- Email: [email protected]
Quantitative Scanning Microwave Microscopy: A calibration flow
Microelectronics Reliability xxx (2014) xxx–xxx
Contents lists available at ScienceDirect
Microelectronics Reliability journal homepage: www.elsevier.com/locate/microrel
Quantitative Scanning Microwave Microscopy: A calibration flow T. Schweinböck ⇑, S. Hommel Infineon Technologies, Failure Analysis, Am Campeon 1-12, 85579 Neubiberg, Germany
a r t i c l e
i n f o
Article history: Received 25 June 2014 Accepted 8 July 2014 Available online xxxx
a b s t r a c t A procedure for doping concentration calibration using Scanning Microwave Microscopy (SMM) is presented. Calibration measurements are performed at a purpose-built sample comprising a wide range of doping concentrations of both n- and p-type implants at the sample surface. In order to evaluate the criticality of parameters affecting the measured signal, a transmission line model is used that takes into account the physical elements forming the evaluated network including modelling parameters for the tip-sample-interaction. Using a vertical DMOS sample a showcase flow for quantification is exercised and the attainable degree of accuracy is discussed. Ó 2014 Elsevier Ltd. All rights reserved.
1. Introduction In the large variety of methods for characterizing semiconductor devices, the two-dimensional carrier profiling plays a major role since it helps detecting defects which are located e.g. in the active transistor area. Scanning Capacitance Microscopy/Spectroscopy (SCM/SCS), mainly used for carrier profiling, lacks from the disability to control the frequency of the used resonant circuit [1]. In order to overcome this, Scanning Microwave Microscopy (SMM) uses a transmission line coaxial resonator connected to a performance network analyser (PNA).1 Here, a frequency signal between 1 and 20 GHz can be chosen such that the network is close to resonance, which results in a higher sensitivity and better controlled measurement conditions [2–4]. Moreover, the system design is built such that influences of stray capacitors and electrical instabilities are minimized which makes it easier to compare quantitative results of different samples. In this paper, a Transition Line Model is presented, which takes into account the physical effective parameters for the MOS capacitor formed by the tip-sample system, like e.g. oxide thickness, tip radius and material, and majority carrier concentration [4,5]. The influence of these parameters on the model is discussed and compared to measurements on semiconductor samples. Furthermore, a new type of calibration sample is shown, which comprises several implant dose areas at the sample surface of both, n- and p-type, on one sample, as well as p–n-junctions of various levels, which can be used to calibrate the SMM system. A calibration flow is presented, which gives a guideline for determining quantitative car⇑ Corresponding author. Tel.: +49 (89) 234 23284; fax: +49 (89) 234 955 6637. E-mail address: Thomas.schweinboeck@infineon.com (T. Schweinböck). Here, a 5600LS AFM by Agilent combined with a N5232A PNA-L Microwave Network analyser is used. 1
rier concentrations on unknown samples with an uncertainty of better than 50 per cent on small areas in the order of 100 nm. 2. Transition Line Model The resonating circuit within the SMM is modelled using a circuitry of line segments shown in Fig. 1, where the tip-sampleinterface is described by a MOS capacitor. In order to simulate an additional DC voltage applied to the system, a voltage source is included in series to the MOS capacitor. The software ‘‘Advanced Design System 2009’’ by Agilent is used to calculate the S11 parameter of the system which corresponds to the measured PNA signal. The differentiated calculated signal from the MOS capacitor corresponds to the dC/dV signal, which reflects the carrier concentration and is proportional to the measured dopant profile measurement module (DPMM) signal. 2.1. Variation of physical parameters The MOS capacitor used in the transmission line model considers most of the influences of a real MOS capacitance. This is why a variation of majority carrier concentration, oxide thickness, and tip radius at various DC voltages offsets can be simulated to investigate those important parameters. The signal increases with decreasing oxide thickness; for practical use, homogeneously grown native oxide is most preferred. The signal increases with larger tip area, however, for practical use a trade-off between desired spatial resolution, S/N-ratio, and mechanical stability of the tip has to be found. Majority carrier concentration: taking a real MOS-capacitor for modelling the tip-sample-interaction one obtains a behaviour that is non-monotonous over the whole range of carrier concentration
http://dx.doi.org/10.1016/j.microrel.2014.07.024 0026-2714/Ó 2014 Elsevier Ltd. All rights reserved.
