- Email: [email protected]
Effect of sputtering process parameters on film properties of molybdenum back contact
Solar Energy Materials & Solar Cells 100 (2012) 1–5
Contents lists available at SciVerse ScienceDirect
Solar Energy Materials & Solar Cells journal homepage: www.elsevier.com/locate/solmat
Effect of sputtering process parameters on film properties of molybdenum back contact Shirish A. Pethe n, Eigo Takahashi, Ashwani Kaul, Neelkanth G. Dhere Florida Solar Energy Center, University of Central Florida, 1679 Clearlake Road, Cocoa, FL 32922, USA
a r t i c l e i n f o
abstract
Available online 29 January 2012
Molybdenum back contact in CuIn1 xGaxSe2 ySy (CIGSeS) solar cells is usually deposited using DC magnetron sputtering. Properties of thin films are dependent on process parameters. Films deposited at high power and low pressure, tend to be more conductive. However, such films exhibit poor adhesional strength since the films are under compressive stress. Films deposited at low power and high pressure tend to be under tensile stress and exhibit higher roughness and resistivity, while the films adhere very well to the sodalime glass substrate. Therefore, it has been a practice to deposit multi-layered Mo back contact to achieve properties of good adhesion and higher conductivity. Deposition of multi-layered back contact results in either increase in deposition time if a single target is used or increase in foot print if multiple targets are used resulting in an increase in the total cost of production. Experiments were carried out to understand the effects of working pressure, sputtering power and working distance on molybdenum film properties with the final aim to develop a process recipe for deposition of a single molybdenum film with acceptable properties of both good adhesion and higher conductivity. Experiments were carried out by varying the working pressure and keeping the sputtering power constant and then varying the sputtering power keeping the working pressure constant. Adhesive tape test was performed on each film to determine the adhesional strength of the films. Moreover, the sheet resistance and the average roughness for each film were measured using a four probe measurement setup and the Dektak Profilometer, respectively. All experiments were also carried out on narrow and long glass strips in order to estimate the residual stress in the film using the bend test method. Published by Elsevier B.V.
Keywords: Mo back contact CIGS thin film solar cells DC magnetron sputtering
1. Introduction With the increasing cost of electricity from conventional sources and the effect of global warming it is essential to shift to renewable energies to meet the global energy demand. Photovoltaic (PV) is one of the most promising of all the renewable technologies. Among the various PV technologies, such as c-Si, a-Si:H, CdTe and CIGS; CIGS absorber based heterojunction solar cell has the large potential for cost reduction with respect to power generation efficiency. The highest reported efficiencies of a CIGS solar cell and module are 20.3% and 16.7%, respectively [1,2]. Several candidate materials have been experimented with for ideal back contact material for CIGS thin film solar cells. Molybdenum is found to be the most appropriate back contact material because of its inertness and high conductivity. Generally, the molybdenum film is deposited using DC magnetron sputtering. It is well known that sputtering process parameters such as sputtering power, working gas pressure and working distance control the properties of sputter-deposited Mo thin films [3–5]. For example, films deposited at lower working gas pressures
n
Corresponding author. E-mail address: [email protected] (S.A. Pethe).
0927-0248/$ - see front matter Published by Elsevier B.V. doi:10.1016/j.solmat.2011.11.038
generally have poor adherence to the substrate, higher conductivity and are under compressive stress, whereas those deposited at higher gas pressures tend to have good adherence, high resistivity and are under tensile stress [3]. Although it is essential to have a well adherent and low resistive back contact for a stable and highly efficient solar cell, it is very difficult to achieve this with a single Mo thin film layer. The state of art process solves this issue by deposition of a multi layered molybdenum back contact. Because this multi layer process requires several deposition steps and specialized equipment, it is not an ideal process in terms of both time and cost. The motivation of this investigation is to find out how DC magnetron sputtering deposition system parameters such as working gas pressure and sputtering power influence the properties of Mo thin film layers and the performances of CIGS solar cells, and eventually develop a suitable process to achieve a qualified Mo thin film single layer with the requisite properties.
