Silicon surface passivation by PEDOT: PSS

Nanotechnology

Related content

PAPER

Silicon surface passivation by PEDOT: PSS functionalized by SnO2 and TiO2 nanoparticles

- Planar n-Si/PEDOT:PSS hybrid heterojunction solar cells utilizing functionalized carbon nanoparticles synthesized via simple pyrolysis route Yoon-Ho Nam, Dong-Hyung Kim, Sambhaji S Shinde et al.

To cite this article: M García-Tecedor et al 2018 Nanotechnology 29 035401

- Surface passivation and antireflectance performances for atomic-layer-deposited Al2O3 films Li Qiang Zhu, Hui Xiao and Yang Hui Liu

View the article online for updates and enhancements.

- One dimensional Si/Ge nanowires and their heterostructures for multifunctional applications—a review Samit K Ray, Ajit K Katiyar and Arup K Raychaudhuri

This content was downloaded from IP address 158.227.173.83 on 14/12/2017 at 09:26

Nanotechnology Nanotechnology 29 (2018) 035401 (10pp)

https://doi.org/10.1088/1361-6528/aa9c9e

Silicon surface passivation by PEDOT: PSS functionalized by SnO2 and TiO2 nanoparticles M García-Tecedor1,2,4 , S Zh Karazhanov2, G C Vásquez1,2, H Haug2, D Maestre1, A Cremades1, M Taeño1,3, J Ramírez-Castellanos3 , J M González-Calbet3, J Piqueras1, C C You2 and E S Marstein2 1

Departamento de Física de Materiales, Facultad de CC. Físicas, Universidad Complutense, 28040, Madrid, Spain 2 Department for Solar Energy, Institute for Energy Technology (IFE), PO BOX 40, 2027, Kjeller, Norway 3 Departamento de Química Inorgánica I, Facultad de CC. Químicas, Universidad Complutense, 28040, Madrid, Spain E-mail: [email protected] Received 4 July 2017, revised 21 November 2017 Accepted for publication 23 November 2017 Published 13 December 2017 Abstract

In this paper, we present a study of silicon surface passivation based on the use of spin-coated hybrid composite layers. We investigate both undoped poly(3,4-ethylenedioxythiophene)/poly(styrenesulfonate) (PEDOT:PSS), as well as PEDOT:PSS functionalized with semiconducting oxide nanomaterials (TiO2 and SnO2). The hybrid compound was deposited at room temperature by spin coating—a potentially lower cost, lower processing time and higher throughput alternative compared with the commonly used vacuum-based techniques. Photoluminescence imaging was used to characterize the electronic properties of the Si/PEDOT:PSS interface. Good surface passivation was achieved by PEDOT:PSS functionalized by semiconducting oxides. We show that control of the concentration of semiconducting oxide nanoparticles in the polymer is crucial in determining the passivation performance. A charge carrier lifetime of about 275 μs has been achieved when using SnO2 nanoparticles at a concentration of 0.5 wt.% as a filler in the composite film. X-ray diffraction (XRD), scanning electron microscopy, high resolution transmission electron microscopy (HRTEM), energy dispersive x-ray in an SEM, and μ-Raman spectroscopy have been used for the morphological, chemical and structural characterization. Finally, a simple model of a photovoltaic device based on PEDOT:PSS functionalized with semiconducting oxide nanoparticles has been fabricated and electrically characterized. Supplementary material for this article is available online Keywords: silicon surface passivation, PEDOT:PSS, functionalization by SnO2 or TiO2, hybrid materials (Some figures may appear in colour only in the online journal) 1. Introduction

research, such as photovoltaic technology, due to their ease of processing, low cost, their ability to be used in flexible substrates and scalability [1, 2]. However, the use of organic materials also involves some disadvantages which can be overcome by their combination with inorganic materials [3], thus achieving better stability, electrical behaviour and tunability of optical and electrical properties. As an example,

The use of organic materials has recently enabled breakthroughs to be achieved in different fields of technological 4

Present address: Institute of Advanced Materials (INAM), Universitat Jaume I, 12006 Castelló, Spain.