Please cite this article in press as: Schweinböck T, Hommel S. Quantitative Scanning Microwave Microscopy: A calibration flow. Microelectron Reliab (2014), http://dx.doi.org/10.1016/j.microrel.2014.07.024
2
T. Schweinböck, S. Hommel / Microelectronics Reliability xxx (2014) xxx–xxx
Fig. 1. Transmission line model used to describe the SMM network. The interacting area between tip and sample is modelled by the MOS capacitor shown.
between 1014 and 1020 cm 3 whereas the extreme value can be shifted by applying different DC bias values (see Fig. 2). This dependency can be explained considering the influence of the doping concentration not only on the MOS structure’s ideal capacitance but also on the flat band voltage of the system [5]. 2.2. Further measurement conditions influencing the system response In order to minimize stray capacitance and charging of the sample while applying a bias offset it is crucial to ensure conducting adhesion of the sample to the chuck. The used cantilever and tip material and radius determine the achieved resolution and measurement stability. Here, metal coated SiN cantilever tips are used (Nanosensors PtSi-FM-10). Full metal tips are not recommended since these tips suffer from fast degradation during scan, thereby changing the interaction behaviour and thus the calibrated value. 3. Calibration sample Calibration samples currently available on the market (e.g. IMEC) usually consist of a series (up to five different carrier concentrations) of epitaxial grown layers. For SCM/SMM measurements a cross-section is performed, in this way forming a staircase of either n- or p-type-doped stripes [5]. Here, a new approach is chosen to design a calibration sample which can be accessed directly without preparation.2 Using a mask stepping approach, up to twenty different implant steps have been performed on a low doped (1015 cm 3) p-Si substrate forming ten n- and p-implant areas, each, with nominal majority carrier concentrations from 4 1015 to 1 1020 cm 3 in a stripe-shape design (see Fig. 3). A set of larger areas, which can be used for Secondary Ion Mass Spectroscopy (SIMS) or Scanning Resistance Profiling (SRP) measurements has been created in the same process. Hence, a re-calibration is possible using methods with higher sensitivity but worse spatial resolution than SMM. SIMS measurements performed show that the profiles are homogeneous within the tip-sample-interaction volume. Only at a depth of more than 200 nm from the surface, a significant decrease of the carrier concentration is found. The measured carrier concentrations are fairly close to the values intended as shown in Table 1, with the exception of the two highest p-type boron implant, where an elevation near the surface was observed. However, comparison to SRP measurement results indicates that the carriers near the surface might only be partially electrically active. 2
Patent pending.
Fig. 2. Simulated dC/dV signal in dependence of carrier concentration for different DC bias values from 0 V (green) to 1 V (red) (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.).
