Significant changes over time have been observed in surface and near surface phosphorus concentration for low dose phosphorus implants measured using secondary ion mass spectrometry (SIMS) with cesium bombardment. These variations in concentration affect the ability to make reproducible dose measurements. Phosphorus measurements have been documented for samples from wafers kept at ambient conditions and also for those stored in a range of other conditions including heat, high humidity, low humidity, and liquid nitrogen. An initial study of wafers ion implanted over a range of doses showed that the change in the surface phosphorus concentration was most apparent for the lowest phosphorus dose (1×1012 atoms/cm2) and that heating the sample resulted in the most significant change (increase in almost three orders of magnitude) in apparent surface P concentration. Only the specimens stored at liquid nitrogen temperature showed no change in P surface concentration. SIMS analysis conditions were optimized and a second set of analyses were performed on wafers that included different wafer processing (thicker surface oxide, preamorphization, and anneal) in an effort to reduce the change in surface phosphorus. Significant improvement in P dose reproducibility was noted for the wafer that had been implanted through a 5 nm oxide and stored in dry conditions.

Advanced lithography requirements have driven pellicle suppliers to develop improvements in both structure and materials. For example, thinner pellicle membranes now improve transmission of light, shortened aluminum frames now reduce image shadowing and advanced adhesives now reduce outgassing. The quality control and evaluative techniques employed to develop and monitor these improvements include ion chromatography (for cation and anion), gas chromatography-mass spectrometry (for organic volatile and semi-volatile compounds) and ultraviolet-visible spectroscopy (for film transmission). As pellicle materials continue to evolve and diversify, additional analysis methods are essential to solve problems both in the development and semiconductor manufacturing phases. This paper summarizes a body of work that was completed in a manufacturing environment in response to an ongoing contamination issue ultimately traced to a particular KrF pellicle supplier. This contamination adversely impacted subsequent pellicles and photomasks undergoing the pellicle demounting process. In an attempt to identify the root source of the contamination, pellicles exhibiting successful demounting characteristics were compared to those that induced unacceptable contamination levels. Direct chemical analysis of respective mounting adhesives via time-of-flight secondary ion mass spectrometry identified mixtures of chemical components unique to each adhesive formulation. This information was fed back to the pellicle vendor who ultimately acknowledged an adhesive formulation change from a styrene-type to an acryl-type compound, after which the resulting pellicle product quality was verified and led to the return of contamination levels to within acceptable limits.

Passive voltage contrast (PVC) is a well-known fault isolation technique in differentiating contrast at via/metal/contact levels while focused ion beam (FIB) is a destructive technique specifically used for cross sectioning once a defect is identified. In this study, we highlight a combination technique of PVC and progressive FIB milling on advanced node fin field-effect transistor (FinFET) for root cause analysis. This combo technique is useful when applied on high-density static random access memory (SRAM) structure, especially when it is difficult to view the defect from top-down inspection. In this paper, we create a FA flow chart and FIB deposition/milling recipe for SRAM failure and successfully apply them to three case studies.

Articles you may be interested inMechanism of β-FeSi2 precipitates growth-and-dissolution and pyramidal defects' formation during oxidation of Fe-contaminated silicon wafersMonitoring metal contamination of silicon by multiwavelength room temperature photoluminescence spectroscopy AIP Advances 2, 042164 (2012); 10.1063/1.4769746The effect of oxygen during irradiation of silicon with low energy Cs + ions Highly sensitive time-of-flight secondary-ion mass spectroscopy for contaminant analysis of semiconductor surface using cluster impact ionization Appl. Phys. Lett. 86, 044105 (2005); 10.1063/1.1852715 Metallic contamination in hydrogen plasma immersion ion implantation of siliconIn order to improve the understanding of unintended cesium (Cs) contamination that occurs during SIMS depth profiling, Cs concentrations on sample surfaces were measured before analysis and at various distances from a depth profile crater after analysis. Cs concentrations in excess of 1 at. % were found directly adjacent to the depth profile analysis site. Cs was also detected at significant concentrations hundreds of micrometers from the depth profile measurement location. This Cs contamination can originate from a number of sources including Cs beam tails, Cs neutral beam, and secondary sputtering from instrument optics and other structures. Since the presence of cesium significantly affects the secondary ion yield of electronegative elements (e.g., phosphorus) in silicon, the unintended presence of cesium on the surface of a previously analyzed sample can strongly affect the reproducibility and accuracy of low dose electronegative element measurements, especially at the surface.

For a specific IDDQ failure only around SRAM cell boundary, we conducted a systematic investigation in the lab involving electrical, physical, and chemical analysis. Following electrical test locating the failure area according to PEM (photon emission microscopy) and physical defect analysis resulting in NDF (no defect found), we explored an alternative method to define the failure. In this paper, we demonstrated the success of using tunneling AFM (TUNA) in diagnosing such an IDDQ failure occurring in FinFET devices. AFM (TUNA) analysis was able to visualize clearly the dopant discrepancies in comparison between the IDDQ fail and pass references in FinFET transistors. The dopant abnormalities indicated the current IDDQ fail was caused by processes that impaired the dopant implantation.