absorptive nature of the intrinsic a-Si:H (a-Si:H(i)) layer. The lack of thermal stability of the MoOx/a-Si:H(i)/c-Si contact is a significant limitation in the potential manufacturing of the SHJ-like solar cell, since an annealing temperature of about 200 °C is often required to recover from sputtering damage after deposition of the transparent conductive oxide (TCO) layer and is an essential back-end treatment during metallization and TCO post-crystallization. While a MoOx/a-Si:H(i)/c-Si contact provides excellent surface passivation properties, the insertion of an a-Si:H(i) interlayer poses several transport issues which could result in the formation of the S-shaped IV curve.

In search of alternative passivating interlayers, ultrathin SiO2 layers (<2 nm) are a potential candidate in replacing a-Si:H(i) due to their ability to provide excellent surface passivation and contact selectivity when combined with doped poly-Si contacts [15]. However, unlike poly-Si contacts where excellent passivation properties can be achieved, the insertion of an oxide interlayer results in apparent hole collection issues when combined with MoOx even though sufficient band bending is typically obtained at the c-Si interface [1619]. Conversely, the selectivity of MoOx contacts with an Al2O3 interlayer has shown to be promising with good surface passivation and contact resistance properties [20]. In addition, we recently showed that an ultrathin atomic layer deposited (ALD) Al2O3/SiOy interlayer stack does not impede the hole selectivity provided by the MoOx contact, resulting in good contact selectivity and cell performance [21]. Note that the ultrathin SiOy is naturally formed at the c-Si surface during the initial cycles of the ALD Al2O3 process, as has been repeatedly documented elsewhere [22,23].

The aim of this work is to further understand the interaction between the salient factors of the passivating interlayer, or interlayer stack, and MoOx contact that influence contact selectivity. A comparative study between different interlayers, notably a-Si:H(i), SiO2, and Al2O3/SiOy is made where the non-ideal carrier extraction behavior, caused by the insertion of an interlayer on our MoOx-based contact, is addressed. Since the carrier transport mechanisms of the aforementioned contacts involve different transport mechanisms, such as band-to-band tunneling, defect-assisted transport, thermionic emission, and pinhole aided transport, a simple model is developed to encompass the different transport properties of these interlayers. In this model, the carrier transport through the interlayer is represented by a thin layer with limited effective charge carrier mobility. This simplified approach facilitates a meaningful comparative interpretation of the impact on majority carrier transport across the interlayers in relation to the observed loss in carrier selectivity.

2 Methodology

2.1 Solar cells fabrication and characterization

To investigate the impact of the different interlayers on the solar cell performance, n-type c-Si solar cells with MoOx-based contacts at the front are studied. A rear poly-Si(n+) contact is used because of its excellent surface passivation quality, contact resistance, and thermal stability, such that the front MoOx-based contact is limiting in the measurements and not the rear poly-Si contact.

The manufacturing steps of the solar cells with front side thermal SiO2, Al2O3/SiOy or a-Si:H(i) interlayers and the cell schematic are shown in Figure 1. The 6-inch, 180 μm-thick pseudo-square Cz c-Si(n) substrates with a resistivity of about 3 Ω cm were processed as follows: textured in a KOH solution, pre-gettered with POCl3 diffusion followed by phosphosilicate removal, surface smoothing etch, and finally cleaned in RCA 1 and 2, and nitric acid oxidation step (NAOS) solutions. Note that the surface smoothing is only a minor surface treatment in the sense that the textured morphology is still preserved overall. The substrates were dipped in a 1% HF solution prior to the formation of the rear SiO2/poly-Si(n+) contact. The SiO2 interlayer (∼1.3 nm) was thermally grown in a low pressure chemical vapor deposition (LPCVD) chamber at 610 °C using an O2 and N2 mixture. Subsequently, the poly-Si(n+) layer was formed through a two-step process: initial deposition of a 20 nm thick a-Si:H(n+) layer using plasma-enhanced chemical vapor deposition (PECVD), followed by a 20 minute thermal annealing in N2 at 900 °C to induce crystallization. After this poly-Si(n+) formation process, the contact was subjected to a hydrogenation process using PECVD with NH3 plasma at 375 °C to reduce interface defects and improve passivation. In the case of an ultrathin SiO2 interlayer, the thermal oxide at the front side was preserved. The samples with a-Si:H(i) and Al2O3/SiOy interlayers were dipped in a 1% HF bath to remove the front thermal oxide. Subsequently, the a-Si:H(i) and Al2O3/SiOy interlayers were deposited using PECVD and spatial ALD tools, respectively. 8 cycles of spatial ALD Al2O3 were used to deposit a 1.5 nm thick Al2O3/SiOy stack. These layers were deposited using a Levitrack deposition tool without any subsequent post-deposition annealing treatment. The samples without interlayer only received an HF dip prior to MoOx deposition. Next, the samples were transferred to an electron beam physical vapor deposition (PVD) tool where a 5 nm MoOx layer was deposited in a high vacuum (7 × 10−6 mbar) environment. Note that minimal X-ray damage to the surface passivation of the MoOx contacts was observed after the electron beam deposition, likely due to the low deposition power used, as shown in Appendix A. Finally, indium tin oxide (ITO) films were deposited in a sputtering tool on both sides of the samples and a full area Ag sheet was also deposited at the rear side by sputtering. To finalize the solar cells, a front Ag grid was screen printed using a low temperature Ag paste and the device was cured in air at different temperatures (190–250 °C).

The photoconductance of charge carrier lifetime samples was measured by using a Sinton WCT-120 system, as well as the internal voltage expressed in terms of the implied open circuit voltage iVoc. The external Voc of the half-fabricated cells were measured by a SunsVoc Sinton tool which does not require metal contacts due to the conductivity of the ITO films, as shown in Figure 2. The solar cells were characterized by IV measurements in a Wacom AAA solar simulator at standard test conditions. The results were corrected for spectral mismatch. Dark IV measurements were performed at varying temperatures between 25 and 65 °C using the Wacom solar simulator. The interlayer films were deposited on single-side polished c-Si wafers with unpolished backsides to eliminate back reflections during ellipsometry measurement. The thickness of the interlayers was determined using a spectroscopic ellipsometer (J.A. Woollam Co., Inc.). Ellipsometric data was collected at three angles of incidence (60°, 70°, and 80°) and over a photon energy range of 1.1 to 5 eV. The Cauchy model was used to analyze oxide layers.

thumbnail Fig. 1

Process flow for the manufacturing of the solar cells with different interlayers (left) and a schematic of the solar cells (right).

thumbnail Fig. 2

Solar cell precursors with front MoOx and rear poly-Si(n+) contacts for iVoc and SunsVoc measurements.

2.2 Simulations and calculations

2D simulations were performed by using the Atlas package of Silvaco [24]. Figure 3 depicts the cell structure used in the simulations. The front contact consists of an ultrathin interlayer with limited carrier mobility, hole-selective layer with a varying work function (WF), and a metal electrode. The rear electron contact is built similarly to the hole contact but consists of optimized parameters that minimize the recombination and contact resistivity.