0957-4484/18/035401+10$33.00

1

© 2017 IOP Publishing Ltd Printed in the UK

Nanotechnology 29 (2018) 035401

M García-Tecedor et al

hybrid solar cells have gained increasing interest as emerging low cost photovoltaic devices with potential competitive efficiency [4–6]. One of the most important and challenging problems that this technology faces is the understanding and control of the interface between the silicon (Si) substrate and the organic material. Among other factors, an improvement of the interface between components is crucial in the optimization of the performance of microelectronic, photovoltaic and sensing devices. It is therefore an important challenge to develop efficient processes and materials, leading to an improvement of the Si surface passivation without additional large costs. At present, inorganic materials such as hydrogenated Si nitride (SiNx:H) [7], amorphous Si (a-Si) [8], aluminum oxide (Al2O3) [9] and silicon dioxide (SiO2) [10], are commonly employed for Si passivation, providing high minority carrier lifetime values of several milliseconds. However, these materials are processed either using vacuumbased techniques or at high temperatures, which increases the operation costs and could lead to process damage (see, e.g. [7–10]). Therefore, developing low temperature passivation methods and materials, involving simple procedure and low costs—as those proposed in this work—could be highly beneficial. Polymers have the potential to satisfy this requirement, as they can be deposited at room temperature, and, furthermore, upon processing at low temperatures, they can retain their main functionalities (see, e.g. [11]). The quality of the Si surface passivation strongly depends on the type of polymer used for the deposition on Si. The most commonly used conductive polymer so far is poly(3,4-ethylenedioxythiophene)/poly-(styrenesulfonate) (PEDOT:PSS), which has a high p-type electrical conductivity, as well as good chemical stability and optical transparency in the visible range. An added advantage is that PEDOT:PSS can be easily processed in aqueous solution [12–14]. In the present paper, we investigate the surface passivation of Si substrates by using bare PEDOT:PSS and PEDOT: PSS functionalized with either titanium oxide (TiO2) or tin oxide (SnO2) nanoparticles. By including optimal concentrations of the nanomaterials, we obtain good Si surface passivation, as evidenced by charge carrier lifetime values of up to ∼275 μs. This is a promising and competitive value considering that the employed method can still be further optimized. In addition, the employed spin-coating process allows for good homogeneity of the deposited layers by a fast and low cost method, which avoids vacuum-based techniques and complex pre-treatment of the Si surface.

