AFM Measurements of Novel Solar Cells

Studying Electronic Properties of Si-Based Radial Junctions

  • AFM Measurements of Novel Solar Cells - Studying Electronic Properties of Si-Based Radial JunctionsAFM Measurements of Novel Solar Cells - Studying Electronic Properties of Si-Based Radial Junctions
  • AFM Measurements of Novel Solar Cells - Studying Electronic Properties of Si-Based Radial Junctions
  • Fig. 1: Solar cell with radial junctions based on Si nanowires coated by intrinsic a Si:H absorber layer, thin doped a-Si:H and a top TCO. a) SEM micrograph of a complete cell (without the top TCO), b) scheme of the PF AFM measurements of the Si NWs based radial junction solar cell.
  • Fig. 2: Measurement of local conductivity via C-AFM using solid platinum tip (RMN-25PT300B) operating in the semi-contact PF TUNA mode on the sample of radial junctions on Si nanowires. a) Conductivity map of the 5x5 µm2 scan area with borders of radial junctions brought out. b) More detailed local conductivity signal in 3x3 µm2 area shown superposed on the height sensor signal in 3D image. Wires with different conductivity can be recognized.

In this paper, we demonstrate the use of atomic force microscopy (AFM) with a conductive cantilever to study local electronic properties of silicon ­nanostructures: p-i-n radial junctions of amorphous Si grown on Si nanowires. We have observed variations of the conductivity of the radial junction ­solar cells based on Si nanowires. Finally, we discuss possibilities of comparing the local photoresponse to local photovoltaic conversion efficiency.


In pursuit of a higher photovoltaic conversion efficiency, modern thin film solar cells utilize microscopic structures to improve light trapping abilities and photogenerated charge collection, thus increasing photovoltaic conversion efficiency. Among these, cells based on semiconductor nanowires (NWs) play an important role, since they offer a possibility of surpassing the Shockley-Queisser limit [1-4].

The cells are typically composed of a huge number of radial junctions layered around Si nanowires with sizes comparable to visible light wavelengths. Optical and electronic properties of individual junctions differ from each other, either as a natural effect of cell structure variations, or due to structural defects.

While the industry aims to improve the uniformity and efficiency of large-area thin film panels, research takes place at another frontier: measurement of conversion efficiency on a microscopic scale [5]. On this scale, even small fluctuations can significantly impact the resulting photovoltaic conversion efficiency.

Thus we need to measure the efficiency on the scale of a single structural element, e.g. a single nanowire-based radial junction.

Atomic force microscopy offers the resolution required for this task. In the past, local electronic properties of various photovoltaic materials were studied using conductive cantilever (C-AFM), as reported in [6,7]. Here we describe the use of the C-AFM on the Si NWs radial junction solar cells for the first time, employing a new PeakForce mode and full-metal platinum tips.

For further details, see our recent paper in Sol. Energy Mater.

Sol. Cells. [8].


Radial p-i-n junctions were obtained by depositing intrinsic (80 nm) and n-type (10 nm) layers of hydrogenated amorphous Si on p-type Si nanowire array resulting from vapour-liquid-solid growth [9], using Sn particles as a catalyst; see [2] for more details. The underlying substrate was glass coated by ZnO as a bottom transparent contact. The solar cells were completed by circular (∅ ~ 5 mm) ITO top contacts defined by a mask during sputter-deposition, leaving areas in between for C-AFM measurements. The conversion efficiency for this particular sample was ~6% (Voc = 0.708 V, Jsc = 13.719 mA/cm2, FF=61.95 %). An SEM image of the complete structure is shown in figure 1a).

Bruker Icon AFM was used in PeakForce (PF) mode to observe both the topography and local conductivity of the Si NWs sample using PF Tunnelling AFM (TUNA) [10] mode (see fig. 1b for illustration of the measurement). For the measurements of local conductivity we used bulk platinum cantilevers (RMN-25PT300B, spring constant 18N/m), of which we expected higher durability and a more stable conductivity map.


The local conductivity maps measured by PF TUNA on Si NWs samples are shown in figure 2. The PF setpoint was 0.1 V, corresponding to the applied force ~200 nN. Higher forces tend to deform the fragile Si NWs while scanning. Too low forces do not allow for good contact, leading to inconclusive conductivity maps. Stable contact during measurement can be achieved by using solid metal probes, in our case, platinum.

