Diffusion Process Research Articles

A New Brain Network Construction Paradigm for Brain Disorder via Diffusion-Based Graph Contrastive Learning.

Brain network analysis plays an increasingly important role in studying brain function and the exploring of disease mechanisms. However, existing brain network construction tools have some limitations, including dependency on empirical users, weak consistency in repeated experiments and time-consuming processes. In this work, a diffusion-based brain network pipeline, DGCL is designed for end-to-end construction of brain networks. Initially, the brain region-aware module (BRAM) precisely determines the spatial locations of brain regions by the diffusion process, avoiding subjective parameter selection. Subsequently, DGCL employs graph contrastive learning to optimize brain connections by eliminating individual differences in redundant connections unrelated to diseases, thereby enhancing the consistency of brain networks within the same group. Finally, the node-graph contrastive loss and classification loss jointly constrain the learning process of the model to obtain the reconstructed brain network, which is then used to analyze important brain connections. Validation on two datasets, ADNI and ABIDE, demonstrates that DGCL surpasses traditional methods and other deep learning models in predicting disease development stages. Significantly, the proposed model improves the efficiency and generalization of brain network construction. In summary, the proposed DGCL can be served as a universal brain network construction scheme, which can effectively identify important brain connections through generative paradigms and has the potential to provide disease interpretability support for neuroscience research.

Integrated impacts of surrounding building density and gap permeability on pollutant dispersion in four-way street networks.

Gaps between buildings facing the streets can effectively accelerate the natural removal of pollutants in street canyons by enhancing ventilation and diffusion processes. This removal process is closely related to gap permeability (Pg, ratio of gap width to street length) and building density (λn) surrounding the streets. However, the integrated effects of these two parameters on pollutant dispersion remain unclear, because of the limited computational resources and the difficulty of traditional modeling methods to discrete the numerous buildings. Numerical simulations coupled with the porous media approach are performed for this purpose. Results indicate that for isolated streets surrounded by empty land (λn = 0), the high-speed airflow through the gaps causes significant flushing effects on pollutants in the street, and thus larger Pg induces better street air quality than the case without gaps (Pg = 0). Conversely, when the streets are surrounded by buildings (λn ≠ 0), the air quality is good for both smaller Pg = 0 and larger Pg = 52 %, and the worst case occurs in Pg = 24 %. Moreover, because of the blocking effect of surrounding buildings, the percentage of pollutant removal by the gaps decreases with increasing λn, which can reach up to 50 % for λn = 0, decreasing to 30 % for medium density (λn = 0.22). Notably, for high density (λn = 0.44), only when Pg > 52 % the pollutant concentration can be reduced compared to Pg = 0. The present findings can serve as principles for determining appropriate building gap layouts at different building densities, thereby improving the urban air quality more effectively.

Vision-guided robot for automated pixel-level pavement crack sealing

Automated pavement crack sealing plays a crucial role in road maintenance. However, challenges remain in refining crack segmentation and sealing control accuracy. This article proposes an automated pavement crack sealing robot(APCSbot), which employs a crack refinement network(CrackSegRefiner) and a crack sealing controller(LQR). Specifically, CrackSegRefiner is based on a denoising diffusion model to refine the coarse mask of crack through a diffusion process. Additionally, the LQR controller integrates weight matrices Q and R to ensure control and state of APCSbot based on visual servo, facilitating the delivery of emulsified asphalt for sealing through the end effector. Extensive experiments conducted on the DeepCrack, CFD, and S2TCrack datasets confirm the effectiveness of APCSbot, which achieved a segmentation precision of 84.48% and mIoU of 79.28%. Furthermore, the system demonstrated a sealing error of 6.22 mm and speed of 0.0456 m/s when addressing discontinuous cracks, showcasing its excellence and robustness in crack sealing.

