Articles
Book Chapters
1. Carolyn F. H. Gondran, Chapter 9 “Applications: Semiconductor Manufacturing”, in Scanning Auger Electron Microscopy Martin Prutton, Mohamed El Gomati; John Wiley & Sons, Ltd, Chichester, West Sussex, UK 2006.
2. Chris M. Sparks and Alain Diebold, Chapter 15 “Novel Analytical Methods for Cleaning Evaluation”, in Cleaning for Semiconductor Manufacturing: Fundamentals and Applications Karen Reinhardt, Rick Reidy; Martin Scrivener Publishing, Salem, MA, USA expected 2010.
Journal Publications
Abstract
Selective sidewall epitaxy of AlN/GaN is reported on vertical fins of silicon using metallorganic vapor phase epitaxy. Silicon (110) wafers are structured to form fins with {111} sidewall facets. AlN buffer layers are grown with uniform thickness on vertical {111} surfaces, followed by GaN which grows selectively on the AlN to form the sidewall fin structures. Raman measurements of the GaN show very narrow line widths, consistent with excellent material quality. Spatial dependence from microcathodoluminescence mapping of the GaN band gap emission shows compressive strain in the GaN relaxes closer to the fin corners.
http://dx.doi.org/10.1063/1.3310279
Abstract
The authors present experimental evidence on the impact of three different chemical surface treatments on the interface between the GaAs substrate and the aluminum oxide dielectric layer used in the fabrication of metal-oxide-semiconductor capacitors. The three different chemical surface treatments studied prior to atomic layer deposition (ALD) of the dielectric layer include (a) GaAs native oxide removal in a dilute HF solution only, (b) HF etch followed by a NH4OH treatment, and (c) HF etch followed by a (NH4)2S treatment. Moreover, interfacial self-cleaning of nontreated GaAs wafers upon ALD of aluminum oxide using trimethyl aluminum precursor was examined. Transmission electron microscopy, electron energy loss spectroscopy (EELS) and capacitance-voltage (C-V) data showed slight differences among the nontreated, HF-only, and NH4OH treated samples. However the (NH4)2S treated sample showed improved capacitance-voltage characteristics as well as an improved aluminum oxide/GaAs interface compared to the other three samples. Additionally, the characteristic oxygen K EELS peak suggests the presence of a thin additional layer close to the center of the high- layer containing oxygen, tantalum, and aluminum, as a consequence of probable plasma damage to the high- layer during the TaN metal gate deposition.
http://dx.doi.org/10.1116/1.3256229
Abstract.
One of the challenges for the nanoscale device fabrication of III-V semiconductors is controllable postdeposition doping techniques to create ultrashallow junctions. Here, we demonstrate nanoscale, sulfur doping of InAs planar substrates with high dopant areal dose and uniformity by using a self-limiting monolayer doping approach. From transmission electron microscopy and secondary ion mass spectrometry, a dopant profile abruptness of nm ∼ 3.5/decade is observed without significant defect density. The n+/p+ junctions fabricated by using this doping scheme exhibit negative differential resistance characteristics, further demonstrating the utility of this approach for device fabrication with high electrically active sulfur concentrations of ∼ 8×1018cm −3.
http://dx.doi.org/10.1063/1.3205113
Abstract
In this study, aluminum nitride (AlN) was grown by molecular layer deposition on HfO2 that had been deposited on 200 mm Si (100) substrates. The AlN was grown on HfO2 using sequential exposures of trimethyl-aluminum and ammonia (NH3) in a batch vertical furnace. Excellent thickness uniformity on test wafers from the top of the furnace to the bottom of the furnace (across the furnace load) was obtained. The equivalent oxide thickness was 16.5–18.8 Å for the AlN/HfO2 stack on patterned device wafers with a molybdenum oxynitride metal gate with leakage current densities from low 10− 5 to mid 10− 6 A/cm2 at threshold voltage minus one volt. There was no change in the work function with the AlN cap on HfO2 with the MoN metal gate, even with a 1000 °C anneal.
http://dx.doi.org/10.1016/j.tsf.2008.10.032
Abstract.
In this study, the capabilities of total reflection X-ray fluorescence spectroscopy characterization for both the photovoltaic industry and advanced semiconductor processing were investigated. Analysis of single crystal silicon coupon samples from various cleans during photovoltaic processing showed that certain clean steps were more effective in removing trace metal contamination. The multicrystalline photovoltaic silicon sample also had detected, but difficult to quantify, metallic contamination. Changes in the silicon dioxide content of hafnium silicate films used in semiconductor processing were also characterized by total reflection X-ray fluorescence spectroscopy analysis.
http://dx.doi.org/10.1016/j.sab.2008.10.032
Abstract.