Figure 2a shows a conductivity map of a 5x5 µm2 scan area with borders of radial junctions accented for better orientation. Differences between the conductivity of individual NWs are clearly visible. Since negative voltage was applied on the sample, more conductive areas appear darker in the image. Figure 2b shows a conductivity map of a smaller area (3x3 µm2) superimposed on a 3D representation of the topography.


The results shown in figure 2 demonstrated the possibility of measuring the Si NWs-based radial junctions by PF TUNA. The maps of local conductivities shown in figure 2 were stable and reproducible, which was not the case for the C-AFM measurements in contact mode. In this case, the lateral forces acting on the radial junctions led to their bending even with very small contact force setpoints, smearing the local conductivity maps (not shown here).

We should note that more stable conductivity maps were obtained using the bulk Pt cantilevers, although the abrasion of the tip led to visibly worse lateral spatial resolution. Closer examination of the local conductivity maps shows increased currents at the edges of the radial junction heads. This is not expected of bulk Pt cantilevers, and further study is required for a proper explanation.

Differences in conductivity of individual NWs cannot be satisfactorily explained by structural variations, such as the tilting or crossing of NWs. For example, from a closer study of figure 2a we can see that there is no correlation between the conductivity and vertical/horizontal position of the NWs. Further probes into this problem are needed, since understanding the aspects affecting the conductivity of NWs can lead to the creation of more uniform nanostructured solar cells with higher efficiency.

The final goal of this type of experiments is to observe the variations of local photovoltaic conversion efficiency, but this requires two additional procedures. First of all, measurements of local photo I-Vs need to be performed at each scan point, or more simply, we can compare the local photocurrents measured by PF TUNA and local photovoltages measured by scanning probe Kelvin microscopy. We also need to take into account the variations of local light absorption. According to [11], it may be possible to map the local absorptance by using the photothermal AFM with an intermittent light to check the local heating of the individual radial junctions. Here the Si NWs-based solar cells offer an additional benefit of well-defined structural units, i.e. individual radial junctions from which we expect a more pronounced photothermal response than from solid thin films or bulk samples.


The recent progress in scanning-probe methods allows mapping the local electronic properties for arrays of radial junctions, based on Si nanowires with conformal coating of amorphous silicon absorber layer. We have demonstrated the first use of the PeakForce atomic force microscopy to map the local conductivity of the radial junction arrays, discovering differences in conductivity seemingly unexplainable by the NWs' structural disorder.

This research was partly supported by Czech Science Foundation projects 13-12386S and 13-25747S and the MSMT infrastructure project LM2011026.

[1] Kayes B. M. et al.: J. Appl. Phys. 97, 114302 (2005)
[2] Yu L. et al.: Nanotechnology 23, 194011 (2012)
[3] Krogstrup P. et al.: Nat. Photonics 7, 306-310 (2013)
[4] Wallentin J. et al.: Science 339, 1057-1060 (2013)
[5] Bonnell D. A. and Kalinin S. V.: Scanning Probe Microscopy for Energy Research, World Scientific (2013)
[6] Rezek B. et al.: J. Appl. Phys. 92, 587-593 (2002)
[7] Mates T. et al.: J. Non-Cryst. Solids. 352, 1011-1015 (2006)
[8] Fejfar A. et al.: Sol. Energy Mat. Sol. Cells 119, 228-234 (2013)
[9] Wagner R. S. et al.: Appl. Phys. Lett. 4, 89-90 (1964)
[10] Li. C. et al.: Bruker Appl. Note 132 (2011)
[11] Hara K. et al.: Appl. Phys. Express. 5, 022301 (2012)

Matěj Hývl, Bc
(corresponding author via e-mail request button below)
Antonín Fejfar
Soumyadeep Misra
Linwei Yu
Pere Roca i Cabarrocas
Institute of Physics
Czech Academy of Sciences
Thin Films and Nanostructures
Prague, Czech Republic


Czech Academy of Sciences
Cukrovamicka 10/112
16200 Praha 6
Czech Republic
Phone: +420 220 318 510

Register now!

The latest information directly via newsletter.

To prevent automated spam submissions leave this field empty.