This model intends to simulate the effects of the surface passivating interlayers and MoOx layer on the carrier selectivity of the hole contact. Therefore, for the sake of simplicity, we mimic the effect of MoOx by using the properties of a p-type material based on a poly-silicon contact as shown in [25]. In this regard, the transport of holes is simplified at the hole contact and metal electrode. This assumption is only valid for comparison purposes as we focus on the study of interlayer with different WF for the hole contact layer. In this simplification, the variation in the WF of this contact is represented by changing the doping level of the p-type poly-Si hole contact. We assume the carrier mobility in the interlayer as the parameter affecting the selectivity of the hole contact [26,27]; the hole mobility (µh) value characterizes the actual physical mechanism of the charge transport across the interlayer [25]. This simplified approach facilitates a meaningful comparative interpretation of the impact on majority carrier transport across the interlayers in relation to the observed loss in carrier selectivity, while avoiding unnece - ssary complexity introduced by incorporating different possible transport mechanisms throughout the interlayers.

Accordingly, in this study, simulations with different mobility values emulate the behavior of the device using different interlayer materials. Table 1 shows the values of the simulation parameters for the c-Si(n) absorber as well as the hole and electron contacts. The interlayer is modelled as a 1.5 nm c-Si(i) thick layer where µh is varied. Mobility values are varied between 10−2 and 10−7 cm2 V−1 s−1 in this simulation work. Mobility values around 10−2 cm2 · V−1 · s−1 are indicative of an interlayer with minimal transport resistance. Conversely, an exceedingly low value of 10−1 cm2 · V−1 · s−1 or lower corresponds to resistances calculated for the quantum tunnelling of holes through an ideal, defect-free SiO2 interlayer [25,28]. The electron mobility of the interlayer is set at 10−5 cm2 · V−1 · s−1 which represents electron carrier mobility for a tunneling SiO2 and a-Si:H(i) interlayers. The effective surface recombination velocity (Seff) of the hole contact is a crucial parameter in our simulation model, as it directly influences the recombination losses at the interface. In our analysis, the Seff value is systematically varied over a wide range, from 10 cm/s (representing a nearly ideal passivated surface) to 105 cm/s (indicating a highly defective interface). By simulating the device performance across this range, we can assess the sensitivity of ΔVoc and FF to variations in Seff, and therefore quantify the impact of the hole contact quality on overall the solar cell efficiency.

thumbnail Fig. 3

Schematic of the cell structure used in Atlas to simulate the contact selectivity of the cell with varying interlayer µh and hole WF.

Table 1

Atlas simulation parameters for the different layers.

3 Results

3.1 Interlayer surface passivating properties of MoOx contacts

In this section, we investigate and compare the influence of different interlayers, i.e. a-Si:H(i), ultrathin spatial ALD Al2O3/SiOy, and thermally grown SiO2, on the surface passivation provided by the contact structures. The thickness of the thermal oxide − measured by ellipsometry − is around 1.3 nm and the oxide layer is combined with our poly-Si contacts. In comparison, the combined thickness of the Al2O3/SiOy stack is around 1.5 nm after 8 spatial ALD cycles to grow Al2O3. To analyze the surface passivation properties of the interlayers on our MoOx contacts, the iVoc value is monitored after the deposition of MoOx and ITO layers, and a subsequent annealing at 190 °C, as shown in Figure 4. The distribution of iVoc values for each group, represented by box and whisker plots, is based on measurements from five samples per group. The MoOx contact without an interlayer shows poor surface passivation, which can be mainly attributed to the poor surface passivation properties of the sub-stoichiometric oxide formed during the initial growth of the evaporated MoOx layer [31]. However, the surface passivation of our MoOx contacts improves by introducing the thermally grown SiO2 and ALD grown Al2O3/SiOy interlayers. On the other hand, excellent surface passivation is achieved by using an a-Si:H(i) interlayer. iVoc above 700 mV is achieved after an annealing treatment at 190 °C. Subsequently, ITO layers were sputtered on the front and rear contacts. Interestingly, iVoc improves for cell precursors with an Al2O3/SiOy interlayer and without interlayer. In contrast, iVoc decreases for precursors with SiO2 and a-Si:H(i) interlayers due to the sputtering damage originated from the ITO deposition [32,33]. However, the sputtering damage can be partially recovered after an annealing treatment. Samples with an Al2O3/SiOy interlayer and with no interlayer show no change in iVoc after annealing at 190 °C.

thumbnail Fig. 4

iVoc of half-fabricates after deposition of MoOx and ITO layers and subsequent annealing at 190 °C. Note that the order of the interlayer configurations (no interlayer, Al2O3/SiOy, a-Si:H(i), and SiO2) is consistent across all figures.

3.2 Effect of interlayer properties on the contact selectivity of MoOx contacts

To investigate the contact selectivity of our MoOx contacts, we use the difference in internal and external VocVoc = iVoc-Voc) as a simple figure of merit with low values signifying a high carrier selectivity [34]. In case of low contact selectivity, the external Voc of the cell is much lower than the internal Voc, resulting in a high ΔVoc value. This implies that the transport of majority carriers to the electrode is hindered. Note that the cell precursor used to measure the iVoc and Voc is shown in Figure 2.

Figure 5 shows ΔVoc for different MoOx contacts in their as-deposited states and as a function of annealing temperature. These samples were annealed in air at a starting temperature of 190 °C − which represents the standard SHJ annealing conditions − followed by cumulative annealing up to 250 °C, with a 20 °C temperature step. The insertion of a-Si:H(i) and Al2O3/SiOy interlayers does not affect the ΔVoc prior to annealing and results in comparable ΔVoc to the MoOx/c-Si contact. In the case of a thermal SiO2 interlayer, a high ΔVoc value of about 258 mV is observed prior to annealing and no major change in ΔVoc is observed after subsequent annealing. For the MoOx/a-Si:H(i) contact, a steady increase in ΔVoc from 15 to 30 mV is measured upon an increase in thermal budget, consistent with previous literature [14,35]. This decrease in selectivity is attributed to a reduction in the induced band bending at the MoOx contact. This reduction in band bending is likely due to a decrease in the MoOx WF, potentially caused by hydrogen effusion from the a-Si:H(i) interlayer and/or the formation of a parasitic layer at the MoOx/a-Si:H(i) interface [14,35]. In comparison, the contact selectivity of the MoOx/Al2O3/SiOy contact improves upon annealing at 190 °C and remains stable after further increases in annealing temperature; an average ΔVoc of about 5 mV is measured after an annealing treatment at 230 °C with a slight increase observed following annealing at 250 °C.

thumbnail Fig. 5

ΔVoc behavior of different MoOx contacts as a function of cumulative annealing (190–250 °C).