2. Experimental setup Mo thin film deposition was carried out by DC magnetron sputtering. The working distance between the sputtering target
2
S.A. Pethe et al. / Solar Energy Materials & Solar Cells 100 (2012) 1–5
and the substrate was maintained at 90 mm for all depositions. Argon was the working gas during deposition. To achieve thickness uniformity over a 150 mm 100 mm substrate, the substrate was slowly moved over the sputtering target with dimension of 305 mm 102 mm using an in-house developed moving mechanism. The rate at which the substrate moves determines the total thickness of the film. Care was taken to make sure that the thickness of the Mo film was 3 mm. Experiments were carried out by systematically changing the working gas pressure while keeping the sputtering power constant and then changing the sputtering power while keeping the working gas pressure constant. The working gas pressure was varied in the range of 0.1–5 mTorr and the sputtering power was varied in the range from 200 to 300 W. Very thin and narrow glass strips of dimensions 0.15 mm 10 mm 150 mm were attached to the main glass substrate during deposition for stress measurements. Since the glass strips were extremely thin and very narrow, any stress developed by the deposited film resulted in a change in the curvature of the glass strips. The residual stress was then measured by the bending beam method. Internal stress can be written as [6]: " # 2 4Es hs d S¼ hc 3ð1uÞL20 where Es is the Young’s modulus of the glass strip, hs is the thickness of the strip, hc is the film thickness, u is the Poisson’s ratio of the strip, S is the average stress of the film, L0 is the length of the glass strip and d is the deflection. X-ray diffraction (XRD) with CuKa radiation was used to determine the preferred orientation and crystal structure of the films. Sheet resistance of Mo thin films was measured using a four-point probe measurement technique. Dektak profilometer was used to determine film thickness and surface topography. Atomic force microscope (AFM) was used to study the surface morphologies. The degree of adhesion was qualitatively assessed using the ‘‘Scotch-tape’’ test [6].
adhere poorly to the glass substrate, and those in tensile stress have a good adhesion. Since all of the films obtained in this investigation were in a tensile stress state, they also exhibited good adherence to the glass substrate. One of the reasons that only tensile stress was observed on our films could be that the working distance is more than the mean free path of sputtered atoms for the power and pressure range used in this work.
3.2. Crystal structure Fig. 1 shows the XRD pattern for films deposited at various working gas pressures and sputtering powers of 200 and 250 W. The XRD pattern for all films indicated the most intense peak at 2y ¼40.51 corresponding to /110S preferred orientation of BCC structure. The calculated lattice parameters of the Mo films are ˚ which is slightly smaller than the ranging from 3.141 to 3.148 A, ˚ This change in the lattice parameter is bulk value of 3.150 A. attributed to the strain associated with the residual stress in the Mo thin film. As seen in Fig. 2 the sharpness of the peak increases with decreasing working gas pressure. Hence it is concluded that the lower the working gas pressure the better the crystallinity of the film. At lower working gas pressures the energy of the incident atoms is higher because of lesser scattering. The energetic incident atoms impart higher momentum to depositing molybdenum atoms so that they can fill up microvoids or vacancies resulting in better crystallinity and enhancing grain growth. Also at lower working gas pressures, backscattered energetic neutral Argon atoms can get embedded in the Mo film.
3. Results and discussions In one case, molybdenum films were deposited while the working gas pressure remained constant and the sputtering power was varied. In the other case, the sputtering power was kept constant and the working gas pressure was varied. The resulting films were then analyzed to determine the effect of each parameter on the residual stress, the crystal structure and the resistivity of the film. 3.1. Adhesion All the films deposited in this investigation passed the Scotchtape test. It has been reported that Mo films in compressive stress
Fig. 2. Normalized XRD profiles for (110) reflection peak of Mo thin films deposited at 200 W.
Fig. 1. XRD pattern for films deposited at various working gas pressure at sputtering power of 200 and 250 W.