Furthermore, an area where n- and p- type stripes are arranged in a cross-shaped manner is available, allowing for imaging of p–njunctions of different kinds. 4. Calibration flow For the quantification of carrier concentrations within an unknown sample, the following calibration procedure consisting of three main steps is used. At first, the calibration sample is imaged using three different DC bias voltage values. The results obtained can be interpolated to gain the calibration curves for the dS11/dV signal against doping concentration. The second step is to measure the investigated sample using the same probe tip and applying the same three DC bias voltages which were chosen for the calibration curves. In order to ensure comparability of the results, the oxide layer of the samples should have a similar thickness and the connectivity of the substrates should be comparable. Finally, the resulting values of the dS11/dV Signal, observed at a sample location of interest, can be assigned to their corresponding values in the calibration curves. The interval matching for all three calibration curves leads directly to the interval for the doping concentration. Therefore, it is not necessary to have a monotonous behaviour of the relation between signal and majority carrier concentration over the complete investigated doping range. For a better accuracy, the result can be improved by fitting a linear approach through the measured values in the possible range. A typical calibration flow can be seen in Fig. 4. The calibration sample is recorded three times with DC bias of 0.5 V,+0.5 V and +1.5 V (Fig. 4a). The same procedure is repeated with the device under test at the same conditions. The example in Fig. 4b shows a vertical DMOS transistor. The results are compared to the values of the calibration sample as shown in Fig. 4c. The intersection of the measured dC/dV signal at the region of interest with the corresponding calibration curve determines the carrier concentration. For a non-monotonous behaviour two intersection points and hence two carrier concentration values are possible using only one bias voltage. By comparison to the resulting values for the other bias curves, however, it is possible to obtain a defined value for the carrier concentration, and moreover to get a value for the accuracy of the measurement. In the shown example, a linear fit was performed at the surrounding of the intersection (shown in Fig. 4d). As a possible doping range, a value between 2.1 1019 ND/cm3 and 3.9 1019 ND/ cm3 for the n doped source implants marked in Fig. 4b, can be
Please cite this article in press as: Schweinböck T, Hommel S. Quantitative Scanning Microwave Microscopy: A calibration flow. Microelectron Reliab (2014), http://dx.doi.org/10.1016/j.microrel.2014.07.024
T. Schweinböck, S. Hommel / Microelectronics Reliability xxx (2014) xxx–xxx
3
Fig. 3. Top: Optical overview image of the calibration sample designed and fabricated at Infineon Technologies sized 5.5 3.5 mm2 (patent pending); Center: SMM dC/dVimage of the stripe-shaped area; an area of 50 lm comprises twenty different carrier concentration areas, from left to right: n-type increasing from 4 1015 to 1 1020 cm 3, p-type decreasing from 1 1020 to 4 1015 cm 3; Bottom: cross-shaped stripes forming p–n-junctions. Table 1 Nominal values of surface carrier concentration (left column) and values determined by SIMS measurement of the n-type (centre column) and p-type (right column) implants. Simulated values (cm 20
1 10 4 1019 1 1019 4 1018 1 1018 4 1017 1 1017 4 1016 1 1016 4 1015
3
)
SIMS values nP (cm 20
1.5 10 5.5 1019 1.2 1019 5 1018 1.1 1018 4 1017 1 1017 4 1016 1 1016 4 1015
3
)
SIMS values nB (cm
3
)
obtained. Similarly, for the p-body area a carrier concentration between 6.0 1016 NA/cm3 and 1.2 1017 NA/cm3 can be obtained. Noticeably, the method works also for rather small doping areas; the n-doped source implant in this example has a width in the order of 100 nm.
20
9 10 8 1019 1.4 1019 5 1018 1.2 1018 3.5 1017 1 1017 4 1016 1 1016 4 1015
5. Conclusion Using SMM and a calibration sample containing a large number of different implant areas, it is possible to achieve highly reliable quantitative values of carrier concentrations of small implant areas that cannot be accessed by other methods like SIMS or Spreading Resistance Profiling (SRP). By applying the procedure shown, it is
Please cite this article in press as: Schweinböck T, Hommel S. Quantitative Scanning Microwave Microscopy: A calibration flow. Microelectron Reliab (2014), http://dx.doi.org/10.1016/j.microrel.2014.07.024
4
T. Schweinböck, S. Hommel / Microelectronics Reliability xxx (2014) xxx–xxx
Fig. 4. (a) dS11/dV images of the calibration sample recorded at DC bias values of 0.5 V (left), 0.5 V (center) and 1.5 V (right); (b) dS11/dV images of an investigated DMOS sample recorded at DC bias values of 0.5 V (left), 0.5 V (center) and 1.5 V (right); (c) Calibration curves corresponding to the dS11/dV images of the calibration sample; (d) Linear fit through the surrounding values of the calibration curves to identify the possible doping range.
possible to determine implant concentrations of unknown samples. By measurements at various DC bias values a better confidence is obtained and the accuracy of the achieved dopant concentration value can be estimated. Furthermore, SMM measurements at the calibration sample can help to characterize and localize p–n-junctions and generate a better understanding of the interpretation of dC/dV.