3.3 Effects of passivating interlayers on IV characteristics

In this section, we investigate the influence of the passivating interlayers on the light IV parameters. The light IV curve and characteristics of the solar cells are shown in Figure 6 and Table 2. As expected, a MoOx/a-Si:H(i) contact results in a high Voc due to the excellent surface passivation of the a-Si:H(i) interlayer, but is limited by the contact selectivity loss after annealing. On the other hand, while ultrathin Al2O3/SiOy and SiO2 interlayers have shown a similar surface passivation quality (as shown in Appendix B), the high carrier selectivity loss of the MoOx/SiO2 contact results in a lower Voc and FF in comparison to the MoOx/Al2O3/SiOy contact. Finally, solar cells with a-Si:H(i) and Al2O3/SiOy interlayers at the hole contact result in comparable conversion efficiencies just above 18%; the Jsc and FF values are higher for the MoOx/Al2O3/SiOy contact due to superior transparency and carrier selectivity, respectively.

thumbnail Fig. 6

Influence of passivating a-Si:H(i), Al2O3/SiOy, and SiO2 interlayers on the light IV characteristics of the solar cells with MoOx contacts and different interlayers (right).

Table 2

Measured IV characteristics for solar cells with different interlayers and without an interlayer.

3.4 Interlayer transport: temperature-dependent dark IV

Further insights to explain the difference in ΔVoc associated with different oxide interlayers can be acquired by performing a temperature-dependent dark IV analysis. Series resistance (Rs) is extracted from the 2-diode model for temperatures in the 25–65 °C range for the SiO2 and Al2O3/SiOy interlayers, as shown in Figure 7. An indirect measurement of the band offsets between the c-Si and the interlayer can be made by extracting the activation energy Ea from the slope of the temperature-dependent series resistance Rs by assuming an Arrhenius dependency [36]. As a result, we obtain Ea values of 117 meV and 2390 meV for the MoOx/Al2O3/SiOy and MoOx/SiO2 contacts, respectively. Ea of the Al2O3/SiOy interlayer is considerably lower than the thermally grown oxide which indicates that an inefficient hole carrier transport exists for the thermally grown SiO2 interlayer. To further determine the effect of hole majority transport on the carrier selectivity of the contact, we simulate the effect of the interlayer hole mobility and contact WF on the hole contact properties.

thumbnail Fig. 7

Dark JV measurements were used to extract Rs as a function of cell temperature (25–65 °C). The fitted lines are used to calculate the activation energy.

3.5 Simulation of MoOx contacts

3.5.1 Effect of surface passivation on the carrier selectivity

In this section, the influence of the surface passivation and µh properties of the interlayers on the carrier selectivity are simulated with respect to varying hole contact WF. We first investigate the influence of the surface recombination properties of the interlayer on the hole selectivity. Figures 8a and 8b show the simulation results of the dependence of the hole selectivity on Seff, contact WF, and µh. As expected, a lower hole contact WF leads to a loss in ΔVoc due to a decreased hole concentration near the interface. The decrease in concentration elevates the hole resistance (Rh) across the contact, which is inversely proportional to the both carrier concentration and the hole mobility [26].

Additionally, Seff of the interlayer has a significant influence on the ΔVoc of the contact, particularly for interlayers with low mobility (μh = 10−7 cm2 V−1 s−1). Interestingly, an increase in ΔVoc is observed with decreasing Seff. This counterintuitive behavior can be explained by the competing contributions of hole resistance within the interlayer (Rh,int) and the absorber (Rh,abs), as highlighted by Onno et al. [37]. ΔVoc depends on both components, and the relationship is described by:

ΔVoc=(Jr,surf+Jr,m)Rh,abs+Jr,mRh,int(1)

where Jr, surf and Jr,m are recombination current density at the Si surface and metal contact, respectively. Consequently, for an interlayer with μh = 10−7 cm2 V−1 s−1, decreasing Seff shifts the recombination dominance towards the surface, increasing Jr,surf at the expense of Jr,m.

thumbnail Fig. 8

Simulated ΔVoc as a function of Seff for the hole contact shown for WF varying from 5.09 to 5.33 eV and for µh of (a) 10−5 cm2V−1s−1 and (b) 10−7 cm2V−1s−1.

3.5.2 Effect of interlayer hole mobility on the carrier selectivity

Figures 9a and 9b show the simulated interacting effect of µh and contact WF on the cell ΔVoc and FF, respectively. Seff of the contact was set to 33 cm/s (as shown in Appendix B) which approximately represents the surface passivation of the MoOx contacts with Al2O3/SiOy and SiO2 interlayers. At WF > 5.25 eV, ΔVoc is minimal since the majority concentration is high enough which effectively reduces Rh,abs (referring to Eq. (1)), even for a interlayer with low mobility (µh = 10−7 cm2 V−1 s−1). This reduction in Rh,abs allows for efficient hole extraction, minimizing recombination losses and maintaining a high Voc, even with less optimal interlayer transport properties. However, for this high WF, a steep decrease in FF is noted at µh = 10−7 cm2 V−1 s−1. At moderate WF (5.1–5.2 eV), noticeable selectivity and FF losses are observed with a strong dependence on µh; a decrease in µh of the interlayer yields higher ΔVoc and FF losses. However, at WF < 5.1 eV, contact selectivity cannot be maintained anymore even for high µh of the interlayer. The simulation results indicate that for both the SiO2 and Al2O3 interlayers the MoOx WF in the range of 5.1–5.2 eV is found. This estimation is supported by the close agreement between the simulated ΔVoc values and the experimental measurements presented in Figures 5a and 9a, respectively. The moderate WF range is defined as the values where a good selectivity can be achieved with sufficient hole mobility provided by the interlayer. The significantly high ΔVoc and FF losses, measured for our MoOx/SiO2/c-Si(n) contact, suggest that the SiO2 interlayer corresponds to a µh in the vicinity of 10−7 to 10−8 cm2 V−1 s−1. This observation is consistent with the characteristics of the thermal SiO2 interlayer and supports the proposed mobility for a defect-free SiO2 interlayer, with a thickness ranging from 1.1 to 1.5 nm [25]. On the other hand, the evident reduction in both ΔVoc and FF within our MoOx/a-Si:H(i)/c-Si(n) contact, as influenced by an elevated thermal budget treatment, exhibits a good correlation with an interlayer mobility of about 10−5 cm2 V−1 s−1 or slightly higher. Finally, the good ΔVoc and FF values obtained with the MoOx/Al2O3/SiOy stack imply that the interlayer presents minimal resistance to hole carriers, and likely possesses a high µh value of about 10−2 cm2 V−1 s−1.

To investigate the impact of µh on recombination losses and cell efficiency, we analysed the simulated JV curves and recombination current distributions. Figures 10a and 10b depict the simulated JV curves and recombination current densities at the p- and n-contacts, and within the absorber (bulk) for µh values of 10−5 and 10−7 cm2V−1s−1, respectively, at a hole contact WF of 5.21 eV. At µh = 10−5 cm2 V−1 s−1, the total current (extracted current minus recombination current) follows a diode-like behaviour, resulting in a high FF. Here, recombination within the bulk absorber is the primary efficiency-limiting factor. However, when µh is reduced to 10−7 cm2 V−1 s−1, the FF decreases significantly, and the JV curve exhibits an S-shape. This indicates that while bulk recombination remains dominant, it no longer follows a simple diode behaviour. The high Rh at the low-mobility interlayer impedes hole transport, forcing the majority hole carriers holes to recombine within the c-Si bulk, as shown by the Jr in Figure 10b.

thumbnail Fig. 9

Simulated effect of varying µh (10−2–10−7cm2V−1s−1) and hole WF contact (4.92–5.34 eV) (a) on ΔVoc, (b) and on FF. Seff is set to 33 cm/s which is representative of the surface passivation quality of Al2O3/SiOy and SiO2 interlayers with a MoOx contact. Electron mobility of the interlayer is set to 10−5 cm2V−1s−1.

thumbnail Fig. 10

Simulated JV curves, and recombination currents at the p- and n- contact, and in the absorber (bulk) as a function of cell voltage, for µh of (a) 10−5 and (b) 10−7 cm2V−1s−1.