S.A. Pethe et al. / Solar Energy Materials & Solar Cells 100 (2012) 1–5
3
3.3. Grain size
3.4. Surface morphology
Using the XRD patterns the full width half max (FWHM) was estimated, which was then used to determine the crystallite size using the Scherrer equation, which is given by [7]
The average roughness of the films was determined using AFM and the average roughness is tabulated in Table 1. It was found that the films deposited at lower gas pressures were rougher than those deposited at intermediate pressures and again the roughness increased as the gas pressured increased. The increase in the average roughness at lower pressures was attributed to the coarse grains and at higher pressure it was attributed to the selfshadowing effect resulting in porous films. Fig. 5 shows the relationship between pseudo-kinetic energy of incident atoms and average roughness. The kinetic energy of the incident atoms was calculated as follows:
D¼
0:94l W cos y
where D is the mean grain size, l is the X-ray wavelength, W is the FWHM measured in radians and y is the Bragg angle. The grain size was found to increase from 115 to 325 A˚ by increasing the sputtering power and decreasing the gas pressure as shown in Fig. 3. It can be deduced that grain growth was facilitated by higher momentum of atoms as mentioned above. The higher momentum of atoms results in an increase in the coalescence of the grains. Fig. 4 shows the atomic force microscopy (AFM) images of several films and it can be clearly seen that the AFM images support the relationship established between the grain size and the process parameters.
spl Ein ¼ Eout exp kT where Ein is the kinetic energy of the incident atom, Eout is the initial kinetic energy of the atom, l is the distance between the substrate and the target, s is the cross section for momentum transfer collision with background gas atom, k is the Boltzmann constant, p is the working gas pressure and T is the temperature.
Table 1 Effect of sputtering power and pressure on average roughness as measured by AFM.
Fig. 3. Mean grain size of Mo thin films as a function of working gas pressure.
Power (W)
Pressure (mTorr)
Ra (nm)
250 250 250 200 200 200
2 0.3 1 0.3 1 3
6.79 7.09 10.7 8.04 7.99 6.82
Fig. 4. AFM images of Mo thin film deposited at various sputtering conditions.
4
S.A. Pethe et al. / Solar Energy Materials & Solar Cells 100 (2012) 1–5
Fig. 7. Bending of glass strips at various sputtering conditions. Fig. 5. Average roughness variation of Mo thin film as a function of pseudo-kinetic energy of incident atoms.
Table 2 Residual stresses of Mo thin film deposited at different pressures. Power (W)
200
Pressure (mTorr)
Residual Stress (MPa)
0.1 0.3 1 3
322 358 101 0
Table 3 Residual stresses of Mo thin film deposited at different powers. Power (W)
Fig. 6. Sheet resistance of Mo thin films as a function of working gas pressure.
It should be noted that an assumption made here is that Eout is equal to the sputtering voltage, which is in turn dependant on the sputtering power. In lower energy region, porous microstructure is grown due to low kinetic energy of incident atoms resulting in rougher surface. The roughness decreases with increasing the kinetic energy of incident atoms; however, it increases again in higher energy region. Rougher surfaces in higher energy region are attributed to the large coarse grains. 3.5. Sheet resistance As can be seen from Fig. 6 the working gas pressure strongly influences the sheet resistance of the molybdenum films. The sheet resistance does change with sputtering power but change in the working gas pressure is the more dominant mechanism in determining the sheet resistance of the film. The lowest sheet resistance of 0.25 O/& was obtained in films prepared at a working gas pressure of 0.1 mTorr and a power of 300 W. The observed increase in sheet resistance shown at higher gas pressure is deduced to be the direct result of the porous microstructure and small grain size as mentioned previously. 3.6. Stress analysis Fig. 7 shows the images of glass strips, which were bent by the residual stress of deposited Mo thin film. It should be noted that the image is not specific to any sputtering condition and is just a representative of some of the sputtering conditions that are experimented in this work. The image is provided to show that a significant amount of bending is caused in the thin and narrow glass strips. This enables to measure the radius of curvature easily
200 225 250 275 300
Pressure (mTorr)
0.1
Residual stress (MPa) 322 249 213 204 152