Ti(OBu)4 for SnO2 and TiO2 nanoparticles, respectively) which was dissolved in distilled water. Then, approximately 100 ml of distilled water was added and placed in a heating plate, with constant stirring, keeping the temperature between 100 °C and 110 °C. The desired product was only obtained in basic medium; therefore, to neutralize the acid solution due to the presence of chlorides, a neutralizing agent, Li2CO3 in our case, was required. Once the desired pH was obtained, the reaction was left until the hydrolysis ceasesd. Then, (100)-oriented Czochralski Si wafers of n-type electrical conductivity with a nominal resistivity between 1–3 (Ωcm) and a thickness of 300 μm were used as substrates in this work. A 40 nm thick layer of hydrogenated amorphous Si (a-Si:H) was deposited by plasma-enhanced chemical vapor deposition (PECVD) on the back side of the wafers as a reference passivation layer. A thin layer of either bare PEDOT:PSS or PEDOT:PSS functionalized with rutile SnO2 or TiO2 nanoparticles was deposited on top of the Si substrates by spin coating. Ultrasonication of the dispersion was performed before the spin coating in order to achieve a good dispersion of the nanomaterials in the PEDOT:PSS solution. Ethylene glycol (EG) was used to enhance the dispersion of the nanoparticles and avoid aggregation during the spin coating, which results in a higher homogeneity of the deposited films. The use of EG could also improve the electrical conductivity of the polymer due to alignment of the polymer chains [15, 16]. However, when no EG was added to the dispersion, no significant variations in the passivation behaviour were measured, while good homogeneity was still reached in the spin-coated layers. After spin coating, a thermal annealing was carried out on a hotplate at 120 °C for 20 min to evaporate the water from the PEDOT:PSS. The thickness of the films, 120 nm on average, was measured using an Alpha-Step profilometer. X-ray diffraction (XRD) measurements were performed with a Philips X’Pert Pro diffractometer using Cu Kα radiation. High resolution transmission microscopy (HRTEM) was performed using a JEOL 3000 FEG electron microscope. For the HRTEM measurements, the SnO2 or TiO2 nanopowders were dispersed in isopropanol and deposited on TEM grids. Absorption spectra were measured for bare PEDOT:PSS and hybrid PEDOT:PSS/nanoparticles deposited on glass substrates using a UV–vis-NIR spectrophotometer (Ocean Optics QE65000). A photoluminescence (PL) imaging system was used to measure the surface recombination at the Si surface. The effective charge carrier lifetime values were calculated from the PL intensity based on the quasi-steady state photoconductance (QSS-PC) measurements [17–19]. Measurements were carried out using a LIS-R1 PL imaging setup from BT Imaging with an excitation wavelength of 808 nm and a constant illumination intensity of 4.2×10−2 W cm−2. Optical micrographs were acquired with a Leica DFC295 optical microscope. Compositional analysis by energy dispersive spectroscopy (EDS) was carried out using a Bruker AXS Quantax system working at 15 kV and 1.5 nA in a Leica 440 Stereoscan SEM. μ-Raman spectroscopy was performed with a confocal microscope (Horiba Jobin Yvon LabRAM HR 800) using an He–Ne red laser with a wavelength of 633 nm.

2. Methods An aqueous PEDOT:PSS dispersion at 1, 3% v/v (SigmaAldrich), presenting a sheet resistance below 100 (Ω sq−1) and p-type conductivities up to σ=1000 S cm−1 was used as the source of a conductive polymer in the composite layer. The SnO2 and TiO2 nanoparticles used as a filler in the hybrid composite were synthesized by a soft chemistry route based on the hydrolysis of a certain precursor (SnCl2·2 H2O and 2



Figure 1. (a) XRD of SnO2 and A-TiO2 nanoparticles, (b) HRTEM image and electron diffraction patterns of SnO2 nanoparticles. (c), (d) FFT patterns along [101] and [111] respectively, corresponding to SnO2 nanoparticles.

llum. from PEDOT side

Illum. from a-Si side 50

PEDOT:PSS

25 22 20 18 16 14 12 10 8 6 4 2 0

40 30 20 10 0

Figure 2. PL image of passivated Si wafer illuminated from (a) PEDOT:PSS and (b) a-Si:H sides. The bars at the right side of each image

represent a colour scale corresponding to carrier lifetime values in μs.

For the characterization of the photovoltaic device, a WACOM solar simulator with a short arc Xe lamp was used in order to simulate standard AM1.5 solar conditions. I–V measurements were carried out with a Keithley 4200-SCS.

nanoparticles was 50 nm, as estimated by the Scherrer formula. The dimensions and crystallinity of the nanoparticles were further analyzed by TEM. Figure 1(b) displays the HRTEM micrograph for SnO2 nanoparticles, as an example, that confirms the high crystallinity of the as-grown nanoparticles, as well as their homogeneity and reduced dimensions. In this case, interplanar distances of 3.35 y 4.74 Å, corresponding to (110) and (001) planes in rutile SnO2, respectively, are indicated in the image. FFT patterns along the [101] and [111] axes are also marked in figures 1(c) and (d), respectively.

