No abstract

We report a thin device layer ( -l p ) framfer technique that allows wafer scale fransplanting of firllyfabricafed circuits ofSOI on to a hosf subsfrate fo produce 3-D integrated circuifs. This 3-D Parallel Lqxring Process (3-0 PLP) uses temperature below 330 C and incorporates BCB as the dielectric bonding layer. The fechnique is parficularly suitable for 3-0 mired-signal or heterogeneous infegration applications where digital and RF/analog circuits benefitfiom separafe manufacturing. Device layer to l q e r alignment of 3 p is demonstrafed Elechical measurement of the tramisfors on the SOI donor n,afer before and afrer fransfer process is presented. IntroductionThere has been great interest in 3-D integration due to its numerous performance advantages as 2-D device scaling reaches its limits [1],[2]. The technique of 3-D integration allows dense integration of CMOS circuits with planar transistors in multiple layers that can be integrated vertically. This permits novel integration approaches and applications. 3-D mixedsignal and 3-D heterogeneous technology integration, where different types of circuit implementations on different materials can be integrated [3], are two examples of directions possible. Two distinct approaches for 3-D integration exist to date: sequential processing where single crystal silicon is

We describe a low-temperature polymer-based 3-D integration technique for wafer-scale transplantation of micrometer thick circuit and device layers onto another host wafer. The maximum temperature of this approach is 340 • C. It incorporates a low-k semiconductor compatible dielectric bonding media, employs tools that are readily available within a fabrication environment, and is very simple to implement. Another unique characteristic of the approach is the simultaneous separation of the transplanting layer from the donor assembly with the bonding to the host assembly. Alignment registration of several micrometers between device layers is demonstrated. Electrical results of 3-D inverter circuit along with demonstration of four-device-layer 3-D integrated stack are presented. Index Terms-Silicon-on-insulator (SOI), system-on-chip integration, 3-D IC, 3-D integration.

chemoradiation for anal cancer. Multiple studies have suggested that irradiation of the lumbosacral spine (LSBM) may be the primary cause of HT. We hypothesized that irradiation of the LSBM may be a superior predictor of absolute neutrophil count (ANC) nadir than the PBM or other anatomically-defined subsites. Materials/Methods: The cohort included 45 patients with non-metastatic anal cancer who were treated with definitive chemoradiation with mitomycin-C and 5-fluorouracil. PBM was defined as the external contour of the pelvic bones from the top of the L5 vertebral body to the bottom of the ischial tuberosity. PBM was divided into three subsites: LSBM included the entire sacrum and L5 vertebral body, iliac subsite (IBM) extended from the iliac crests to the superior border of the femoral head, and lower pelvis (LPBM) included the pubic bones, ischia, acetabula, and proximal femora. The primary endpoint was the absolute neutrophil count (ANC) nadir during or within 2 weeks of completion of treatment. Generalized linear modeling was used to analyze the correlation between the equivalent uniform doses (EUD) with "a" value of 0.5 to the PBM, LSBM, IBM, and LPBM structures and the ANC. Age, BMI, sex, baseline ANC count, immunosuppression were analyzed as potential covariates. Results: Mean ANC nadir was 0.77 x 10 9 /L (standard deviation: 0.66 x 10 9 /L). Grade 3 and 4 ANC toxicity occurred in 26.7% and 44.4% of patients, respectively. Baseline ANC was the only significant covariate and was included in all models. EUD of PBM, LSBM and IBM were significantly associated with ANC nadir (Table). Model performance (R2) was not superior for the LSBM compared to the IBM and PBM models. Grade 4 ANC toxicity was significantly associated with the EUD of PBM (odds ratio per Gy (OR)): 1.04; P Z 0.038), LSBM (OR: 1.02; P Z 0.029), and IBM (OR: 1.04; P Z 0.035). Conclusion: The LSBM model was not superior to the PBM model, likely because of the functional importance of the IBM region. Our results support further investigation of techniques designed to reduce irradiation of both the LSBM and IBM subsites. Whether the LPBM should be considered functionally important for avoidance of HT requires further study.

No abstract