Process and mechanism of Zn and Co leaching from cobalt xanthate slag enhanced by oxidative roasting

Cobalt xanthate slag is the waste slag produced by the use of xanthate cementation and decontamination before the wet refining and electrowinning of Zn, in which Zn, Co, Cd, etc., and butyl xanthate exist in the form of hydrophobic complexes, which are poorly soluble with water. In addition, most of the cobalt xanthate in the slag exists in the form of high-valent Co, which lowers the extraction rate when directly leaching the metal elements. In this study, the atmosphere of the roasting process is regulated to destroy the hydrophobic complexes and change the mineral phase structure of the cobalt xanthate slag. In the cobalt xanthate slag, the Co and Fe valence states change during the roasting and leaching steps; specifically, during the roasting process, the complexes first decompose into CS2, reducing part of the Co3+. Following the disappearance of this sulfur-rich environment, the oxygen in the flowing air oxidises Fe2+ to Fe3+, but the valence state of Co remains unchanged. During the subsequent leaching process, Co3+ and Fe2+ further react to realise the benign transformation of Co3+ to Co2+ and Fe2+ to Fe3+ without an added redox agent. The rapid redox reaction between Co3+ and Fe2+ controls Co3O4 dissolution; thus, the reaction rate is controlled by the slower process of internal diffusion. Under the optimal roasting conditions, the Co3+ content in cobalt xanthate slag decreases to 21.10% and that of Fe2+ to 48.78%. Consequently, the optimum Zn and Co leaching rates reach 99.92% and 99.57%, respectively, and only 0.86% of Fe in the final leaching solution exists in the form of Fe2+.

Reverse stable diffusion: What prompt was used to generate this image?

Text-to-image diffusion models have recently attracted the interest of many researchers, and inverting the diffusion process can play an important role in better understanding the generative process and how to engineer prompts in order to obtain the desired images. To this end, we study the task of predicting the prompt embedding given an image generated by a generative diffusion model. We consider a series of white-box and black-box models (with and without access to the weights of the diffusion network) to deal with the proposed task. We propose a novel learning framework comprising a joint prompt regression and multi-label vocabulary classification objective that generates improved prompts. To further improve our method, we employ a curriculum learning procedure that promotes the learning of image-prompt pairs with lower labeling noise (i.e. that are better aligned). We conduct experiments on the DiffusionDB data set, predicting text prompts from images generated by Stable Diffusion. In addition, we make an interesting discovery: training a diffusion model on the prompt generation task can make the model generate images that are much better aligned with the input prompts, when the model is directly reused for text-to-image generation. Our code is publicly available for download at https://github.com/CroitoruAlin/Reverse-Stable-Diffusion.

Mathematical modeling of bioactive glass degradation

Bioactive glasses (BGs) are promising for bone tissue engineering (BTE). Mathematical modeling is a powerful tool for understanding BTE scaffold degradation. We developed mathematical functions based on chemical reaction equations governing dissolution and diffusion processes to model the degradation of 45S5 BGs. An empirical mathematical model was employed to characterize the formation process of hydroxycarbonate apatite (HCA). Two sets of numerical simulations with BG powder and bulk samples immersed in simulated body fluid were compared with in vitro experiments to validate and parameterize the model. The model could accurately predict BG degradation and HCA formation. Our findings indicate that the proposed parameters K1=2600mm/(μmol·h), K2=2.0 mm/h and K3=0.001mm/h are suitable for simulating the degradation of silicate-based BGs, specifically 45S5 BG. The proposed model successfully predicted the degradation behavior and subsequent HCA formation over 21 days. The proposed mathematical model serves as a valuable tool for designing degradable BG-containing scaffolds.

Organic-inorganic hybrids cathode with Hydrogen Bonding Network for highly efficient zinc-ion batteries

Organic-inorganic hybridization has been explored to improve the electrochemical performance of electrodes for aqueous zinc-ion batteries. In this work, 4,4′-diamino-2,2′-bipyridine (DB) and 2,6-diaminoanthraquinone (DAAQ) intercalated vanadium pentoxide (HVO-DB and HVO-DAAQ) are prepared by pre-intercalation techniques. The incorporation of DAAQ and DB play a pivotal role in not merely enhancing the interlayer spacing of VOx slabs substantially, which consequently elevates the efficiency of ion diffusion processes, but also in reinforcing the overall structural stability through promoting the establishment of a robust hydrogen bond network. Moreover, the abundant CO redox-active sites present in DAAQ actively coordinate with Zn2+ during the insertion process, leading to a notable enhancement in specific capacity. HVO-DAAQ demonstrates an exceptional specific capacity of 395 mA h g−1 when charged at 0.1 A g−1, coupled with remarkable cycling stability, retaining 87.8 % of its capacity even after enduring 2000 cycles at a high rate of 5 A g−1, and superior rate performance. Ex-situ characterization reveals the exceptional reversibility of Zn2+ insertion, alongside the remarkable capacity of hydrogen bond networks to undergo disruption and subsequent reconstruction, showcasing a dynamic and resilient structural adaptability. Furthermore, this work provides insights into the synergistic energy storage mechanism in the organic-inorganic hybrids cathode for aqueous zinc-ion batteries.