and also more accurately. Tables 2 and 3 show the employed sputtering conditions and corresponding residual stresses. As can be seen from Tables 2 and 3, the residual stress is inversely proportional to the sputtering power. As the sputtering power decreases the energy of the incident sputtered atoms and the neutralized Argon atoms reduces thus causing an open structure. Such an open structure results in an increase in the intergranular spacing, which results in the decrease in the attractive force. Hence, the tensile stress increases with decreasing sputtering power. Similarly, at higher sputtering powers the films are more compactly packed thus resulting in the reduction in the tensile stress developed in the films. At higher working gas pressures, the number of collisions that the sputtered atom and the neutralized Argon atom encounter before getting deposited on the substrate increases. The increase in number of collisions results in the reduction of the kinetic energy of the sputtered atoms as well as that of the neutralized Argon atoms. This leads to an open structure resulting in an increase in the tensile stress in the film. Attractive force strength, which is responsible for tensile stress, among the grains is inversely proportional to the intergranular spacing. Therefore, by decreasing the target voltage, the films developed a tensile stress with an increase in the intergranular spacing. Reversely, increasing the discharge voltage results in a lowering of the number of open sites inside the film, and thereby reducing the tensile stress. If the film is deposited at higher pressures, it can be expected that collisions with background Argon gas atoms reduce the kinetic energy of sputtered Mo atoms and reflected neutral Argon gas atoms, resulting in more porous and less densely packed Mo film with a significant tensile stress state. The discussion is valid below the gas pressure of 0.3 mTorr; however, further increase of gas pressure beyond this value
S.A. Pethe et al. / Solar Energy Materials & Solar Cells 100 (2012) 1–5
results in attenuation of the tensile stress down to zero. This tensile stress reduction is explained by the model proposed by ¨ Muller [8]. Residual tensile stress is related to the defect and void size distribution; however, the intergranular attractive force works effectively only when the void size is smaller than a certain critical size. It is interesting to note that in this work we have observed only tensile stress has been developed for all the sputtering conditions that were experimented with. However, other groups have reported compressive stress at similar sputtering conditions. One of the possible reasons for this discrepancy could be difference in the working distance. Therefore, it is likely that in our case the working distance is more than the mean free path of the sputtered atoms thus resulting in films only exhibiting tensile stress.
4. Conclusions It can be concluded that the sputtering process parameters determine the film properties. The residual stress is strongly influenced by the kinetic energy of the incident sputtered atoms and the backscattered Argon atoms, which in turn is determined by the sputtering power and working gas pressure. Resistivity was found to be dependent on working gas pressure. Low sheet resistance ranging from 0.25 to 0.6 O/& was obtained at 0.1 to 0.3 mTorr between 200 and 300 W. Coarse grains tend to grow at lower gas pressure and higher sputtering power conditions, which make the film surface rougher. At higher pressure and lower power columnar grains with porous microstructure are
5
observed. Based on the work carried out in this paper, the sputtering parameters for Mo deposition were determined and a Mo back contact layer was deposited. Further, 14% CIGS thin film solar cells were prepared on this back contact.
Acknowledgment Authors are thankful to Dr. Helio Moutinho, National Renewable Energy Laboratory for AFM and XRD analyses. References [1] M. Powalla, Solar thin film reaches 20.3% efficiency, 2010. ZSW (Center for Solar Energy and Hydrogen Research) Baden-Wurttemberg press release. [2] M.A. Green, K. Emery, Y. Hishikawa, W. Warta, Solar cell efficiency tables (version 35), Progress in Photovoltaics: Research and Applications 18 (2010) 144–150. [3] J.H. Scofield, A. Duda, D. Albin, B.L. Ballard, P.K. Predecki, Sputtered molybdenum bilayer back contact for CIS based polycrystalline thin-film solar cells, Thin Solid Films 260 (1995) 26–31. [4] K. Orgassa, H.W. Schock, J.H. Werner, Alternative back contact materials for thin film Cu(In,Ga)Se2 solar cells, Thin Solid Films 431-432 (2003) 387–391. [5] G. Gordillo, M. Griza´lez, L.C. Hernandez, Structural and electrical properties of DC sputtered molybdenum films, Solar Energy Materials and Solar Cells 51 (1998) 327–337. [6] L.I. Maissel, R. Glang (Eds.), Handbook of thin film technology, McGraw-Hill, NY, USA, 1983. [7] E.F. Kaeble, Handbook of X-rays, McGraw-Hill, NY, USA, 1967. ¨ [8] K.H. Muller, Stress and microstructure of sputter-deposited thin films: molecular dynamics investigations, Journal of Applied Physics 62 (1987) 1796–1799.