4 Discussion

While the mobility model employed in this study simplifies carrier transport across the MoOx contacts, it is essential to acknowledge the complexity of the actual contacts. In reality, several transport mechanisms, such as thermionic emission, band-to-band tunneling, and trap-assisted tunneling, exist at the interfaces of the MoOx contacts. Additionally, the influence of the TCO layer and interlayer formation at the interfaces were not considered in this model, potentially introducing additional transport limitations. Nevertheless, as a comparative study, the presented model proves valuable in discerning differences in observed selectivity losses and recognizing the limitations imposed by the interlayers on the MoOx contacts. The following section explains the differences in contact selectivity of the MoOx contacts.

The combination of an a-Si:H(i) interlayer with MoOx contact shows excellent surface passivation properties, but results in a decreasing hole selectivity with increasing annealing temperature compared to an Al2O3/SiOy interlayer. Several factors can contribute to this difference: (1) the degradation of induced band bending with annealing temperature which is also exacerbated by the presence of an a-Si:H(i) interlayer − possibly attributed to a pronounced Fermi level pinning effect [38,39]; (2) high contact resistance resulting from an intermixed oxide region formed at the interface between the MoOx and a-Si:H(i) interlayer, combined with the sensitive alignment between the MoOx conduction band and valence band of the a-Si:H(i) interlayer [40,41]. The latter arises from the necessity of closely aligning the conduction band of the MoOx layer with the valence band of the a-Si:H(i) layer for efficient tunneling transport.

By omitting the a-Si:H(i) interlayer, good contact selectivity and FF were obtained but the MoOx/c-Si contact lacks surface passivation properties. The surface passivation and carrier selectivity of the MoOx contact improve by inserting an ultrathin Al2O3/SiOy stack, resulting in higher FF and Voc values, low Ea, and improved contact thermal stability. This improvement can be attributed to the high hole mobility of approximately 10−2 cm2 V−1 s−1 of the Al2O3/SiOy interlayer, which does not impede the extraction of majority hole charge carriers. This effect is linked to the amorphous and non-stoichiometric nature of the ultrathin SiOy layer formed at the c-Si surface, as shown in [21]. Furthermore, the considerably low Ea suggests a smaller valence band offset (VBO) between SiOy and c-Si, thereby facilitating the transport of holes. While both thermionic emission and tunneling can contribute to carrier transport through thin oxide layers, the low Ea value for the MoOx/Al2O3/SiOy contact is consistent with typical thermionic emission barriers reported for similar structures [42,43]. Similarly, several studies show that oxygen incorporation in a-SiOx:H interlayers result in an inefficient hole transport and consequently in an S-shaped IV curve [4,44]. In addition, Al2O3/SiOy films on c-Si substrates typically consist of high negative fixed charge properties which can promote the collection of holes as majority carriers; an inversion layer near the c-Si surface is created which increases the hole concentration. The negative fixed charged of our Al2O3/SiOy film was, indeed, detected by conducting a corona charge experiment. However, further work is required to quantify the magnitude of this fixed charge of the layer due to the quick dissipation of charges after corona charge deposition. Further details can be found in Appendix B.

In comparison, a high carrier selectivity loss is apparent for the thermally grown SiO2 interlayer although surface passivation properties similar to the Al2O3/SiOy interlayer were achieved; surface passivation affects the hole resistance near the c-Si surface which ultimately influences the hole selectivity [25,37]. Additionally, the large disparity in the calculated Ea between the two interlayers suggests differences in transport mechanisms. The high Ea for SiO2 indicates a larger energy barrier, likely hindering thermionic emission and suggesting that tunneling is the dominant transport mechanism, which is less efficient than thermionic emission. This is in contrast to the Al2O3/SiOy interlayer, where the low Ea suggests a smaller barrier resulting in a more effective thermionic emission. The observed ΔVoc and FF losses of the MoOx/SiO2/c-Si contact are likely the result of a significantly lower interlayer µh in the range of 10−7 to 10−8 cm2 V−1 s−1. This difference in layer mobility is likely caused by a large VBO with c-Si (4.7 eV) of the SiO2 interlayer which creates a large barrier for holes. As a result, a significant step in the quasi-Fermi level of the holes (EFp) is introduced at the interface, as illustrated in Figure 11, thereby reducing the current towards the hole contact. This step acts as an additional barrier to hole extraction, further impeding the flow of holes through the contact resulting in an increase in Jr,abs. This increases Rh,int which consequently contributes to the observed loss in carrier selectivity.

Although the contact selectivity loss can be reduced by increasing the contact WF, such high contact WF is often not feasible. In practice, a more effective approach is to enhance the interlayer mobility. For instance, in the case of poly-Si(p+) contacts, a post-deposition annealing step at high temperature is usually required for boron diffusion from the poly-Si to the SiO2 and c-Si absorber. The increase in boron concentration in the poly-Si(p+) increases the contact WF, but also causes an increase in boron diffusion inside the Si substrate, leading to a higher defect density at the interface. Nevertheless, this post-deposition annealing step is crucial in enhancing the hole mobility across the SiO2 interlayer; this process allows for the creation of pinholes and/or to reduce the interlayer thickness, thereby improving the transport of holes [45]. For MoOx/SiO2 contacts, a high temperature treatment is not viable due to the lack of thermal stability of the MoOx layer [46].

Additionally, the thermal instability of the MoOx contact poses challenges in implementing post-hydrogenation techniques to enhance the surface passivation at the Si/interlayer interface. Conventionally, the diffusion of hydrogen to the interface of poly-Si contacts can be achieved in several different schemes such as hydrogen-rich capping layers or a remote hydrogen plasma treatment [47,48]. However, in the case of MoOx contacts, similar hydrogenation techniques are challenging since the MoOx layer interacts with hydrogen thereby degrading the WF value of the MoOx layer [49]. The introduction of an ALD Al2O3 interlayer addresses some of these issues, allowing for improvement of the surface passivation without compromising on the contact selectivity. While further layer optimizations and post-deposition treatments on the Al2O3 interlayer can be developed to enhance the surface passivation properties, it must be ensured that these processes do not compromise the interlayer hole mobility, keeping it above 10−5 cm2 V−1 s−1.

thumbnail Fig. 11

Schematic band diagrams of hole contacts with Al2O3/SiOy (left) and SiO2 (right) interlayers, illustrating the contrasting energy barriers and hole and electron quasi-Fermi levels (EFp,n) alignment.