Acknowledgements The authors want to give special thanks to Ferry Kienberger, Agilent Technologies Linz, and Jürgen Smoliner, TU Vienna, for measurement support and valuable discussions. Moreover, support on the transition line model from Fritz Schumann and Sébastien Mézière as well as valuable discussions with
Please cite this article in press as: Schweinböck T, Hommel S. Quantitative Scanning Microwave Microscopy: A calibration flow. Microelectron Reliab (2014), http://dx.doi.org/10.1016/j.microrel.2014.07.024
T. Schweinböck, S. Hommel / Microelectronics Reliability xxx (2014) xxx–xxx
Andreas Altes (all from Infineon Technologies) are greatly acknowledged. References
5
[2] Huber HP et al. Rev Sci Instrum 2010;81:113701. [3] Smoliner J, Huber HP, Hochleitner M, Moertelmaier, Kienberger F. JAP 2010;108:064315. [4] Huber HP, Humer I, Hochleitner M, Fenner M, Moertelmaier M, Rankl C, et al. JAP 2012;111:014301. [5] Torigoe K, Arita M, Motooka T. JAP 2012;112:104325.
[1] Morita Seizo. Roadmap of scanning probe microscopy. Springer; 2007.
Please cite this article in press as: Schweinböck T, Hommel S. Quantitative Scanning Microwave Microscopy: A calibration flow. Microelectron Reliab (2014), http://dx.doi.org/10.1016/j.microrel.2014.07.024
Contents lists available at ScienceDirect
Microelectronics Reliability journal homepage: www.elsevier.com/locate/microrel
Quantitative Scanning Microwave Microscopy: A calibration flow T. Schweinböck ⇑, S. Hommel Infineon Technologies, Failure Analysis, Am Campeon 1-12, 85579 Neubiberg, Germany
a r t i c l e
i n f o
Article history: Received 25 June 2014 Accepted 8 July 2014 Available online xxxx
a b s t r a c t A procedure for doping concentration calibration using Scanning Microwave Microscopy (SMM) is presented. Calibration measurements are performed at a purpose-built sample comprising a wide range of doping concentrations of both n- and p-type implants at the sample surface. In order to evaluate the criticality of parameters affecting the measured signal, a transmission line model is used that takes into account the physical elements forming the evaluated network including modelling parameters for the tip-sample-interaction. Using a vertical DMOS sample a showcase flow for quantification is exercised and the attainable degree of accuracy is discussed. Ó 2014 Elsevier Ltd. All rights reserved.