A trade-off between the diffusion depth, the thickness of the Tb-rich shell and the surface grain coarsening during the grain boundary diffusion of sintered Nd-Ce-Fe-B magnets

The grain boundary diffusion process (GBDP) has become one of the main methods to enhance the coercivity of Nd-Ce-Fe-B magnets. In this study, we examined how the magnetic properties of sintered Nd-Ce-Fe-B magnets are influenced by the combined impacts of diffusion depth, Tb-rich shell thickness, and surface grain coarsening after conducting grain boundary diffusion. There exists a trade-off between achieving a desired diffusion depth and avoiding excessive surface grain coarsening. To examine this trade-off, samples with varying diffusion depths were prepared through controlled diffusion time. Results revealed that compared to the original annealed magnets, the coercivity increments of the magnets diffused for 1 h and 3 h were 148 kA/m and 290 kA/m, respectively, while the coercivity of the magnet diffused for 9 h remained nearly the same as that diffused for 3 h. Microstructural analysis indicated that surface grain coarsening intensified with increasing diffusion time, leading to a reduction in the surface diffusion channels, thereby diminishing diffusion efficiency. In addition, strong mutual diffusion was observed between the magnet and the diffusion source. Furthermore, micromagnetic simulation studies revealed that severe surface grain coarsening limits the enhancement of coercivity even with increased depth of diffusion and thickness of the Tb-rich shell layer. This study offers valuable insights into the correlation between diffusion depth, Tb-rich shell thickness, surface grain coarsening, and the ultimate magnetic properties in sintered Nd-Ce-Fe-B magnets after GBDP, providing guidance for enhancing the efficiency of GBDP.

A signal-diffusion-based unsupervised contrastive representation learning for spatial transcriptomics analysis.

Spatial transcriptomics allows for the measurement of high-throughput gene expression data while preserving the spatial structure of tissues and histological images. Integrating gene expression, spatial information, and image data to learn discriminative low-dimensional representations is critical for dissecting tissue heterogeneity and analyzing biological functions. However, most existing methods have limitations in effectively utilizing spatial information and high-resolution histological images. We propose a signal-diffusion-based unsupervised contrast learning method (SDUCL) for learning low-dimensional latent embeddings of cells/spots. SDUCL integrates image features, spatial relationships and gene expression information. We designed a signal diffusion microenvironment discovery algorithm, which effectively captures and integrates interaction information within the cellular microenvironment by simulating the biological signal diffusion process. By maximizing the mutual information between the local representation and the microenvironment representation of cells/spots, SDUCL learns more discriminative representations. SDUCL was employed to analyze spatial transcriptomics datasets from multiple species, encompassing both normal and tumor tissues. SDUCL performed well in downstream tasks such as clustering, visualization, trajectory inference, and differential gene analysis, thereby enhancing our understanding of tissue structure and tumor microenvironments. https://github.com/WeiMin-Li-visual/SDUCL. Supplementary data are available at Bioinformatics online.

Theoretical Model and Experimental Application for Solute Diffusion in Leaching Test With Time-Dependent Boundary Conditions

Abstract Measuring the diffusion coefficient of clay-based liner materials is important in estimating and predicting long-term barrier performance in waste containment facilities. Various theoretical models, including the finite cylindrical model, have been commonly used to determine the diffusion properties of clay-based liner materials in leaching tests. However, the assumption of zero-concentration boundary conditions of the traditional finite cylindrical model contradicts the measured variation of concentration in real leaching tests, likely resulting in (1) underestimated and unconservative diffusion coefficient, or (2) requirement of a relatively large liquid-to-soil ratio and frequent leachate replacement in the experiment to maintain the zero-concentration boundary condition. In this study, a theoretical model was developed to evaluate the solute diffusion process within a soil specimen under arbitrary, time-dependent concentration boundary conditions. The proposed model, incorporating the time-dependent boundary conditions, provides efficient calculations of the concentration distribution and the cumulative fraction leached of solute across the soil specimen. The example application of the proposed model to experimental data demonstrates the capability of the proposed model to determine apparent diffusion coefficients of clay-based liner materials without introducing errors associated with the assumption of a zero concentration boundary condition. The proposed model provides a comprehensive method to investigate the dynamic transport behaviors of solutes through clay-based liner materials in future studies.