Contents lists available at SciVerse ScienceDirect
Solar Energy Materials & Solar Cells journal homepage: www.elsevier.com/locate/solmat
Effect of sputtering process parameters on film properties of molybdenum back contact Shirish A. Pethe n, Eigo Takahashi, Ashwani Kaul, Neelkanth G. Dhere Florida Solar Energy Center, University of Central Florida, 1679 Clearlake Road, Cocoa, FL 32922, USA
a r t i c l e i n f o
abstract
Available online 29 January 2012
Molybdenum back contact in CuIn1 xGaxSe2 ySy (CIGSeS) solar cells is usually deposited using DC magnetron sputtering. Properties of thin films are dependent on process parameters. Films deposited at high power and low pressure, tend to be more conductive. However, such films exhibit poor adhesional strength since the films are under compressive stress. Films deposited at low power and high pressure tend to be under tensile stress and exhibit higher roughness and resistivity, while the films adhere very well to the sodalime glass substrate. Therefore, it has been a practice to deposit multi-layered Mo back contact to achieve properties of good adhesion and higher conductivity. Deposition of multi-layered back contact results in either increase in deposition time if a single target is used or increase in foot print if multiple targets are used resulting in an increase in the total cost of production. Experiments were carried out to understand the effects of working pressure, sputtering power and working distance on molybdenum film properties with the final aim to develop a process recipe for deposition of a single molybdenum film with acceptable properties of both good adhesion and higher conductivity. Experiments were carried out by varying the working pressure and keeping the sputtering power constant and then varying the sputtering power keeping the working pressure constant. Adhesive tape test was performed on each film to determine the adhesional strength of the films. Moreover, the sheet resistance and the average roughness for each film were measured using a four probe measurement setup and the Dektak Profilometer, respectively. All experiments were also carried out on narrow and long glass strips in order to estimate the residual stress in the film using the bend test method. Published by Elsevier B.V.
Keywords: Mo back contact CIGS thin film solar cells DC magnetron sputtering
1. Introduction With the increasing cost of electricity from conventional sources and the effect of global warming it is essential to shift to renewable energies to meet the global energy demand. Photovoltaic (PV) is one of the most promising of all the renewable technologies. Among the various PV technologies, such as c-Si, a-Si:H, CdTe and CIGS; CIGS absorber based heterojunction solar cell has the large potential for cost reduction with respect to power generation efficiency. The highest reported efficiencies of a CIGS solar cell and module are 20.3% and 16.7%, respectively [1,2]. Several candidate materials have been experimented with for ideal back contact material for CIGS thin film solar cells. Molybdenum is found to be the most appropriate back contact material because of its inertness and high conductivity. Generally, the molybdenum film is deposited using DC magnetron sputtering. It is well known that sputtering process parameters such as sputtering power, working gas pressure and working distance control the properties of sputter-deposited Mo thin films [3–5]. For example, films deposited at lower working gas pressures
n
Corresponding author. E-mail address: [email protected] (S.A. Pethe).
0927-0248/$ - see front matter Published by Elsevier B.V. doi:10.1016/j.solmat.2011.11.038
generally have poor adherence to the substrate, higher conductivity and are under compressive stress, whereas those deposited at higher gas pressures tend to have good adherence, high resistivity and are under tensile stress [3]. Although it is essential to have a well adherent and low resistive back contact for a stable and highly efficient solar cell, it is very difficult to achieve this with a single Mo thin film layer. The state of art process solves this issue by deposition of a multi layered molybdenum back contact. Because this multi layer process requires several deposition steps and specialized equipment, it is not an ideal process in terms of both time and cost. The motivation of this investigation is to find out how DC magnetron sputtering deposition system parameters such as working gas pressure and sputtering power influence the properties of Mo thin film layers and the performances of CIGS solar cells, and eventually develop a suitable process to achieve a qualified Mo thin film single layer with the requisite properties.