Hafnium silicate is a so-called high-k material, which is a new key material in the semiconductor field. This material is difficult to analyze by a conventional W-Lβ1TXRF source due to the high background originating from Hf-Lα lines. In this paper, the capability of Ir source TXRF analysis on hafnium silicate films is investigated with intentional contamination of Ti, Cr, Fe, Ni and Cu elements. The spectral fitting is discussed where X-ray resonant Raman scattering and escape peak of Ir-Lα overlap with Ni-Kα peak. The detection limits are estimated to 0.9 × 1010 to 2 × 1010 atoms/cm2 for the transition metals.
http://dx.doi.org/10.1016/j.sab.2008.10.002
Abstract.
Total reflection X-ray fluorescence (TXRF) has undergone key adaptations in recent years to stay relevant to the semiconductor industry for contamination analysis. With the change from silicon dioxide-based dielectrics and polysilicon gates to new materials, there are not only challenges in processing these materials but also in measuring the effectiveness of the process. Traditionally, the metals removal efficiency of a cleans process is monitored in the fab with TXRF. However, these systems are typically optimized for analyzing a bare silicon or silicon dioxide wafer. If the surface of the wafer has a layer of material with an absorption edge below the energy of the excitation source (typically W-Lb1 or Mo-Ka), the spectrum can have a high background in the energy region of interest making quantification difficult. To evaluate the level of contamination on hafnium-based films, a new iridium source was developed for TXRF that allows greater differentiation of transition metals. Another recent advancement in TXRF analysis is the development of the ability to map contamination across a wafer. This development facilitates contamination control when there is a non-uniform distribution of metallics across the surface of the wafer.
http://www.scientific.net/SSP.134.285
Abstract
Next generation planar and non-planar complementary metal oxide semiconductor (CMOS) structures are three-dimensional nanostructures with multi-layer stacks that can contain films thinner than ten atomic layers. The high resolution of transmission electron microscopy (TEM) is typically chosen for studying properties of these stacks such as film thickness, interface and interfacial roughness. However, TEM sample preparation is time-consuming and destructive, and TEM analysis is expensive and can provide problematic results for surface and interface roughness. Therefore, in this paper, we present the use of direct measurements of sidewall surface structures by conventional atomic force microscopy (AFM) as an alternative or complementary method for studying multi-layer film stacks and as the preferred method for studying FinFET sidewall surface roughness. In addition to these semiconductor device applications, this AFM sidewall measurement technique could be used for other three-dimensional nanostructures.
http://dx.doi.org/10.1088/0957-4484/18/33/335303
Abstract.
SIMS has been applied to the characterization of high-k and low-k materials used in the semiconductor manufacturing process. Profiles of thin high-k films, particularly HfO2 and HfSiO, exhibit preferential sputtering that affects ion yields and sputter rates in the films as well as in the Si substrate. These artifacts make it difficult to quantify major constituents and dopants in the films and substrate. The ion yields of B were observed to vary by as much as 3× as a function of Si content in HfSiO films, while As ion yield variations are not as great. Evidence for B penetration from a highly-doped Si substrate into the high-k films was also observed and quantified. First generation low-k films are not as susceptible to charging as the newer, porous materials. Backside SIMS was used to show Ti migration into the open pores of a low-k film during the metal deposition step.
http://dx.doi.org/10.1016/j.apsusc.2006.02.087
Abstract.
High-k dielectrics are under study as part of the effort to continually reduce semiconductor device dimensions and hafnium silicate (HfSixOy) is one of the most promising high-k materials. A requirement of the dielectric is that the constituent elements cannot diffuse into adjacent device regions during thermal processing. Analysis for inter-diffusion using front side SIMS of high-k dielectrics has been complicated by matrix and sputtering effects.
Use of a back side analysis sample preparation procedure that was successful for copper diffusion and site specific studies produced a HfSiO specimen that has less than 250 nm silicon remaining and minimal slope over the analysis region. Magnetic Sector (CAMECA IMS-6F) SIMS analysis of this specimen with low energy O2+ bombardment does not show the matrix and sputtering effects noted in the front side data. Sufficient depth resolution was obtained to define the interface between the silicon substrate and the HfSiO layer and indicate what appears to be an interfacial layer. There is no indication of hafnium diffusion into the silicon substrate.
http://dx.doi.org/10.1016/j.apsusc.2006.02.099
Abstract.
An international round-robin study was undertaken under the auspices of ISO TC201/SC6 to determine the best analytical conditions and the level of interlaboratory agreement for the determination of the implantation dose of arsenic in silicon by secondary ion mass spectrometry (SIMS). Fifteen SIMS laboratories, as well as two laboratories that performed low energy electron-induced X-ray emission spectrometry (LEXES) and one that made measurements by instrumental neutron activation analysis (INAA) were asked to determine the implanted arsenic doses in three unknown samples using as a comparator NIST Standard Reference Material® 2134. The use of a common reference material by all laboratories resulted in better interlaboratory agreement than was seen in a previous round-robin that lacked a common comparator. The relative standard deviation among laboratories was less than 4% for the medium-dose sample, but several percent larger for the low- and high-dose samples. The high-dose sample showed a significant difference between point-by-point and average matrix normalization because the matrix signal decreased in the vicinity of the implant peak, as observed in a previous study. The dose from point-by-point normalization was in close agreement with that determined by INAA. No clear difference in measurement repeatability was seen when comparing Si2− and Si3− as matrix references with AsSi−.
http://dx.doi.org/10.1016/j.apsusc.2006.02.152
Abstract.