5 Conclusion

In this work, we highlight the importance of high hole contact WF to create a strong induced band bending near the c-Si interface and the necessity of a sufficient hole mobility through the interlayer to achieve an effective hole-selective contact. An a-Si:H(i) interlayer can provide excellent surface passivation, but the MoOx WF loss upon a thermal annealing treatment results in observable contact selectivity loss. On the other hand, a dense, stoichiometric, thermally grown SiO2 interlayer will cause considerable contact selectivity losses if no post-treatment is performed to improve the hole mobility. An ultrathin Al2O3/SiOy interlayer provides better transparency, hole transport, and thermal stability when combined with MoOx. This is because the sub-stoichiometric SiOy layer does not hinder the transport of holes across the Al2O3/SiOy interlayer. Hydrogenation strategies prior to the MoOx deposition can be explored to improve the surface passivation provided by the Al2O3/SiOy interlayer stack to ultimately improve the quality of MoOx-based contacts in c-Si solar cells.

Funding

The authors would like to thank Martien Koppes, Eelko Hoek, and Benjamin Kikkert for the fabrication of the solar cells. This work was financially supported by Top consortia for Knowledge and Innovation (TKI) Solar Energy programs “COMPASS” (TEID215022), “RADAR” (TEUE116905) and MOMENTUM (TKI Energy PPS Toeslag project number: 1821101) of the Ministry of Economic Affairs of The Netherlands.

Conflicts of interest

The authors have no conflicts to disclose.

Data availability statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Author contribution statement

Conceptualization, M.T.S.K. Ah Sen, P. Bronsveld, and G. Janssen.; Methodology, M.T.S.K. Ah Sen and G. Janssen.; Software, M.T.S.K. Ah Sen and G. Janssen.; Validation, M.T.S.K. Ah Sen and G. Janssen.; Formal Analysis, M.T.S.K. Ah Sen and G. Janssen.; Investigation, M.T.S.K. Ah Sen.; Resources, M.T.S.K. Ah Sen, P. Bronsveld, and A. Weeber; Data Curation, M.T.S.K. Ah Sen; Writing − Original Draft Preparation, M.T.S.K. Ah Sen.; Writing − Review & Editing, M.T.S.K. Ah Sen, G. Janssen, A. Mewe, P. Bronsveld, J. Melskens, F. Hashemi, P. Procel, and A. Weeber; Visualization, M.T.S.K. Ah Sen; Supervision, P. Bronsveld, J. Melskens, and A. Weeber; Project Administration, P. Bronsveld and A. Weeber; Funding Acquisition, P. Bronsveld, J. Melskens.

Appendix: A

We investigate the impact of x-ray-induced surface passivation damage, occurring during MoOx deposition, on the surface passivating properties of the interlayers. 6 nm thick AlOx layers were deposited on both sides of a c-Si(n) substrate, followed by post-deposition annealing (PDA) at 600 °C. To assess the effect of X-ray induced damage from MoOx e-beam deposition, a glass sheet is placed on top of the AlOx passivated sample, preventing MoOx deposition while allowing x-ray emission to pass through. Figure A.1 shows the effective lifetime (τeff) measured at a carrier concentration of 1015 cm−3 for both a reference sample and a sample subjected to the x-rays. AlOx samples with X-ray induced damage show similar τeff as the AlOx reference sample. The negligible difference between the samples implies that the induced X-ray damage has minimal effect on the surface passivating quality. This is possibly because of the low e-beam power used during the MoOx deposition; MoOx has a low sublimation point and therefore only requires little energy to evaporate.

thumbnail Fig. A.1

Comparing the surface passivation quality between passivated AlOx reference and x-ray exposed samples.

Appendix: B

To assess the impact of the passivating interlayers on the induced band bending, we use a corona charging setup by Delft Spectral Technologies for samples shown in Figure B.1. Further details about the corona charging tool can be found in references [4953]. The samples with AlOx, SiO2, and a-Si:H(i) interlayers were capped with MoOx, a thin layer of indium tin oxide (ITO), and 6 nm of AlOx to reduce the dissipation of charges. To ensure of more representative stack, a thin ITO layer was added and further capped with AlOx to minimize the potential charge dissipation.

thumbnail Fig. B.1

Symmetric samples consisting of AlOx, SiO2, and a-Si:H(i) interlayers, capped with MoOx, ITO, and AlOx.

Note here that Seff,max was calculated from the wafer thickness W and the effective minority carrier lifetime τeff (Seff,max = W / 2τeff) after conducting quasi-steady state photoconductance (QSSPC) measurements using a Sinton WCT-120TS setup in the generalized (1/64) mode. For the QSSPC measurements we assumed n-type substrates, a wafer thickness of 200 μm, an optical constant of 0.55 (for chemically polished substrates) and the τeff values at an injection level of 1 · 1015 cm−3 were used for the calculation of Seff,max.

We have conducted positive charging on both sides of all samples in an attempt to derive the fixed charge density (Qf). Figure B.2  shows the effect of the cumulative induced positive charges on the Seff values of the samples. However, the fixed charge density could not be reliably determined for these samples due to a minimal change in passivation quality after charge deposition combined with leaky behavior. As a result, no increase in Seff,max is observed with increasing cumulative corona charging time. For samples with ITO, no significant changes in Seff,max are observed throughout the experiment, which can be understood as mirror charges that appear in the top part of the ITO layer upon corona charge deposition such that the silicon surface is effectively shielded to the point that no change in the field effect passivation can be realized. For samples without ITO, a more pronounced degradation of Seff can be observed by which the Seff,max is determined after 600 s. Nevertheless, the amplitude of the curve is rather small resulting in a high uncertainty.

thumbnail Fig. B.2

Seff against cumulative charge deposition time for MoOx samples.

To estimate the amount of deposited corona charge up until the maximum in Seff,max is reached, the increase in the Kelvin probe voltage VKP over time is linearly fitted, as is shown in Figure B.3. Although the resulting fit does not clearly follow a linear trend, this kind of approximation is anyway used to estimate the change in VKP that is required for the evaluation of fixed charged density (Qf). The poor quality of the linear fit further illustrates that the error on the Qf value that will be calculated below should only be taken as a lower limit. Furthermore, it is interesting how the slope of the fitted line corresponds to the approximate slope corresponding to the samples with ITO/AlOx capping, while beyond 600 s the uncapped MoOx layer is not able to retain any additional charge due to leaky behavior. Following the plotted fitting approach, the total amount of deposited corona charge is estimated from the difference in VKP between the value at the start of the experiment and the value after 600 s of charging using the slope of the fitted line: ΔVKP = 7.075  ·  10−4 * 600 = 0.424 V. Note here that 600 s is the point in the experiment where the maximum in Seff,max is reached that can in turn be used to calculate Qf, as follows:

Qcorona=ϵrϵ0d.VKPe(A1)

thumbnail Fig. B.3

VKP as a function of cumulative corona charging time for MoOx contacts with different interlayers.

where ε0 is the vacuum permittivity (8.854 × 10−12 Fm−1), εr is the relative permittivity of MoOx (18), d is the thickness of the layer stack (5 nm), and e is the elementary charge. Since the deposited corona charge counteracts the fixed charge that is initially present in the layer stack, it holds that Qf = –2.0 ± 1.0 × 1012 cm−2. This moderately negative fixed charge could be associated with traps in the MoOx layer that are being filled in the corona charging experiment. If this is correct, other variations in the MoOx layer properties, such as what is induced by different growth temperatures, and their possible impact on Qf could become detectable by further corona charging experiments.