2. Experimental setup Mo thin film deposition was carried out by DC magnetron sputtering. The working distance between the sputtering target
2
S.A. Pethe et al. / Solar Energy Materials & Solar Cells 100 (2012) 1–5
and the substrate was maintained at 90 mm for all depositions. Argon was the working gas during deposition. To achieve thickness uniformity over a 150 mm 100 mm substrate, the substrate was slowly moved over the sputtering target with dimension of 305 mm 102 mm using an in-house developed moving mechanism. The rate at which the substrate moves determines the total thickness of the film. Care was taken to make sure that the thickness of the Mo film was 3 mm. Experiments were carried out by systematically changing the working gas pressure while keeping the sputtering power constant and then changing the sputtering power while keeping the working gas pressure constant. The working gas pressure was varied in the range of 0.1–5 mTorr and the sputtering power was varied in the range from 200 to 300 W. Very thin and narrow glass strips of dimensions 0.15 mm 10 mm 150 mm were attached to the main glass substrate during deposition for stress measurements. Since the glass strips were extremely thin and very narrow, any stress developed by the deposited film resulted in a change in the curvature of the glass strips. The residual stress was then measured by the bending beam method. Internal stress can be written as [6]: " # 2 4Es hs d S¼ hc 3ð1uÞL20 where Es is the Young’s modulus of the glass strip, hs is the thickness of the strip, hc is the film thickness, u is the Poisson’s ratio of the strip, S is the average stress of the film, L0 is the length of the glass strip and d is the deflection. X-ray diffraction (XRD) with CuKa radiation was used to determine the preferred orientation and crystal structure of the films. Sheet resistance of Mo thin films was measured using a four-point probe measurement technique. Dektak profilometer was used to determine film thickness and surface topography. Atomic force microscope (AFM) was used to study the surface morphologies. The degree of adhesion was qualitatively assessed using the ‘‘Scotch-tape’’ test [6].
adhere poorly to the glass substrate, and those in tensile stress have a good adhesion. Since all of the films obtained in this investigation were in a tensile stress state, they also exhibited good adherence to the glass substrate. One of the reasons that only tensile stress was observed on our films could be that the working distance is more than the mean free path of sputtered atoms for the power and pressure range used in this work.
3.2. Crystal structure Fig. 1 shows the XRD pattern for films deposited at various working gas pressures and sputtering powers of 200 and 250 W. The XRD pattern for all films indicated the most intense peak at 2y ¼40.51 corresponding to /110S preferred orientation of BCC structure. The calculated lattice parameters of the Mo films are ˚ which is slightly smaller than the ranging from 3.141 to 3.148 A, ˚ This change in the lattice parameter is bulk value of 3.150 A. attributed to the strain associated with the residual stress in the Mo thin film. As seen in Fig. 2 the sharpness of the peak increases with decreasing working gas pressure. Hence it is concluded that the lower the working gas pressure the better the crystallinity of the film. At lower working gas pressures the energy of the incident atoms is higher because of lesser scattering. The energetic incident atoms impart higher momentum to depositing molybdenum atoms so that they can fill up microvoids or vacancies resulting in better crystallinity and enhancing grain growth. Also at lower working gas pressures, backscattered energetic neutral Argon atoms can get embedded in the Mo film.
3. Results and discussions In one case, molybdenum films were deposited while the working gas pressure remained constant and the sputtering power was varied. In the other case, the sputtering power was kept constant and the working gas pressure was varied. The resulting films were then analyzed to determine the effect of each parameter on the residual stress, the crystal structure and the resistivity of the film. 3.1. Adhesion All the films deposited in this investigation passed the Scotchtape test. It has been reported that Mo films in compressive stress
Fig. 2. Normalized XRD profiles for (110) reflection peak of Mo thin films deposited at 200 W.
Fig. 1. XRD pattern for films deposited at various working gas pressure at sputtering power of 200 and 250 W.