Two-dimensional geometric models and experimental data are used to evaluate the effect of tip shape artifacts on atomic force microscopy images, roughness values, and power spectra. The effects of tip size are studied as a function of the surface feature height, spacing, and width. The models demonstrate that the need for sharp tips is dictated by the sample roughness and the size and spacing of surface features. In addition, the magnitude and direction of the error in the measured roughness parameters are related to artifact-induced changes in the skew in the distribution of the data points about the surface mean. Significantly, it is shown that either higher or lower roughness values can be measured using a smaller sized tip, depending on the surface character. These model results are supported by sample data obtained with 2 and 10 nm radii of curvature tipped probes on films used in semiconductor research and development. Sharp-tipped probes are clearly beneficial for imaging the surface microstructure of low-k materials. Tip size has a marked impact on the roughness values obtained on a somewhat smoother metal film sample. However, the high-k sample studied was smooth enough that no benefit was realized with the use of a sharper-tipped probe.
http://dx.doi.org/10.1116/1.2172936
Abstract.
Sputter-induced artifacts can create the appearance of an interfacial reaction between Hf-based films and the underlying SiO2 in sputter depth profiles. A combination of front and back side Auger electron spectroscopy depth profile analysis is used to distinguish between a native interfacial reaction at the interface between HfN and SiO2 and a potential sputtering artifact. Experimental results show that a native reaction occurs; however, the extent of the reaction may be over-represented in the front side profile due to a sputter artifact component.
http://dx.doi.org/10.1116/1.2232380
15. “Focused ion beam grounding to alleviate sample charging for scanning Auger electron spectroscopy", C. F. H. Gondran and E. Morales, Proceedings of 32nd Int. Symposium Test and Failure Analysis (2006) 293.
Abstract
It is shown that a focused beam ion (FIB) grounding technique can be used to alleviate charge buildup on samples that would otherwise charge in the electron beam to the point where analysis by Auger electron spectroscopy (AES) was limited or impossible. FIB grounding alleviates the sample charging and permits AES analysis. The grounding technique is quick, easy and well understood as it has been used extensively for voltage-contrast analysis. The technique is shown to be useful for enabling analysis on electrically isolated conductive features as well as insulating samples.
16. "Buttons and threads: Tailoring defect analysis", C. F. H. Gondran, D. F. Paul, S. K. Das, B. Foran, and M. H. Clark, Proceedings of 32nd Int. Symposium Test and Failure Analysis (2006) 293.
Abstract.
In this study, a BioDot BioJet dispensing system was investigated as a nanoliter sample deposition method for total reflection X-ray fluorescence (TXRF) analysis. The BioDot system was programmed to dispense arrays of 20 nL droplets of sample solution on Si wafers. Each 20 nL droplet was approximately 100 μm in diameter. A 10 × 10 array (100 droplets) was deposited and dried in less than 2 min at room temperature and pressure, demonstrating the efficiency of the automated deposition method. Solutions of various concentrations of Ni and Ni in different matrices were made from stock trace element standards to investigate of the effect of the matrix on the TXRF signal. The concentrations were such that the levels of TXRF signal saturation could be examined. Arrays were deposited to demonstrate the capability of drying 100 μL of vapor phase decomposition-like residue in the area of a typical TXRF detector.
http://dx.doi.org/10.1016/j.sab.2006.09.014
Abstract.
The versatility of confocal micro X-ray fluorescence (MXRF) in analyzing thin films on semiconductor wafers is demonstrated. Unlike conventional MXRF, confocal MXRF can depth profile sample layers and reduce spectral background. Non destructive quantification of the silicon dioxide (SiO2) concentration in hafnium silicate (HfSiO) thin films is an example of one application demonstrating the advantage of confocal MXRF. Additionally, the growth of titanium nitride (TiN) films on various high-k gate dielectric substrates was analyzed with confocal MXRF due to its ability to detect sub-nm film thickness changes.
http://www.icdd.com/resources/axa/VOL49/V49_32.pdf
http://dx.doi.org/10.1017/S1431927606062532
Abstract.