The paradox of product scarcity: Catalyzing the speed of innovation diffusion

AbstractProduct shortages are known to slow down the diffusion process. However, we counterintuitively theorize and empirically demonstrate that under specific conditions of social influence, the diffusion process may be accelerated by early product scarcity. Using an Agent-Based framework and Genetic Algorithm-based estimation, we analyzed 20 product categories to identify the critical trade-off influencing diffusion: the interplay of the social influence ratio of waiting customers to adopters, the external influence, and level of product scarcity. Strategic managerial actions can accelerate the adoption of products. For example, in the case of fitness trackers, we were able to simulate speed-up by up to two years compared to the standard Bass model. Importantly, we introduce a novel framework to study competition dynamics, analyzing how the timing of market entry and the production capacity of competitors, along with the initial installed capacity of the pioneering firm affect diffusion speed. This acceleration, whether due to managerial foresight or serendipity, necessitates careful orchestration to harness the enthusiasm of waiting customers and strategically allocate marketing spending on social media platforms, thereby differentially amplifying the influence of adopters and potential customers.

Development of evaluation method for radiocesium availability in soil by biomimetic approach.

Applicability of biomimetic approach with simulation of plant uptake for assessment of radiocesium availability in soil was investigated. The soil spiked with 137Cs tracer was contacted with wicking material and copper-substituted prussian blue (Cu-PB), which simulate transpirationally induced mass flow and concentration gradient-induced diffusion of radiocesiumin the soil, respectively. Comparison of the removed 137Cs to the wick and the wick + Cu-PB from the soil during the contact period of 12weeks suggested that the diffusion process has larger contribution than the mass flow process in radiocesium dynamics in root zone. The change of the removed rate of 137Cs from the soil was reflected that its availability decreased with the time elapsing and with subjecting repeated wet-dry treatment. The results suggest that the biomimetic approach can be applicable to the realistic evaluation of the availability of radiocesium in soil.

Photoassisted electron emission from planar-type electron source based on graphene/oxide/silicon structure

This paper reports on photoassisted electron emission from a planar-type electron emission source based on a graphene/oxide/p-Si structure under visible laser light irradiation. The electron source is expected to be useful as a photocathode with a high quantum efficiency and an ultrafast pulsed electron beam. The electron emission efficiency is found to be 12%, independent of laser light irradiation. Without a negative electron affinity surface, the emission current generated by irradiation shows an increase of several orders of magnitude compared with that in the dark, and a quantum efficiency of 0.3% is achieved. This electron source exhibits advanced photoassisted electron emission characteristics with high photosensitivity in the order of milliamps per watt. The photoassisted emission from the device shows a photoresponse with a rise and fall time of 70 μs at a wavelength of 633 nm, which is determined by the diffusion process of photoexcited electrons in the bulk. The use of a heavily doped p-type silicon substrate provides a practical route for further improvement.

Diffusion Behavior and Kinetics for the Vapor Phase Infiltration of Trimethylaluminum in Poly(ethylene oxide): An In Situ Quartz Crystal Microgravimetry Study.

Vapor phase infiltration (VPI) facilitates the incorporation of inorganic components into organic polymers, emerging as an effective technique for fabricating organic-inorganic hybrid materials. However, the complexity of diffusion behavior during the VPI process presents challenges in studying diffusion kinetics, particularly for highly reactive precursor-polymer systems such as trimethylaluminum (TMA) and poly(ethylene oxide) (PEO). In this study, we investigate the VPI process of TMA in PEO using in situ quartz crystal microgravimetry (QCM), which enables measurement of diffusion behavior and kinetics with high precision due to its high temporal resolution. Our results indicate that the VPI process consists of two main regions: a rapid diffusion process, corresponding to the initial penetration of the precursor into the film, followed by a slower relaxation process, attributed to the ongoing chemical reaction. The equivalent diffusion coefficient (De) was estimated to be on the order of 10-9 cm2/s and decreased with increasing aluminum content. Using energy application as a proof-of-concept, when optimized, VPI-modified PEO films were successfully utilized as solid polymer electrolytes (SPEs) for lithium metal batteries (LMBs), showcasing superior performance in mitigating lithium dendrite growth. This study offers valuable insights into the VPI process for PEO-TMA systems and provides guidance for optimizing VPI conditions to enhance the performance of advanced materials.