References

  1. H. Lin, M. Yang, X. Ru, G. Wang, S. Yin, F. Peng, C. Hong, M. Qu, J. Lu, L. Fang, C. Han, P. Porcel, O. Isabella, P. Gao, Z. Li, X. Xu, Silicon heterojunction solar cells with up to 26.81% efficiency achieved by electrically optimized nanocrystalline-silicon hole contact layers, Nat. Energy 8, 789 (2023) [CrossRef] [Google Scholar]
  2. JinkoSolar Holding Co., Ltd, JinkoSolar's High-efficiency N-Type Monocrystalline Silicon Solar Cell Sets New Record with Maximum Conversion Efficiency of 26.89% [Press release]. Available at https://www.prnewswire.com/news-releases/jinkosolars-high-efficiency-n-type-monocrystalline-silicon-solar-cell-sets-new-record-with-maximum-conversion-efficiency-of-26-89- 3019 71256.html (access: 13.02.2024) [Google Scholar]
  3. A. Richter, R. Müller, J. Benick, F. Feldmann, B. Steinhauser, C. Reichel, A. Fell, M. Bivour, M. Hermle, S.W. Glunz, Design rules for high-efficiency both-sides-contacted silicon solar cells with balanced charge carrier transport and recombination losses, Nat. Energy 6, 429 (2021) [CrossRef] [Google Scholar]
  4. J. Peter Seif, A. Descoeudres, M. Filipič, F. Smole, M. Topič, Z. Charles Holman, S. De Wolf, C. Ballif, Amorphous silicon oxide window layers for high-efficiency silicon heterojunction solar cells, J. Appl. Phys. 115, 024502 (2014) [Google Scholar]
  5. Z.C. Holman, A. Descoeudres, L. Barraud, F.Z. Fernandez, J.P. Seif, S. De Wolf, C. Ballif, Current losses at the front of silicon heterojunction solar cells, IEEE J. Photovolt. 2, 7 (2012) [CrossRef] [Google Scholar]
  6. J. Melskens, B.W.H. Van De Loo, B. Macco, L.E. Black, S. Smit, W.M.M. Kessels, Passivating contacts for crystalline silicon solar cells: from concepts and materials to prospects, IEEE J. Photovolt. 8, 373 (2018) [CrossRef] [Google Scholar]
  7. L.E. Black, B.W.H. van de Loo, B. Macco, J. Melskens, W.J.H. Berghuis, W.M.M. Kessels, Explorative studies of novel silicon surface passivation materials: considerations and lessons learned, Sol. Energy Mater. Sol. Cells 188, 182 (2018) [CrossRef] [Google Scholar]
  8. C. Battaglia, X. Yin, M. Zheng, I.D. Sharp, T. Chen, S. Mcdonnell, A. Azcatl, C. Carraro, B. Ma, R. Maboudian, R.M. Wallace, A. Javey, Hole selective MoOx contact for silicon solar cells, Nano Lett. 14, 967 (2014) [CrossRef] [PubMed] [Google Scholar]
  9. C. Battaglia, S.M. De Nicolás, S. De Wolf, X. Yin, M. Zheng, C. Ballif, A. Javey, Silicon heterojunction solar cell with passivated hole selective MoOx contact, Appl. Phys. Lett. 104, 1 (2014) [Google Scholar]
  10. J. Bullock, D. Yan, A. Cuevas, Y. Wan, C. Samundsett, n- and p-typesilicon solar cells with molybdenum oxide hole contacts, Energy Proc. 77, 446 (2015) [CrossRef] [Google Scholar]
  11. J. Bullock, M. Hettick, J. Geissbühler, A.J. Ong, T. Allen, C.M. Sutter-Fella, T. Chen, H. Ota, E.W. Schaler, S. De Wolf, C. Ballif, A. Cuevas, A. Javey, Efficient silicon solar cells with dopant-free asymmetric heterocontacts, Nat. Energy 1, 1 (2016) [Google Scholar]
  12. L. Cao, P. Procel, A. Alcañiz, J. Yan, F. Tichelaar, E. Özkol, Y. Zhao, C. Han, G. Yang, Z. Yao, M. Zeman, R. Santbergen, L. Mazzarella, O. Isabella, Achieving 23.83% conversion efficiency in silicon heterojunction solar cell with ultra-thin MoOx hole collector layer via tailoring (i) a-Si:H/MoOx interface, Prog. Photovolt.: Res. Appl. 31, 1245 (2022) [Google Scholar]
  13. J. Geissbühler, J. Werner, S. Martin De Nicolas, L. Barraud, A. Hessler-Wyser, M. Despeisse, S. Nicolay, A. Tomasi, B. Niesen, S. De Wolf, C. Ballif, 22.5% efficient silicon heterojunction solar cell with molybdenum oxide hole collector, Appl. Phys. Lett. 107, 081601 (2015) [Google Scholar]
  14. L. Neusel, M. Bivour, M. Hermle, Selectivity issues of MoOx-based hole contacts, Energy Proc. 124, 425 (2017) [CrossRef] [Google Scholar]
  15. S.W. Glunz, F. Feldmann, A. Richter, M. Bivour, C. Reichel, H. Steinkemper, J. Benick, M. Hermle, The irresistible charm of a simple current flow pattern-25% with a solar cell featuring a full-area back contact, in Proceedings of the 31st European Photovoltaic Solar Energy Conference and Exhibition (München, Germany, 2015), p. 259 [Google Scholar]
  16. M. Boccard, X. Yang, K. Weber, Z.C. Holman, Passivation and carrier selectivity of TiO2 contacts combined with different passivation layers and electrodes for silicon solar cells, in Conference Record of the IEEE Photovoltaic Specialists Conference (2016), pp. 2403–2407 [Google Scholar]
  17. T. Kamioka, Y. Hayashi, Y. Isogai, K. Nakamura, Y. Ohshita, Effects of annealing temperature on workfunction of MoOx at MoOx/SiO2 interface and process-induced damage in indium tin oxide/MoOx/SiOx/Si stack, Jpn. J. Appl. Phys. 57, 076501 (2018) [CrossRef] [Google Scholar]
  18. J. Tong, T.T. Le, W. Liang, M.A. Hossain, K.R. McIntosh, P. Narangari, S. Armand, T.C. Kho, K.T. Khoo, Y. Zakaria, A.A. Abdallah, S. Surve, M. Ernst, B. Hoex, K.C. Fong, Impact of pregrown SiOx on the carrier selectivity and thermal stability of molybdenum-oxide-passivated contact for Si solar cells, ACS Appl. Mater. Interfaces 13, 36426 (2021) [Google Scholar]
  19. M. Bivour, B. Macco, J. Temmler, W.M.M. Kessels, M. Hermle, Atomic layer deposited molybdenum oxide for the hole-selective contact of silicon solar cells, Energy Proc. 92, 443 (2016) [CrossRef] [Google Scholar]
  20. B.E. Davis, N.C. Strandwitz, Aluminum oxide passivating tunneling interlayers for molybdenum oxide hole-selective contacts, IEEE J. Photovolt. 10, 722 (2020) [CrossRef] [Google Scholar]
  21. M.T.S.K. Ah Sen, P. Bronsveld, A. Weeber, Thermally stable MoOx hole selective contact with Al2O3 interlayer for industrial size silicon solar cells, Sol. Energy Mater. Sol. Cells 230, 111139 (2021) [CrossRef] [Google Scholar]
  22. V. Naumann, M. Otto, R.B. Wehrspohn, C. Hagendorf, Chemical and structural study of electrically passivating Al2O3/Si interfaces prepared by atomic layer deposition, J. Vacuum Sci. Technol. A 30, 04D106 (2012) [CrossRef] [Google Scholar]
  23. O. Renault, L.G. Gosset, D. Rouchon, A. Ermolieff, Angle-resolved x-ray photoelectron spectroscopy of ultrathin Al2O3 films grown by atomic layer deposition, J. Vac. Sci. Technol. A 20, 1867 (2002) [CrossRef] [Google Scholar]
  24. A. Sarkar, Device simulation using Silvaco ATLAS tool, Technology Computer Aided Design (CRC Press, Boca Raton, Florida, USA, 2018), p. 203–252 [Google Scholar]
  25. G.J.M. Janssen, M.T.S.K. Ah Sen, P.C.P. Bronsveld, A simplified model to simulate passivating & selective hole-collecting contacts, in Proceeding 36th European Photovoltaic Solar Energy Conference and Exhibition (Marseille, France, 2019), pp. 2–8 [Google Scholar]
  26. U. Wurfel, A. Cuevas, P. Wurfel, Charge carrier separation in solar cells, IEEE J. Photovolt. 5, 461 (2015) [CrossRef] [Google Scholar]