This work introduces inline x-ray fluorescence (XRF) and x-ray reflectometry (XRR) metrology tools to aid the process development of atomic layer deposition on high-k dielectric films. In this approach, XRR monitors the deposition rate of the thickness and identifies the transition from the three-dimensional to the two-dimensional growth regime. XRF is used to monitor the atomic deposition rate. The interpretation of XRR result is verified with transmission electron microscopy (TEM) and Auger electron spectroscopy. Unlike the conventional approach using Rutherford backscattering to monitor deposition rate and TEM to monitor thickness, this proposed scheme is noninvasive and does not require any sample preparation. In addition, the inline approach prevents exposing the film to a nonproduction grade environment and avoids the potential growth of the high-k interface or degradation of the film during the measurement.
http://dx.doi.org/10.1116/1.2009774
Abstract
Transistor gate stack systems consisting of atomic layer deposited HfO2 with polycrystalline silicon or TiN gate electrodes have been characterized by analytical electron microscopy to elucidate underlying physical contributions to electrical performance differences. High-angle annular dark-field scanning transmission electron microscopy was used to determine film and interface thickness dimensions and chemical analysis depth profiling was obtained from electron energy loss spectra and energy dispersive x-ray spectra. The high-k gate dielectric film system is shown to be influenced by the choice of electrode material with the formation of an HfO2-poly-Si interface that increases the dielectric equivalent oxide thickness and may affect electron trapping characteristics.
http://dx.doi.org/10.1063/1.2011827
ABSTRACT
As complementary metal oxide semiconductor (CMOS) devices continue to scale along the rapid pace of Moore’s Law, gate dielectric materials with significantly higher dielectric constant (k = 10 – 25) are being evaluated as replacements for conventional silicon dioxide, SiO2 (k = 3.9), and silicon oxynitride. This allows for the introduction of a physically thicker film with lower leakage current and with capacitance equivalent to a thinner (1.0 nm and below) SiO2 layer [1-3]. Although binary metal oxide films such as HfO2 and ZrO2 exhibit higher permittivity than their corresponding silicates and aluminates, alloyed with various molecular percents of SiO2 or Al2O3, respectively, they are compromised by lower onset of crystallization temperature which contributes a higher degree of interfacial microroughness and increased gate leakage current due to dislocations and oxygen vacancies generated along grain boundaries. Accordingly, development of hafnium silicate has been the subject of intense investigation as an advanced gate dielectric thin film designed to meet the device manufacturing requirements of thermal stability in direct contact with substrate silicon and metal gate electrode materials. In this paper, we present results corresponding to the utilization of total reflection X-ray fluorescence spectroscopy (TXRF) as a quick, accurate, non-destructive technique for hafnium silicate composition determination based on detection of the Hf: Si ratio of (HfO2)x(SiO2)1-x, where x varies over the range 0.2 – 1.0.
http://www.icdd.com/resources/axa/VOL48/V48_39.pdf
Abstract.
We present the first experimental data on secondary ion emission characteristics from the impact of 26 keV Au3+ and 74.6–114.6 keV Au4004+. In particular we show secondary ion yield distributions and secondary ion and coincidental ion yields of molecular cluster ions from single impact events. The target consisted of an amorphous (HfO2)0.6(SiO2)0.4 layer deposited on a Si wafer. Large increases in higher order emission events and both secondary ion and coincidental ion yields within these events were observed for bombardment with Au4004+ even though the energy per atom of this projectile is more than an order of magnitude less than Au3+.
http://dx.doi.org/10.1016/j.ijms.2004.10.009
Abstract.
Ultrashallow p+ junctions are required for next generation electronics. We present a technique for the formation of ultrashallow p+ junctions that increases the thermal stability of the junctions formed by either epitaxy or ion implantation. By using a 10 nm Si1–xGex barrier layer, the diffusion of B is inhibited during high temperature processes. Alloys having a composition from x = 0–0.4 were investigated and it is shown that the most effective barrier had the maximum Ge fraction. The junction depth decreased to 36.7 nm for a 5×1015/cm2 1 kV BF3 plasma implant spike annealed at 1050 °C, compared to a junction depth of 48 nm for a Si control sample having the identical implant and anneal. It is hypothesized that the inhibition of B diffusion in the alloy layer is caused by a reduction of the Si self-interstitials in the alloy.
http://dx.doi.org/10.1116/1.1784825
25. C. F. H. Gondran and D. K. Michelson “Sampling and Reference Considerations for Very High Resolution AFM Analysis", Proceedings International Symposium for Testing and Failure Analysis (2004) 357.
Abstract
The sidewalls of etched Si lines will be the carrier channel surfaces in FinFET devices. These surfaces must be as smooth as possible for optimal device performance. Thus, the ability to quantitatively measure sidewall roughness is essential to process development. A methodology to quantitatively measure Fin sidewall roughness by AFM is presented. The samples were prepared for measurement by cleaving along the length of the Fins or dense-line test structures and by FIB polishing to bring the edge of the sample close to the sidewall of the etched feature. The cleaved and FIB-polished sample was mounted 17 degrees shy of normal. This exposes the sidewall on the top surface while preventing shadowing of the lower part of the sidewall due to contact between the side of the probe support and the cleaved or polished edge. Quantitative AFM measurements taken by this method show meaningful differences in the sidewall roughness for samples that have seen different sidewall smoothing treatments. The average observed rms roughness values for various surface-smoothing treatments range from 0.8 to 1.8 Å for a 50 nm square area.