3. Results and discussion 3.1. Inorganic nanomaterials functionalizing the PEDOT:PSS

Before the analysis of the passivation properties of the composites, a study of the nanoparticles of the inorganic materials used as fillers was carried out. XRD measurements (figure 1(a)) confirm that the nanoparticles consist of a cassiterite SnO2 and TiO2 anatase phase (hereinafter A-TiO2). According to the Scherrer formula, dimensions of about 5–8 nm have been estimated for the nanoparticles. By thermal treatment at 1000 °C for 24 h, the powders in the anatase phase were transformed into a rutile phase (hereinafter R-TiO2). As a consequence of the thermal treatment, grain growth was induced and the average size of the R-TiO2

3.2. Si surface passivation

Prior to the study of the hybrid composites, the passivation behavior of bare PEDOT:PSS was analyzed as a reference. Based on the PL imaging and QSS-PC, values of charge photocarrier lifetimes can be measured for Si substrates passivated by PEDOT:PSS (front side) and by a-Si:H (back side). Therefore, PL images have been acquired by illumination either from the PEDOT:PSS side or from the reference 3



Figure 3. PL images of a composite sample containing (a) SnO2 nanoparticles in 5 wt.% concentration, (b) A-TiO2 nanoparticles in 5 wt.%, and (c) R-TiO2 nanoparticles in a 5 wt.% obtained by illumination from the PEDOT:PSS or from the a-Si:H side. The corresponding QSSPCC measurements are also shown. The bars at the right side of each PL image represent a colour scale corresponding to carrier lifetime values in μs.

passivation layer at the back side, as shown in figures 2(a) and (b). Color bars beside the PL images indicate charge carrier lifetime values. Average lifetimes around 10 μs and 33 μs have been obtained by illumination by the front or rear side, respectively, thus indicating poor passivation behavior associated with bare PEDOT:PSS, as expected. These values are two orders of magnitude lower that those obtained when using a-Si:H at both front and rear sides (τ∼1 ms). Dispersions of PEDOT:PSS and SnO2 or TiO2 (both R-TiO2 and A-TiO2) nanoparticles in a concentration range from 0.25–5 wt% were prepared following the method described in [20, 21]. The obtained hybrid layers were deposited by spin coating on a Si substrate. Figures 3(a)–(c) show the PL images and injection-level dependences of carrier lifetimes for Si passivated by PEDOT:PSS/SnO2, PEDOT:PSS/A-TiO2, and PEDOT:PSS/R-TiO2 respectively, using an initial concentration of nanoparticles of 5% wt. The PL images were obtained for illumination from the hybrid composite and from the a-Si:H side (figure 3). Analysis shows that the average minority carrier lifetime for the functionalization of PEDOT:PSS with SnO2 nanoparticles

is below 200 μs at carrier injection levels of about Δn=1015 cm−3 (figure 3(a)), and clearly increased by around one order of magnitude as compared to the value of bare PEDOT:PSS. Functionalization with A-TiO2 nanoparticles provides a lower average lifetime of ∼30 μs (figure 3(b)), similar to those obtained for bare PEDOT:PSS. Contrary to the A-TiO2 nanoparticle composite case, the use of R-TiO2 nanoparticles significantly improves the average lifetime values (see figure 3(c)) although the values of the SnO2 composite are not reached. This shows that, regarding the good homogeneity of the films and high charge carrier lifetime values, layers of PEDOT:PSS with SnO2 or R-TiO2 nanoparticles provide better passivation performance of Si surface than bare PEDOT:PSS. Analogous measurements (not shown here) have been performed using the same composite layers on the top of p-type Si showing lower lifetime values and no effective passivation. Further analysis showed that the quality of the passivation mainly depends on the concentration of the nanoparticles in the initial dispersion, as confirmed by measurements of the 4



Figure 4. Carrier lifetime as a function of the concentration of (a) SnO2 nanoparticles, and (b) R-TiO2 nanoparticles.