  27. G. Janssen, M. Stodolny, I. Romijn, B. Geerligs, The role of the oxide in the carrier selectivity of metal/poly-Si/oxide contacts to silicon wafers, in Proceeding 33rd European Photovoltaic Solar Energy Conference and Exhibition (Amsterdam, The Netherlands, 2017), pp. 256–261 [Google Scholar]
  28. F. Feldmann, G. Nogay, J.I. Polzin, B. Steinhauser, A. Richter, A. Fell, C. Schmiga, M. Hermle, S.W. Glunz, A study on the charge carrier transport of passivating contacts, IEEE J. Photovolt. 8, 1503 (2018) [CrossRef] [Google Scholar]
  29. D.B.M. Klaassen, A unified mobility model for device simulation—I. Model equations and concentration dependence, Solid-State Electron. 35, 953 (1992) [Google Scholar]
  30. A. Richter, S. Glunz, F. Werner, J. Schmidt, Improved quantitative description of Auger recombination in crystalline silicon, Phys. Rev. B 86, 165202 (2012) [CrossRef] [Google Scholar]
  31. L.G. Gerling, C. Voz, R. Alcubilla, J. Puigdollers, Origin of passivation in hole-selective transition metal oxides for crystalline silicon heterojunction solar cells, J. Mater. Res. 32, 260 (2017) [CrossRef] [Google Scholar]
  32. B.B. Demaurex, S. De Wolf, A. Descoeudres, Z. Charles Holman, C. Ballif, Damage at hydrogenated amorphous/crystalline silicon interfaces by indium tin oxide overlayer sputtering, Appl. Phys. Lett. 101, 171604 (2012) [Google Scholar]
  33. M.T.S.K. Ah Sen, A. Mewe, J. Melskens, J. Bolding, M. Van de Poll, A. Weeber, Soft deposition of TCOs by pulsed laser for high-quality ultra-thin poly-Si passivating contacts, J. Appl. Phys. 134, 154502 (2023) [Google Scholar]
  34. M. Bivour, C. Messmer, L. Neusel, F. Zähringer, J. Schön, S.W. Glunz, M. Hermle, Principles of carrier-selective contacts based on induced junctions, in Proceeding 33rd European PV Solar Energy Conference and Exhibition (Amsterdam, The Netherlands, 2017), pp. 25–29 [Google Scholar]
  35. S. Essig, J. Dréon, E. Rucavado, M. Mews, T. Koida, M. Boccard, J. Werner, J. Geissbühler, P. Löper, M. Morales-Masis, L. Korte, S. De Wolf, C. Balllif, Toward annealing-stable molybdenum-oxide-based hole-selective contacts for silicon photovoltaics, Solar RRL 2, 1700227 (2018) [CrossRef] [Google Scholar]
  36. J.P. Seif, D. Menda, A. Descoeudres, L. Barraud, O. Özdemir, C. Ballif, , S. De Wolf, Asymmetric band offsets in silicon heterojunction solar cells: Impact on device performance, J. Appl. Phys. 120, 054501 (2016) [Google Scholar]
  37. A. Onno, C. Chen, Z.C. Holman, Electron and hole partial specific resistances: A framework to understand contacts to solar cells, in Conference Record of the IEEE Photovoltaic Specialists Conference (2019), pp. 2329–2333 [Google Scholar]
  38. D. Sacchetto, Q. Jeangros, G. Christmann, L. Barraud, A. Descoeudres, J. Geissbuhler, M. Despeisse, A. Hessler-Wyser, S. Nicolay, C. Ballif, ITO/MoOx/a-Si:H(i) hole-selective contacts for silicon heterojunction solar cells: degradation mechanisms and cell integration, IEEE J. Photovolt. 7, 1584 (2017) [CrossRef] [Google Scholar]
  39. J. Cho, N. Nawal, A. Hadipour, M. Recaman Payo, A. van der Heide, H.S. Radhakrishnan, M. Debucquoy, I. Gordon, J. Szlufcik, J. Poortmans, Interface analysis and intrinsic thermal stability of MoOx based hole-selective contacts for silicon heterojunction solar cells, Sol. Energy Mater. Sol. Cells 201, 110074 (2019) [CrossRef] [Google Scholar]
  40. C. Messmer, M. Bivour, J. Schön, S.W. Glunz, M. Hermle, J. Schon, S.W. Glunz, M. Hermle, Numerical simulation of silicon heterojunction solar cells featuring metal oxides as carrier-selective contacts, IEEE J. Photovolt. 8, 456 (2018) [CrossRef] [Google Scholar]
  41. S.M. Sze, K.K. Ng, Physics of Semiconductor Devices, 3rd edn. (Wiley-Interscience, 2006) [Google Scholar]
  42. R.T. Tung, Recent advances in Schottky barrier concepts, Mater. Sci. Eng. R Rep. 35, 1 (2001) [CrossRef] [Google Scholar]
  43. M. Liebhaber, M. Mews, L. Korte, T.F. Schulze, B. Rech, K. Lips, Valence band offset and hole transport across a-SiOx (0<x<2) passivation layers in silicon heterojunction solar cells, in 31st European Photovoltaic Solar Energy Conference and Exhibition (i), (2015) pp. 770–775 [Google Scholar]
  44. J. Bullock, A. Cuevas, T. Allen, C. Battaglia, Molybdenum oxide MoOx: a versatile hole contact for silicon solar cells, Appl. Phys. Lett. 105, 232109 (2014) [Google Scholar]
  45. R. Peibst, U. Römer, Y. Larionova, M. Rienäcker, A. Merkle, N. Folchert, S. Reiter, M. Turcu, B. Min, J. Krügener, D. Tetzlaff, E. Bugiel, T. Wietler, R. Brendel, Working principle of carrier selective poly-Si/c-Si junctions: is tunnelling the whole story? Sol. Energy Mater. Sol. Cells 158, 60 (2016) [CrossRef] [Google Scholar]
  46. J. Cho, N. Nawal, A. Hadipour, M. Recaman Payo, A. van der Heide, H.S. Radhakrishnan, M. Debucquoy, I. Gordon, J. Szlufcik, J. Poortmans, Interface analysis and intrinsic thermal stability of MoOx based hole-selective contacts for silicon heterojunction solar cells, Sol. Energy Mater. Sol. Cells 201, 110074 (2019) [CrossRef] [Google Scholar]
  47. B.W.H. van de Loo, B. Macco, M. Schnabel, M.K. Stodolny, A.A. Mewe, D.L. Young, W. Nemeth, P. Stradins, W.M.M. Kessels, On the hydrogenation of Poly-Si passivating contacts by Al2O3 and SiNx thin films, Sol. Energy Mater Sol. Cells 215, 110592 (2020) [Google Scholar]
  48. F. Feldmann, M. Simon, M. Bivour, C. Reichel, M. Hermle, S.W. Glunz, Solar energy materials & solar cells efficient carrier-selective p- and n-contacts for Si solar cells, Sol. Energy Mater. Sol. Cells 131, 100 (2014) [CrossRef] [Google Scholar]
  49. M.T.Greiner, L. Chai, M.G. Helander, W.M. Tang, Z.H. Lu, Metal/metal-oxide interfaces: How metal contacts affect the work function and band structure of MoO3, Adv. Funct. Mater. 23, 215 (2013) [Google Scholar]
  50. W.J.H. Berghuis, M. Helmes, J. Melskens, R.J. Theeuwes, W.M.M. Kessels, B. Macco, Extracting surface recombination parameters of germanium-dielectric interfaces by corona-lifetime experiments, J. Appl. Phys. 131, 195301 (2022) [CrossRef] [Google Scholar]
  51. W.J.H. Berghuis, J. Melskens, B. Macco, R.J. Theeuwes, L.E. Black, M.A. Verheijen, W.M.M. Kessels, Excellent surface passivation of germanium by a-Si: H/Al2O3 stacks, J. Appl. Phys. 130, 135303 (2021) [CrossRef] [Google Scholar]
  52. J. Melskens, R.J. Theeuwes, L.E. Black, W.J.H. Berghuis, B. Macco, P.C.P. Bronsveld, W.M.M. Kessels, Excellent passivation of n-type silicon surfaces enabled by pulsed-flow plasma-enhanced chemical vapor deposition of phosphorus oxide capped by aluminum oxide, Phys. Stat. Solidi 15, 2000399 (2021) [Google Scholar]
  53. W.J.H. Berghuis, J. Melskens, B. Macco, R.J. Theeuwes, M.A. Verheijen, W.M.M. Kessels, Surface passivation of germanium by atomic layer deposited Al2O3 nanolayers, J. Mater. Res. 36, 571 (2021) [CrossRef] [Google Scholar]