http://www.mrs.org/s_mrs/sec_subscribe.asp?CID=2659&DID=106650&action=detail
Abstract
CD-mode AFM data contains information about the roughness of the line sidewalls. In this paper, sidewall roughness information from an in-line CD-mode AFM will be compared to AFM measurements taken on the same samples off-line using a destructive sample preparation. Correlation between in-line and off-line measurements shows that there are applications for which the CD-mode AFM could be used to gain information about sidewall roughness. With software and tip development, it is likely that the in-line technique will be useful for etch development applications. However, due to noise floor and tip shape limitations, applications involving smoother sidewalls require the off-line technique.
http://dx.doi.org/10.1109/ASMC.2004.1309570
Abstract
The versatility of a total reflection X-ray fluorescence (TXRF) spectrometer in the analysis of semiconductor samples will be demonstrated. While TXRF has a well-established place in trace metals analysis on silicon wafers, the practice of characterizing new films with TXRF is not routine in the semiconductor industry. In this paper, we will examine the monitoring of high-k film growth on silicon wafers by TXRF. We will show that a linear relationship between cycles of film deposition and TXRF signal is possible with proper analytical conditions. This signal can be converted to film thickness by normalizing to cross-sectional measurement from transmission electron microscopy (TEM). Information about the interface between the deposited high-k layer and the silicon substrate can also be determined from TXRF data. Secondary ion mass spectrometry data of a chlorine species at the interface of the high-k and silicon were collaborated with TXRF data. Critical angle measurements were taken on ruthenium and ruthenium dioxide films to extract physical characteristics and these results were compared to those from other techniques.
http://dx.doi.org/10.1016/j.sab.2004.04.009
Abstract
SIMS depth profiles obtained from a B implanted Si sample capped with a thin HfO2 film revealed the presence of several artifacts that make it problematic to determine the distribution and concentration of elements in the underlying Si substrate. The distribution of the B implant, normalized to 30Si+, was much broader than expected. Variations in the intensities of several matrix species under the HfO2 film were also observed. AFM measurements from the sputtered crater bottoms suggested that sputter-induced surface roughness is present under the high-k/Si interface. The roughness is partially responsible for reducing the sputter rate, which in turn lowers matrix ion intensities. The sputter rate variation can be corrected using the ratio of intensity loss of different matrix species. Use of the variation in the SiO+ intensity to correct the depth scale of the normalized B profile provided the best agreement with the original B implant-only sample. Rotating the sample while profiling improves the shape of the normalized B profile but some sputter rate correction was still necessary. The presence of 10% Hf (atomic) in the Si leads to slight reduction in the SiO+ intensity but the sputter rate does not need correcting. Under these conditions, point-by-point normalization of the B profile to a matrix species cancels out the ion yield variations. At Hf concentrations >10% the combination of ion yield and sputter rate variations complicates the depth profiles.
http://dx.doi.org/10.1016/j.apsusc.2004.03.099
Abstract
In this experiment we investigated two different techniques for carbon quantification and explored the differences in the measurements of volatile and nonvolatile carbon on the surface of wafers. The first technique, thermal desorption gas chromatography mass spectroscopy (TD-GCMS), quantifies the amount of volatile carbon that was absorbed on the surface of the wafer. The second technique, polyencapsulation SIMS, measures nonvolatile carbon or residual interface carbon that has been left behind after high temperature processing steps. Two sets of duplicate wafers were processed at International SEMATECH, one measured by TD-GCMS and the other set was capped with poly silicon and measured by dynamic SIMS. Both techniques produced similar results for the quantification of carbon, but the accuracy and ability to track the residual interface carbon was best shown by polyencapsulation SIMS.
http://dx.doi.org/10.1016/j.apsusc.2004.03.028
Abstract
Measuring HfO2/Si stacks-by secondary ion mass spectroscopy (SIMS) has become a common task in the semiconductor industry. Although the reliability of the SIMS-profiles from multilayer structures can be influenced by sputter yield and ionization yield variations in the interface region, the part of the dopant profile located beyond the interface can normally be obtained with excellent SIMS-characteristics.
By comparing the SIMS-profiles from B-profiles in Si, measured with and without a HfO2-cap layer, it becomes clear that the high k film has a strong impact on the resulting dopant profiles. With the high k film present, a significant distortion of the depth and intensity scale can be observed and the B-profile (which is, in principle, fully contained in the Si) appears almost twice as deep when profiling through the HfO2-layer as compared to the uncapped sample. Normalizing the depth and concentration scale with the variations of the SiO-signals, seems to lead to an acceptable B-profile. The distortion results from a strong reduction in sputter yield due to the presence of (even small) amounts of Hf. Apparently, Hf segregates towards the surface forming a (fractional) layer, which shields the Si-atoms from sputtering. The sputter yield of this Hf-layer on the other hand is very low leading to a drastic reduction in average erosion rate. For conditions where normally no full oxidation is obtained (high energy, non-normal incidence beams on Si and Ge-substrates), this reduction leads to a temporary enhancement of the oxygen incorporation and the ionization probability.
http://dx.doi.org/10.1016/j.apsusc.2004.03.102
Abstract.