1. Introduction In the large variety of methods for characterizing semiconductor devices, the two-dimensional carrier profiling plays a major role since it helps detecting defects which are located e.g. in the active transistor area. Scanning Capacitance Microscopy/Spectroscopy (SCM/SCS), mainly used for carrier profiling, lacks from the disability to control the frequency of the used resonant circuit [1]. In order to overcome this, Scanning Microwave Microscopy (SMM) uses a transmission line coaxial resonator connected to a performance network analyser (PNA).1 Here, a frequency signal between 1 and 20 GHz can be chosen such that the network is close to resonance, which results in a higher sensitivity and better controlled measurement conditions [2–4]. Moreover, the system design is built such that influences of stray capacitors and electrical instabilities are minimized which makes it easier to compare quantitative results of different samples. In this paper, a Transition Line Model is presented, which takes into account the physical effective parameters for the MOS capacitor formed by the tip-sample system, like e.g. oxide thickness, tip radius and material, and majority carrier concentration [4,5]. The influence of these parameters on the model is discussed and compared to measurements on semiconductor samples. Furthermore, a new type of calibration sample is shown, which comprises several implant dose areas at the sample surface of both, n- and p-type, on one sample, as well as p–n-junctions of various levels, which can be used to calibrate the SMM system. A calibration flow is presented, which gives a guideline for determining quantitative car⇑ Corresponding author. Tel.: +49 (89) 234 23284; fax: +49 (89) 234 955 6637. E-mail address: Thomas.schweinboeck@infineon.com (T. Schweinböck). Here, a 5600LS AFM by Agilent combined with a N5232A PNA-L Microwave Network analyser is used. 1
rier concentrations on unknown samples with an uncertainty of better than 50 per cent on small areas in the order of 100 nm. 2. Transition Line Model The resonating circuit within the SMM is modelled using a circuitry of line segments shown in Fig. 1, where the tip-sampleinterface is described by a MOS capacitor. In order to simulate an additional DC voltage applied to the system, a voltage source is included in series to the MOS capacitor. The software ‘‘Advanced Design System 2009’’ by Agilent is used to calculate the S11 parameter of the system which corresponds to the measured PNA signal. The differentiated calculated signal from the MOS capacitor corresponds to the dC/dV signal, which reflects the carrier concentration and is proportional to the measured dopant profile measurement module (DPMM) signal. 2.1. Variation of physical parameters The MOS capacitor used in the transmission line model considers most of the influences of a real MOS capacitance. This is why a variation of majority carrier concentration, oxide thickness, and tip radius at various DC voltages offsets can be simulated to investigate those important parameters. The signal increases with decreasing oxide thickness; for practical use, homogeneously grown native oxide is most preferred. The signal increases with larger tip area, however, for practical use a trade-off between desired spatial resolution, S/N-ratio, and mechanical stability of the tip has to be found. Majority carrier concentration: taking a real MOS-capacitor for modelling the tip-sample-interaction one obtains a behaviour that is non-monotonous over the whole range of carrier concentration
http://dx.doi.org/10.1016/j.microrel.2014.07.024 0026-2714/Ó 2014 Elsevier Ltd. All rights reserved.
Please cite this article in press as: Schweinböck T, Hommel S. Quantitative Scanning Microwave Microscopy: A calibration flow. Microelectron Reliab (2014), http://dx.doi.org/10.1016/j.microrel.2014.07.024
2
T. Schweinböck, S. Hommel / Microelectronics Reliability xxx (2014) xxx–xxx
Fig. 1. Transmission line model used to describe the SMM network. The interacting area between tip and sample is modelled by the MOS capacitor shown.
between 1014 and 1020 cm 3 whereas the extreme value can be shifted by applying different DC bias values (see Fig. 2). This dependency can be explained considering the influence of the doping concentration not only on the MOS structure’s ideal capacitance but also on the flat band voltage of the system [5]. 2.2. Further measurement conditions influencing the system response In order to minimize stray capacitance and charging of the sample while applying a bias offset it is crucial to ensure conducting adhesion of the sample to the chuck. The used cantilever and tip material and radius determine the achieved resolution and measurement stability. Here, metal coated SiN cantilever tips are used (Nanosensors PtSi-FM-10). Full metal tips are not recommended since these tips suffer from fast degradation during scan, thereby changing the interaction behaviour and thus the calibrated value. 3. Calibration sample Calibration samples currently available on the market (e.g. IMEC) usually consist of a series (up to five different carrier concentrations) of epitaxial grown layers. For SCM/SMM measurements a cross-section is performed, in this way forming a staircase of either n- or p-type-doped stripes [5]. Here, a new approach is chosen to design a calibration sample which can be accessed directly without preparation.2 Using a mask stepping approach, up to twenty different implant steps have been performed on a low doped (1015 cm 3) p-Si substrate forming ten n- and p-implant areas, each, with nominal majority carrier concentrations from 4 1015 to 1 1020 cm 3 in a stripe-shape design (see Fig. 3). A set of larger areas, which can be used for Secondary Ion Mass Spectroscopy (SIMS) or Scanning Resistance Profiling (SRP) measurements has been created in the same process. Hence, a re-calibration is possible using methods with higher sensitivity but worse spatial resolution than SMM. SIMS measurements performed show that the profiles are homogeneous within the tip-sample-interaction volume. Only at a depth of more than 200 nm from the surface, a significant decrease of the carrier concentration is found. The measured carrier concentrations are fairly close to the values intended as shown in Table 1, with the exception of the two highest p-type boron implant, where an elevation near the surface was observed. However, comparison to SRP measurement results indicates that the carriers near the surface might only be partially electrically active. 2
Patent pending.