Cite this article as: Mike Tang Soo Kiong Ah Sen, Gaby Janssen, Agnes Mewe, Paula Bronsveld, Jimmy Melskens, Fatemeh Hashemi, Paul Procel-Moya, Arthur Weeber, Influence of passivating interlayers on the carrier selectivity of MoOx contacts for c-Si solar cells, EPJ Photovoltaics 15, 34 (2024)

All Tables

Table 1

Atlas simulation parameters for the different layers.

Table 2

Measured IV characteristics for solar cells with different interlayers and without an interlayer.

All Figures

thumbnail Fig. 1

Process flow for the manufacturing of the solar cells with different interlayers (left) and a schematic of the solar cells (right).

In the text
thumbnail Fig. 2

Solar cell precursors with front MoOx and rear poly-Si(n+) contacts for iVoc and SunsVoc measurements.

In the text
thumbnail Fig. 3

Schematic of the cell structure used in Atlas to simulate the contact selectivity of the cell with varying interlayer µh and hole WF.

In the text
thumbnail Fig. 4

iVoc of half-fabricates after deposition of MoOx and ITO layers and subsequent annealing at 190 °C. Note that the order of the interlayer configurations (no interlayer, Al2O3/SiOy, a-Si:H(i), and SiO2) is consistent across all figures.

In the text
thumbnail Fig. 5

ΔVoc behavior of different MoOx contacts as a function of cumulative annealing (190–250 °C).

In the text
thumbnail Fig. 6

Influence of passivating a-Si:H(i), Al2O3/SiOy, and SiO2 interlayers on the light IV characteristics of the solar cells with MoOx contacts and different interlayers (right).

In the text
thumbnail Fig. 7

Dark JV measurements were used to extract Rs as a function of cell temperature (25–65 °C). The fitted lines are used to calculate the activation energy.

In the text
thumbnail Fig. 8

Simulated ΔVoc as a function of Seff for the hole contact shown for WF varying from 5.09 to 5.33 eV and for µh of (a) 10−5 cm2V−1s−1 and (b) 10−7 cm2V−1s−1.

In the text
thumbnail Fig. 9

Simulated effect of varying µh (10−2–10−7cm2V−1s−1) and hole WF contact (4.92–5.34 eV) (a) on ΔVoc, (b) and on FF. Seff is set to 33 cm/s which is representative of the surface passivation quality of Al2O3/SiOy and SiO2 interlayers with a MoOx contact. Electron mobility of the interlayer is set to 10−5 cm2V−1s−1.

In the text
thumbnail Fig. 10

Simulated JV curves, and recombination currents at the p- and n- contact, and in the absorber (bulk) as a function of cell voltage, for µh of (a) 10−5 and (b) 10−7 cm2V−1s−1.

In the text
thumbnail Fig. 11

Schematic band diagrams of hole contacts with Al2O3/SiOy (left) and SiO2 (right) interlayers, illustrating the contrasting energy barriers and hole and electron quasi-Fermi levels (EFp,n) alignment.

In the text
thumbnail Fig. A.1

Comparing the surface passivation quality between passivated AlOx reference and x-ray exposed samples.

In the text
thumbnail Fig. B.1

Symmetric samples consisting of AlOx, SiO2, and a-Si:H(i) interlayers, capped with MoOx, ITO, and AlOx.

In the text
thumbnail Fig. B.2

Seff against cumulative charge deposition time for MoOx samples.

In the text
thumbnail Fig. B.3

VKP as a function of cumulative corona charging time for MoOx contacts with different interlayers.

In the text

Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.

Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.

Initial download of the metrics may take a while.