Low-energy B, BF2, and As implants into 20 nm, 50 nm, and 150 nm Si layers on silicon-on-insulator wafers were investigated. Before annealing, a pileup of the B and As was observed at the Si/buried oxide interface in samples where the implant range and straggle approached the Si layer thickness. The absence of 10B from the interface, along with the presence of the pileup in profiles obtained from the back side of the samples, indicates that the pileup is implant induced. The pileup is believed to be due to the blockage of the open channels in the single-crystal Si by the amorphous SiO2 at the interface. In samples pre-amphorized with Ge the pileup is not observed. Examples of representative profiles will be presented along with a discussion of the methods used to reduce or eliminate several secondary ion mass spectrometry artifacts.
http://dx.doi.org/10.1116/1.1621401
Abstract
In this study, the nanoliter dried spot method was applied to semiconductor contamination analysis to enhance vapor phase decomposition processes with total reflection X-ray fluorescence detection. Nanoliter-sized droplets (10 and 50 nl) were deposited onto native silicon oxide wafer surfaces in a clean room environment from both single and multielemental standards containing various concentrations of iron in different matrices. Direct comparisons were made to droplets formed by conventional VPD with similar iron standards. Nanoliter dried spots could be reproducibly deposited and dried in air with typical drying times ranging from 20 s to 2 min depending on the nanoliter volume deposited, compared to VPD spots which have drying times ranging from tens of minutes to several hours. Both types of residues showed a linear relationship between Fe intensity and mass deposited. Variable angle experiments showed that both nanoliter and VPD deposits of single element standards were film-like in character, while residues formed from much more complex matrices and higher mass loadings were particulate in character. For the experimental conditions used in this study (30 kV, 100 mA), typical TXRF spectral Fe limits of detection were calculated to be on the order of picograms or 1×1010 atoms/cm2 for a 0.8 cm2 X-ray excitation beam area for both nanoliter dried spots and VPD spots prepared from single elemental standards. Calculated Fe detection limits for 200 mm diameter silicon wafers used in this study were in the 1×108 atoms/cm2 range. By using nanoliter sized droplets, the required sample volume is greatly reduced resulting in higher sample throughput than with conventional VPD methods.
http://dx.doi.org/10.1016/j.sab.2003.11.010
Abstract. Trace metal contamination from the bevel and edge exclusion region of various types of starting silicon was measured. Analysis of the data suggests that contaminating elements and their concentrations can vary between suppliers and by wafer type although aluminum was detected on all wafer types. Copper contamination results were lower for this quality of wafer than in previously experiments using test-quality wafers.
http://www.electrochem.org/dl/pv/published/2004/PV%202004-05_TOC.pdf
http://journals.cambridge.org/action/displayAbstract?fromPage=online&aid=165667
Abstract. Contamination around the edge exclusion area and bevel of silicon wafers is becoming an increasingly important area to control in semiconductor manufacturing. This is especially relevant in any manufacturing line where portions of the process toolset, e.g. metrology or lithography, are shared between multiple types of materials processed. This could be materials like copper versus aluminum for interconnect lines or more recently the newer high-k gate dielectrics and alternate metal electrodes. There are numerous pathways for contamination; for example, one source could be incomplete etching of a film at the edge during a backside-cleaning process. Also, the move to edge handling of wafers along with wafer alignment and centering pins is a contamination pathway. Even the direct contact between the wafer's edge and the cassette it is in can be a source of cross contamination. Historically, it has been difficult or impossible to quantify metal contamination in the bevel and edge exclusion region of silicon wafers by traditional analytical methods. Total reflection x-ray fluorescence spectroscopy cannot operate close to the edge of a wafer due to scattering affects of the incident radiation and typically has a built in edge exclusion area of 10 mm. Vapor phase decomposition would expose the entire wafer to hydrofluoric acid vapor, which would not be desired on a patterned wafer or a wafer with a film that would reacted with the vapor. Direct acid drop decomposition and vapor phase decomposition both employ a scanning technique where it is impossible to include the bevel area for analysis. Time of flight secondary ion mass spectrometry is also not easily done on the angled bevel edge of a wafer. The relative sample size of a time of flight mass spectrometer analysis is also quite small and therefore would require many analyses to achieve a sampling set that is representative of the entire edge of a wafer. We have developed a technique that will allow us to precisely measure the metallic contamination in this difficult region on both 200 mm and 300 mm wafers. This procedure calls for exactly positioning the wafer in a mechanical jig and collecting the contamination via chemical extraction. The amount of the wafer's edge exclusion analyzed is controllable as well as the ability to analyze the entire circumference of the wafer or any portion thereof. The solution sample is then analyzed for trace metals by inductively coupled plasma mass spectrometry. Knowing the concentration of the metals in solution, the mass of the solution, and the area of the wafer analyzed we will calculate the area concentration in atoms/cm2 for comparison to the traditional techniques mentioned in the previous paragraph. We will show the development of this technique along with data highlighting contamination control in a manufacturing line that processes multiple types of material. Detection limits of this technique and current challenges under development will be discussed. This technique is likely to become an indispensable part of any semiconductor fab’s analytical capabilities.