Fig. 2. Simulated dC/dV signal in dependence of carrier concentration for different DC bias values from 0 V (green) to 1 V (red) (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.).
Furthermore, an area where n- and p- type stripes are arranged in a cross-shaped manner is available, allowing for imaging of p–njunctions of different kinds. 4. Calibration flow For the quantification of carrier concentrations within an unknown sample, the following calibration procedure consisting of three main steps is used. At first, the calibration sample is imaged using three different DC bias voltage values. The results obtained can be interpolated to gain the calibration curves for the dS11/dV signal against doping concentration. The second step is to measure the investigated sample using the same probe tip and applying the same three DC bias voltages which were chosen for the calibration curves. In order to ensure comparability of the results, the oxide layer of the samples should have a similar thickness and the connectivity of the substrates should be comparable. Finally, the resulting values of the dS11/dV Signal, observed at a sample location of interest, can be assigned to their corresponding values in the calibration curves. The interval matching for all three calibration curves leads directly to the interval for the doping concentration. Therefore, it is not necessary to have a monotonous behaviour of the relation between signal and majority carrier concentration over the complete investigated doping range. For a better accuracy, the result can be improved by fitting a linear approach through the measured values in the possible range. A typical calibration flow can be seen in Fig. 4. The calibration sample is recorded three times with DC bias of 0.5 V,+0.5 V and +1.5 V (Fig. 4a). The same procedure is repeated with the device under test at the same conditions. The example in Fig. 4b shows a vertical DMOS transistor. The results are compared to the values of the calibration sample as shown in Fig. 4c. The intersection of the measured dC/dV signal at the region of interest with the corresponding calibration curve determines the carrier concentration. For a non-monotonous behaviour two intersection points and hence two carrier concentration values are possible using only one bias voltage. By comparison to the resulting values for the other bias curves, however, it is possible to obtain a defined value for the carrier concentration, and moreover to get a value for the accuracy of the measurement. In the shown example, a linear fit was performed at the surrounding of the intersection (shown in Fig. 4d). As a possible doping range, a value between 2.1 1019 ND/cm3 and 3.9 1019 ND/ cm3 for the n doped source implants marked in Fig. 4b, can be
Please cite this article in press as: Schweinböck T, Hommel S. Quantitative Scanning Microwave Microscopy: A calibration flow. Microelectron Reliab (2014), http://dx.doi.org/10.1016/j.microrel.2014.07.024
T. Schweinböck, S. Hommel / Microelectronics Reliability xxx (2014) xxx–xxx
3
Fig. 3. Top: Optical overview image of the calibration sample designed and fabricated at Infineon Technologies sized 5.5 3.5 mm2 (patent pending); Center: SMM dC/dVimage of the stripe-shaped area; an area of 50 lm comprises twenty different carrier concentration areas, from left to right: n-type increasing from 4 1015 to 1 1020 cm 3, p-type decreasing from 1 1020 to 4 1015 cm 3; Bottom: cross-shaped stripes forming p–n-junctions. Table 1 Nominal values of surface carrier concentration (left column) and values determined by SIMS measurement of the n-type (centre column) and p-type (right column) implants. Simulated values (cm 20
1 10 4 1019 1 1019 4 1018 1 1018 4 1017 1 1017 4 1016 1 1016 4 1015
3
)
SIMS values nP (cm 20
1.5 10 5.5 1019 1.2 1019 5 1018 1.1 1018 4 1017 1 1017 4 1016 1 1016 4 1015
3
)
SIMS values nB (cm
3
)
obtained. Similarly, for the p-body area a carrier concentration between 6.0 1016 NA/cm3 and 1.2 1017 NA/cm3 can be obtained. Noticeably, the method works also for rather small doping areas; the n-doped source implant in this example has a width in the order of 100 nm.