charge carrier lifetime. The passivation properties were systematically measured for the composite films as a function of the concentration of nanoparticles (0.25, 0.5, 1, 3, and 5 wt% in the starting dispersion). Figures 4(a) and (b) show the dependence of the measured carrier lifetime on the concentration of SnO2 and R-TiO2 nanoparticles, respectively, which is not monotonic. The carrier lifetime, calculated by QSS-PC based on the PL intensity, increases with an increase in the concentration of SnO2 or R-TiO2 at smaller concentrations, and decreases at higher concentrations. Optimal concentrations of nanoparticles have been determined as 0.5 wt.% for SnO2, showing a maximum carrier lifetime of 275 μs, and 1.0 wt.% in the case of R-TiO2 with the largest carrier lifetime value of 160 μs. These results indicate that the initial concentration of nanoparticles is a crucial parameter to control in order to achieve improved passivation performance. Some other authors also reported a dependence of optical and electrical properties of the composite on the weight fraction of the inorganic nanoparticles on the polymeric matrix. Raccis et al [22] analyzed hybrid composites formed by PEDOT:PSS and α-Fe2O3 nanoparticles in a variable ratio for which the maximum conductivity values were achieved when using a 1% wt. ratio of nanoparticles. Agglomeration and percolation effects, which could be caused when using a high concentration of nanoparticles in the composite, would induce a worsening of the passivation properties of the composite. In the present work, the largest charge carrier lifetime achieved in this work by using SnO2 nanoparticles in the composite is smaller than typical values for commonly used passivation layers, such as τ=2.41 ms [23, 24] obtained by passivation with the stack a-SiOxNy:H/SiNx. However, that stack is deposited by PECVD whereas the deposition process used in this work is low cost, easy, fast, and there is still room for further enhancement of the lifetime. In order to achieve a deeper understanding of the improved passivation behavior achieved by using inorganic nanoparticles in the composite, the optical and compositional properties of the layers with the optimized concentration of SnO2 nanoparticles (0.5% wt) have been investigated and compared to the best results using R-TiO2 (1% wt.) and bare PEDOT:PSS. The films have been deposited by spin coating

Figure 5. Absorption spectra of bare PEDOT:PSS and composite

films with SnO2 (0.5% wt.) or R-TiO2 (1% wt.) deposited on glass. The absorption spectrum from glass is also shown for comparison.

on glass substrates for the absorbance measurements, therefore the absorption spectrum from the glass substrate is also included as a reference. All samples show a high level of transparency in the visible range, as expected. As observed in figure 5, where a comparison to the glass substrate is also shown, the optical absorption of the PEDOT:PSS and PEDOT:PSS/SnO2 or R-TiO2 films in the visible range is lower than 10%. Only high absorption is observed in the UV range for all the cases. Since SnO2 and R-TiO2 are wide band gap oxides, upon functionalization of PEDOT:PSS with these nanoparticles, the optical absorption properties of the hybrid material are not expected to significantly change, as shown in figure 5. Hence, the optical transparency to the visible characteristic of the PEDOT:PSS is only slightly modified by adding SnO2 or R-TiO2 nanoparticles, which ensures its suitability for use in devices requiring high transparency, while adding good passivation behavior. Using EDS measurements, the chemical analysis of the PEDOT:PSS/SnO2 or R-TiO2 nanoparticle films was performed, and carbon, Si, oxygen, sulphur and tin or titanium were detected, as observed in the EDS spectra of figure 6. Carbon and sulphur are related to PEDOT:PSS, oxygen is associated with both polymer and nanoparticles, whereas tin or titanium are due 5