http://dx.doi.org/10.1117/12.485233
Abstract
It is shown that the focused ion beam in the Physical Electronics SMART-Tool can be used to create the conditions needed to locate interconnect defects by voltage-contrast analysis. The SMART-Tool is designed for the analysis of small defects on full wafers by Auger electron spectroscopy. These defects are typically located using a defect coordinate map from a light-scattering based inspection tool. The SMART-Tool can be equipped with a focused ion beam for cross-sectioning defects. Stand-alone focused ion beam tools have been used to locate defects by voltagecontrast analysis. Unlike stand-alone tools, the ion beam in the SMART-Tool is situated off the surface normal. This does not hinder its ability to ground interconnect parts to the substrate, creating the conditions for passive voltagecontrastimaging. A defective via chain, identified by high resistivity on parametric test, was grounded to the substrate by focused ion beam milling in the SMART-Tool. The defective via was then identified by voltage contrast in images. The defect was marked and cross-sectioned by the focused ion beam and analyzed by Auger electron spectroscopy, all in the SMART-Tool without breaking or repositioning the wafer. Studies suggest that unbroken wafers can be returned to the manufacturing line to complete processing after focused ion beam milling without compromising unaffected die. Thus, this type of interconnect defect analysis can be performed on defective die without sacrificing non-defective die on the same wafer.
http://dx.doi.org/10.1117/12.487627
Abstract
Depth profiling of thin gate dielectric films was studied. For SiON films profiled with 300 eV Cs+ at 75° roughening was not observed at 3 nm, but depth scale discrepancies suggest that roughening-induced sputter rate variations are present. Profiles of ZrO2 and HfO2 films sputtered with Cs show a unique behavior of sputter-induced roughness going through a maximum in the Si under the oxide film. This roughness influences the Cs concentration, which in turn affects the ion yields. Profiling the ZrO2 and HfO2 films with O2+ appears to be compromised by the presence of radiation-enhanced diffusion that leads to large decay lengths of the Zr+ or Hf+ signals.
http://dx.doi.org/10.1016/S0169-4332(02)00690-6
Abstract
Significant improvement in 157nm optical components lifetime is required for successful implementation of pilot and production scale 157nm lithography. To date, most of the 157nm optics lifetime data has been collected in controlled laboratory conditions by introducing predetermined concentrations of contaminants and monitoring degradation in terms of transmission loss. This publication compliments prior work by documenting field experience with the 157nm Exitech Microstepper currently in operation at International SEMATECH. Failure mechanisms of various optical components are presented and molecular contamination levels in purge gas, tool enclosure, and clean room are documented. Finally the impacts of contaminant deposition and degradation of components on imaging performance is discussed.
http://dx.doi.org/10.1117/12.474621
Abstract
A silver tube anode was used for total reflection x-ray fluorescence (TXRF) spectroscopy to analyze for zirconium contamination on silicon wafers. There was good correlation between results obtained by TXRF and results from vapor phase decomposition inductively coupled plasma mass spectrometry (VPD-ICP-MS) until the detection limits of the TXRF were reached at ~1E10 atoms/cm2. Calibration with either a nickel or a zirconium standard wafer was acceptable. The measurement precision using the silver anode was also acceptable (< 5% RSD).
http://www.icdd.com/resources/axa/VOL45/V45_80.pdf
41. “Blanket Metal Deposition and Etch Studies and Materials Characterization for Advanced Gate Electrode Application”, N. Chaudhary, J. Tamim, P. Lysaght, C. Sparks, M. Beebe, D. Baker, and C. Gondran, International SEMATECH technical transfer document. August (2002).
Abstract.
A magnetic sector secondary ion mass spectrometry (SIMS) instrument has been fitted with a modified hot filament duoplasmatron ion source for generation of SF5+ primary ion beams for SIMS depth profiling applications. The SF5+ primary ion beam has been evaluated by depth profiling of several low energy boron ion implants, boron delta-doped structures and a Ni/Cr metal multilayer depth profiling standard reference material. Using 3.0 keV impact SF5+ bombardment at a 52° impact angle with oxygen flooding gives a trailing edge decay length (1/e) for the boron implants and delta-doped layers of 1.3 nm. Under the same conditions, O2+ bombardment gives a trailing edge decay length (1/e) of 2.3 nm. The use of the SF5+ beam without oxygen flooding gives a substantial increase in decay length that is related to the formation of ripples as determined by atomic force microscopy. In the case of the Ni/Cr reference material, a significant reduction in sputter-induced topography is observed with SF5+ bombardment.
http://dx.doi.org/10.1116/1.591221
44. "Defect Reduction in Leading Edge Interconnect Equipment (CFMA001): Copper Cross Contamination Monitoring", M. Louis, R. Chowdhury, C. Gondran, C. Sparks, and F. Hafiz, SEMATECH TTID: 98033495BENG, Published Date: 03-Mar (1999).