20
9 10 8 1019 1.4 1019 5 1018 1.2 1018 3.5 1017 1 1017 4 1016 1 1016 4 1015
5. Conclusion Using SMM and a calibration sample containing a large number of different implant areas, it is possible to achieve highly reliable quantitative values of carrier concentrations of small implant areas that cannot be accessed by other methods like SIMS or Spreading Resistance Profiling (SRP). By applying the procedure shown, it is
Please cite this article in press as: Schweinböck T, Hommel S. Quantitative Scanning Microwave Microscopy: A calibration flow. Microelectron Reliab (2014), http://dx.doi.org/10.1016/j.microrel.2014.07.024
4
T. Schweinböck, S. Hommel / Microelectronics Reliability xxx (2014) xxx–xxx
Fig. 4. (a) dS11/dV images of the calibration sample recorded at DC bias values of 0.5 V (left), 0.5 V (center) and 1.5 V (right); (b) dS11/dV images of an investigated DMOS sample recorded at DC bias values of 0.5 V (left), 0.5 V (center) and 1.5 V (right); (c) Calibration curves corresponding to the dS11/dV images of the calibration sample; (d) Linear fit through the surrounding values of the calibration curves to identify the possible doping range.
possible to determine implant concentrations of unknown samples. By measurements at various DC bias values a better confidence is obtained and the accuracy of the achieved dopant concentration value can be estimated. Furthermore, SMM measurements at the calibration sample can help to characterize and localize p–n-junctions and generate a better understanding of the interpretation of dC/dV.
Acknowledgements The authors want to give special thanks to Ferry Kienberger, Agilent Technologies Linz, and Jürgen Smoliner, TU Vienna, for measurement support and valuable discussions. Moreover, support on the transition line model from Fritz Schumann and Sébastien Mézière as well as valuable discussions with
Please cite this article in press as: Schweinböck T, Hommel S. Quantitative Scanning Microwave Microscopy: A calibration flow. Microelectron Reliab (2014), http://dx.doi.org/10.1016/j.microrel.2014.07.024
T. Schweinböck, S. Hommel / Microelectronics Reliability xxx (2014) xxx–xxx
Andreas Altes (all from Infineon Technologies) are greatly acknowledged. References
5
[2] Huber HP et al. Rev Sci Instrum 2010;81:113701. [3] Smoliner J, Huber HP, Hochleitner M, Moertelmaier, Kienberger F. JAP 2010;108:064315. [4] Huber HP, Humer I, Hochleitner M, Fenner M, Moertelmaier M, Rankl C, et al. JAP 2012;111:014301. [5] Torigoe K, Arita M, Motooka T. JAP 2012;112:104325.
[1] Morita Seizo. Roadmap of scanning probe microscopy. Springer; 2007.
Please cite this article in press as: Schweinböck T, Hommel S. Quantitative Scanning Microwave Microscopy: A calibration flow. Microelectron Reliab (2014), http://dx.doi.org/10.1016/j.microrel.2014.07.024