S.A. Pethe et al. / Solar Energy Materials & Solar Cells 100 (2012) 1–5
3
3.3. Grain size
3.4. Surface morphology
Using the XRD patterns the full width half max (FWHM) was estimated, which was then used to determine the crystallite size using the Scherrer equation, which is given by [7]
The average roughness of the films was determined using AFM and the average roughness is tabulated in Table 1. It was found that the films deposited at lower gas pressures were rougher than those deposited at intermediate pressures and again the roughness increased as the gas pressured increased. The increase in the average roughness at lower pressures was attributed to the coarse grains and at higher pressure it was attributed to the selfshadowing effect resulting in porous films. Fig. 5 shows the relationship between pseudo-kinetic energy of incident atoms and average roughness. The kinetic energy of the incident atoms was calculated as follows:
D¼
0:94l W cos y
where D is the mean grain size, l is the X-ray wavelength, W is the FWHM measured in radians and y is the Bragg angle. The grain size was found to increase from 115 to 325 A˚ by increasing the sputtering power and decreasing the gas pressure as shown in Fig. 3. It can be deduced that grain growth was facilitated by higher momentum of atoms as mentioned above. The higher momentum of atoms results in an increase in the coalescence of the grains. Fig. 4 shows the atomic force microscopy (AFM) images of several films and it can be clearly seen that the AFM images support the relationship established between the grain size and the process parameters.
spl Ein ¼ Eout exp kT where Ein is the kinetic energy of the incident atom, Eout is the initial kinetic energy of the atom, l is the distance between the substrate and the target, s is the cross section for momentum transfer collision with background gas atom, k is the Boltzmann constant, p is the working gas pressure and T is the temperature.
Table 1 Effect of sputtering power and pressure on average roughness as measured by AFM.
Fig. 3. Mean grain size of Mo thin films as a function of working gas pressure.
Power (W)
Pressure (mTorr)
Ra (nm)
250 250 250 200 200 200
2 0.3 1 0.3 1 3
6.79 7.09 10.7 8.04 7.99 6.82
Fig. 4. AFM images of Mo thin film deposited at various sputtering conditions.
4
S.A. Pethe et al. / Solar Energy Materials & Solar Cells 100 (2012) 1–5
Fig. 7. Bending of glass strips at various sputtering conditions. Fig. 5. Average roughness variation of Mo thin film as a function of pseudo-kinetic energy of incident atoms.
Table 2 Residual stresses of Mo thin film deposited at different pressures. Power (W)
200
Pressure (mTorr)
Residual Stress (MPa)
0.1 0.3 1 3
322 358 101 0
Table 3 Residual stresses of Mo thin film deposited at different powers. Power (W)
Fig. 6. Sheet resistance of Mo thin films as a function of working gas pressure.
It should be noted that an assumption made here is that Eout is equal to the sputtering voltage, which is in turn dependant on the sputtering power. In lower energy region, porous microstructure is grown due to low kinetic energy of incident atoms resulting in rougher surface. The roughness decreases with increasing the kinetic energy of incident atoms; however, it increases again in higher energy region. Rougher surfaces in higher energy region are attributed to the large coarse grains. 3.5. Sheet resistance As can be seen from Fig. 6 the working gas pressure strongly influences the sheet resistance of the molybdenum films. The sheet resistance does change with sputtering power but change in the working gas pressure is the more dominant mechanism in determining the sheet resistance of the film. The lowest sheet resistance of 0.25 O/& was obtained in films prepared at a working gas pressure of 0.1 mTorr and a power of 300 W. The observed increase in sheet resistance shown at higher gas pressure is deduced to be the direct result of the porous microstructure and small grain size as mentioned previously. 3.6. Stress analysis Fig. 7 shows the images of glass strips, which were bent by the residual stress of deposited Mo thin film. It should be noted that the image is not specific to any sputtering condition and is just a representative of some of the sputtering conditions that are experimented in this work. The image is provided to show that a significant amount of bending is caused in the thin and narrow glass strips. This enables to measure the radius of curvature easily
200 225 250 275 300
Pressure (mTorr)
0.1
Residual stress (MPa) 322 249 213 204 152