Abstract:
A design-of-experiments methodology was implemented to assess the commercial equipment viability to fabricate the high-K dielectrics Ta 2O5, TiO 2 and BST (70/30 and 50/50 compositions) for use as gate dielectrics. The high-K dielectrics were annealed in 100% or 10% 02 for different times and temperatures in conjunction with a previously prepared NH 3 nitrided or 14N implanted silicon surface. Five metal electrode configurations-Ta, TaN, W, WN and TiN--were concurrently examined. Three additional silicon surface configurations were explored in conjunction with a more in-depth set of time and temperature anneals for Ta 2O5 . Electrical characterization of capacitors fabricated with the above high-K gate dielectrics, as well as SIMS and TEM analysis, indicate that the post high-K deposition annealing temperature was the most significant variable impacting the leakage current density, although there was minimal influence on the capacitance. Further studies are required, however, to clarify the physical mechanisms underlying the electrical data presented.
http://www.mrs.org/s_mrs/sec_subscribe.asp?CID=17418&DID=223738&action=detail
Abstract.
The electrical characteristics of shallow junction devices are often dependent on tightly controlled ion implants for their performance and yield. However, concerns exist over the ability of existing in-line metrology tools to accurately verify low energy implants, and therefore, determine whether the ion implanter is in control before committing product to it. This article will explore the capabilities of the Therma-Probe TP500, the Tencor OmniMap RS75, and secondary-ion-mass spectroscopy to measure low energy ion implants done on an Applied Materials xR-LEAP with energies 10 keV and below, doses of 1.0E15–3.0E15 cm – 2, and both boron and arsenic species.
http://dx.doi.org/10.1116/1.589818
Model-based FTIR characterization of epitaxial silicon measurements is a promising new technique that provides information about the epi film thickness and doping profile. This new method has many advantages, in terms of precision and accuracy, over the current commercial techniques based on the analysis of the interferogram. Because model-based analysis eliminates the need for matching reference wafers and empirical adjustments, it improves the intermachine measurement precision by a factor of 10, and the accuracy by a factor of 5 to 10. These improvements allow more consistent interpretation and fusion of thickness data from multiple measurement stations or laboratories, and reduce the need for destructive measurements.
Abstract.
Time-of-flight secondary ion mass spectrometry (TOF-SIMS) is an efficient, sensitive method for characterizing semiconductor surfaces. In addition, TOF-SIMS can be applied in a depth profiling mode allowing qualitative characterization of the top 20 nm of material. The utility of TOF-SIMS ultra-shallow depth profiling is demonstrated on GaAs substrates that were passivated with P2S5 solutions and oxidized by exposure to an UV/ozone treatment.
http://dx.doi.org/10.1116/1.587143
Abstract
Auger depth profiles are used to compare sputtered Ti, reactively sputtered TiN, and TiN/Ti bilayer films in submicron (0.8 and 0.4 µm) contact/via holes, deposited with and without a collimator. Samples were cleaved and mounted to minimize difficulties associated with the sputter profiling of and data acquisition from recessed features. The Ti and N concentrations were calculated with a subtract-and-weight routine which uses dN(E)/d(E) peak-to-peak intensities. Transmission electron microscopy film thicknesses measurements were used to confirm that the sputter rates were comparable for the various sample mounting geometries. The thickness of films deposited on the contact base, relative to those on the surrounding oxide was greater when the collimator was used. For 0.8 µm contacts, the film thickness ratio obtained using the collimator was nearly twice that obtained with the uncollimated deposition (0.5 versus 0.25 for TiN; 0.8 versus 0.6 for Ti). The benefits of the collimator, for obtaining good coverage of the base of the hole, are more pronounced in smaller geometries. For films deposited with the collimated on 0.4 µm contacts the film thickness ratio was ten times that for films deposited without the collimator (0.4 versus 0.05 for TiN, 0.35 versus 0.03 for Ti). There was no evidence of variations in the TiN stoichiometry for films deposited in confined geometries versus the surrounding oxide for either deposition configuration. There was an increase in the depth of surface oxidation for the Ti-only film deposited on the sidewalls using the collimator. Shadowing effects cause an increase in film porosity when the direction of collimated metal flux is parallel to the sample surface. On exposure to atmosphere, this porosity allows oxidation below the surface of the film.