Figure 6. EDS spectra acquired on layers formed by (a) PEDOT:PSS and SnO2 nanoparticles (5% wt.) and (b) PEDOT:PSS and R-TiO2

nanoparticles (5% wt.). (c) Raman spectra of PEDOT: PSS (black line), PEDOT: PSS with SnO2 nanoparticles 5 wt%. (blue line) and PEDOT: PSS with R-TiO2 5 wt% (red line). (d) Raman spectra for the samples in the region between 1000 and 1800 cm−1, where the characteristic PEDOT: PSS benzoic structure is located.

to the presence of the nanoparticles. The EDS map confirms a high spatial homogeneity of the spin-coated films (shown in figure S1, in the supporting information is available online at stacks.iop.org/NANO/29/035401/mmedia). By using Raman spectroscopy, the structural configuration has been studied to achieve a better understanding of the composites under study, as well as a fingerprint of the analyzed materials. In figure 6(c), the Raman spectra of bare PEDOT:PSS as well as PEDOT:PSS with SnO2 5% wt. or R-TiO2 5% wt. are shown. The characteristic benzoic structure from bare PEDOT:PSS can be appreciated in the corresponding spectrum, with the main band between 1400–1500 cm−1, associated with the stretching vibration of Cα=Cβ on the five-member ring of PEDOT [25]. In particular, contribution from C–C inter-ring stretching (1258 cm−1), single C–C stretching (1364 cm−1), C–C symmetrical stretching (1441 cm−1), C–C asymmetrical stretching (1510 cm−1) and C–C antisymmetrical stretching (1567 cm−1) can be identified in figure 6(d). In the case of the PEDOT:PSS/nanoparticle films, these bands are also observed, as well as the peaks corresponding to the rutile structure of the SnO2 nanoparticles (Eg=474 cm−1 and A1g=633 cm−1) or R-TiO2 nanoparticles (Eg=447 cm−1 and A1g=611 cm−1). The Eg mode corresponds to the movement of oxygen atoms along the c-axis, while the A1g mode is related to the extension movements of oxygen anions

with respect to tin cations, perpendicularly to the c-axis. In addition, the peak corresponding to crystalline Si (520 cm−1) from the substrate also appears in the Raman spectra. As observed in the Raman spectra of figure 6(d), the TiO2-based composite exhibits the main band of PEDOT:PSS around 1400–1500 cm−1, almost unchanged with respect to the bare PEDOT:PSS, whereas for the SnO2-based composite, this band is blurred out and shifted with respect to the bare PEDOT:PSS. Moreover, a shoulder at about 700 cm−1 can be observed in the spectrum from the PEDOT:PSS/SnO2 (figure 6(c)) which can be associated with the symmetric C–S–C deformation, although activation of the A2u mode (705 cm−1) in rutile SnO2 cannot be excluded. This demonstrates that the properties of the composite film are not merely the addition of the properties of the single counterparts; moreover in the case of the SnO2 composite, interactions between PEDOT:PSS and the SnO2 nanoparticles are revealed. In this work, we propose a possible physical passivation mechanism occurring in our hybrid system, although we do not discard its combination with chemical passivation contribution, which consists of the saturation of free bounds at the Si surface due to the SiO2 native layer on top of the Si wafers [10, 26], as will be described in section 3.3. This kind of physical passivation mechanism, also known as chargeassisted passivation or field-effect passivation, tries to reduce 6



the concentration of one type of charge carrier, usually by adding fix charges or using an extra doped layer [27]. In both cases, the objective is to restrict the arrival of the minority carriers to the surface and to promote a greater separation of carriers, hence reducing surface recombination. According to this mechanism, negative fixed charges would be required to passivate p-type Si, whereas positive charges are required in the n-type case. Metal oxide nanoparticles exhibit a charged surface in aqueous dispersions due to the complete coverage of the surface by hydroxyl groups [28]. The surface charge density depends on the nanoparticle dimensions, the properties of the surface, as well as the pH and molar concentration of the dispersion, as described in [29]. The PEDOT:PSS– nanoparticle aqueous dispersion used as a precursor in this study presents low pH values (pH