and also more accurately. Tables 2 and 3 show the employed sputtering conditions and corresponding residual stresses. As can be seen from Tables 2 and 3, the residual stress is inversely proportional to the sputtering power. As the sputtering power decreases the energy of the incident sputtered atoms and the neutralized Argon atoms reduces thus causing an open structure. Such an open structure results in an increase in the intergranular spacing, which results in the decrease in the attractive force. Hence, the tensile stress increases with decreasing sputtering power. Similarly, at higher sputtering powers the films are more compactly packed thus resulting in the reduction in the tensile stress developed in the films. At higher working gas pressures, the number of collisions that the sputtered atom and the neutralized Argon atom encounter before getting deposited on the substrate increases. The increase in number of collisions results in the reduction of the kinetic energy of the sputtered atoms as well as that of the neutralized Argon atoms. This leads to an open structure resulting in an increase in the tensile stress in the film. Attractive force strength, which is responsible for tensile stress, among the grains is inversely proportional to the intergranular spacing. Therefore, by decreasing the target voltage, the films developed a tensile stress with an increase in the intergranular spacing. Reversely, increasing the discharge voltage results in a lowering of the number of open sites inside the film, and thereby reducing the tensile stress. If the film is deposited at higher pressures, it can be expected that collisions with background Argon gas atoms reduce the kinetic energy of sputtered Mo atoms and reflected neutral Argon gas atoms, resulting in more porous and less densely packed Mo film with a significant tensile stress state. The discussion is valid below the gas pressure of 0.3 mTorr; however, further increase of gas pressure beyond this value
S.A. Pethe et al. / Solar Energy Materials & Solar Cells 100 (2012) 1–5
results in attenuation of the tensile stress down to zero. This tensile stress reduction is explained by the model proposed by ¨ Muller [8]. Residual tensile stress is related to the defect and void size distribution; however, the intergranular attractive force works effectively only when the void size is smaller than a certain critical size. It is interesting to note that in this work we have observed only tensile stress has been developed for all the sputtering conditions that were experimented with. However, other groups have reported compressive stress at similar sputtering conditions. One of the possible reasons for this discrepancy could be difference in the working distance. Therefore, it is likely that in our case the working distance is more than the mean free path of the sputtered atoms thus resulting in films only exhibiting tensile stress.
4. Conclusions It can be concluded that the sputtering process parameters determine the film properties. The residual stress is strongly influenced by the kinetic energy of the incident sputtered atoms and the backscattered Argon atoms, which in turn is determined by the sputtering power and working gas pressure. Resistivity was found to be dependent on working gas pressure. Low sheet resistance ranging from 0.25 to 0.6 O/& was obtained at 0.1 to 0.3 mTorr between 200 and 300 W. Coarse grains tend to grow at lower gas pressure and higher sputtering power conditions, which make the film surface rougher. At higher pressure and lower power columnar grains with porous microstructure are
5
observed. Based on the work carried out in this paper, the sputtering parameters for Mo deposition were determined and a Mo back contact layer was deposited. Further, 14% CIGS thin film solar cells were prepared on this back contact.
Acknowledgment Authors are thankful to Dr. Helio Moutinho, National Renewable Energy Laboratory for AFM and XRD analyses. References [1] M. Powalla, Solar thin film reaches 20.3% efficiency, 2010. ZSW (Center for Solar Energy and Hydrogen Research) Baden-Wurttemberg press release. [2] M.A. Green, K. Emery, Y. Hishikawa, W. Warta, Solar cell efficiency tables (version 35), Progress in Photovoltaics: Research and Applications 18 (2010) 144–150. [3] J.H. Scofield, A. Duda, D. Albin, B.L. Ballard, P.K. Predecki, Sputtered molybdenum bilayer back contact for CIS based polycrystalline thin-film solar cells, Thin Solid Films 260 (1995) 26–31. [4] K. Orgassa, H.W. Schock, J.H. Werner, Alternative back contact materials for thin film Cu(In,Ga)Se2 solar cells, Thin Solid Films 431-432 (2003) 387–391. [5] G. Gordillo, M. Griza´lez, L.C. Hernandez, Structural and electrical properties of DC sputtered molybdenum films, Solar Energy Materials and Solar Cells 51 (1998) 327–337. [6] L.I. Maissel, R. Glang (Eds.), Handbook of thin film technology, McGraw-Hill, NY, USA, 1983. [7] E.F. Kaeble, Handbook of X-rays, McGraw-Hill, NY, USA, 1967. ¨ [8] K.H. Muller, Stress and microstructure of sputter-deposited thin films: molecular dynamics investigations, Journal of Applied Physics 62 (1987) 1796–1799.