Toward a Universal Model of Hyporheic Exchange and Nutrient Cycling in Streams

AbstractIn this paper we demonstrate that several ubiquitous hyporheic exchange mechanisms can be represented simply as a one‐dimensional diffusion process, where the diffusivity decays exponentially with depth into the streambed. Based on a meta‐analysis of 106 previously published laboratory measurements of hyporheic exchange (capturing a range of bed morphologies, hydraulic conditions, streambed properties, and experimental approaches) we find that the reference diffusivity and mixing length‐scale are functions of the permeability Reynolds Number and Schmidt Number. These dimensionless numbers, in turn, can be estimated for a particular stream from the median grain size of the streambed and the stream's depth, slope, and temperature. Application of these results to a seminal study of nitrate removal in 72 headwater streams across the United States, reveals: (a) streams draining urban and agricultural landscapes have a diminished capacity for in‐stream and in‐bed mixing along with smaller subsurface storage zones compared to streams draining reference landscapes; (b) under steady‐state conditions nitrate uptake in the streambed is primarily biologically controlled; and (c) median reaction timescales for nitrate removal in the hyporheic zone are 0.5 and 20 hr for uptake by assimilation and denitrification, respectively. While further research is needed, the simplicity and extensibility of the framework described here should facilitate cross‐disciplinary discussions and inform reach‐scale studies of pollutant fate and transport and their scale‐up to watersheds and beyond.

Analysis of Macromolecular Size Distributions in Concentrated Solutions

AbstractThe solution state of macromolecules in concentrated solutions impacts fields ranging from cell biology, to colloid chemistry and engineering of protein pharmaceuticals. Dependent on the interplay between repulsive and weakly attractive forces, proteins may exhibit oligomerization, aggregation, crystallization, liquid‐liquid phase separation, or the formation of multiprotein complexes. The particle size‐distribution is a key characteristic, but difficult to determine when interparticle distances are on the order of their size and macromolecular motion is coupled through hydrodynamic interactions. Here we extend sedimentation velocity analytical ultracentrifugation to measure macromolecular size distributions under these conditions: We apply results from statistical fluid mechanics for the concentration‐dependence of hindered settling and diffusion, embedded in a mean‐field approximation that can resolve coupled sedimentation and diffusion processes of different sized species given experimental sedimentation data. This is combined with a description of transient optical aberrations from lensing in the refractive index gradients associated with sedimentation boundaries (Wiener skewing). We demonstrate this approach in the application to protein solutions with macromolecular volume fractions up to ≈10 %, for example, resolving monomers and dimers of albumin at 140 mg/ml. This enables size‐distribution analysis of proteins at concentrations of therapeutic antibody formations and close to physiological concentration in serum and cells.

Abstract 4135080: Relation of focal and diffuse myocardial fibrosis to left bundle branch block in patients with dilated cardiomyopathy and structurally normal heart

The role of fibrosis in etiology of left bundle branch block (LBBB) in still not defined. The aim of this study was to assess the relation of focal and diffuse LV fibrosis to LBBB in patients with dilated cardiomyopathy (DCM) and in patients with structurally normal heart. Materials and methods: 60 DCM with LBBB (DCM-LBBB group, 46.7% male, mean age 57[50;63] years), 50 non-LBBB DCM (78% male, mean age 43[32.3;54] years), 15 patients (53.3% male, mean age 39[28;50] years) with LBBB and no signs of structural heart disease and 10 healthy volunteers (HV) were included in the study. Endomyocardial biopsy (EMB) was performed in 15(24,6%) LBBB-DCM and 16(32%) non-LBBB DCM patients and allowed the quantification of collagen volume fraction (CVF, Fig.1A-D). All patients and HV underwent cardiovascular magnetic resonance (CMR) with late gadolinium enhancement (LGE) analysis. Obtained LGE LV images were post-proceed for core scar (CS) and gray zone (GZ) calculation and 3D reconstruction. Diffuse LV interstitial fibrosis was estimated on all LGE-CMR images with the novel introduced marker - diffuse intensity ratio (DIR). Results: DIR was validated on EMB samples: the percentage of CVF (8.2[4.6;10.8]%) correlated with the DIR value (0.71 ± 0.13) in the same segment (r=0.66, p<0.001, Fig.1E-F). The value of CVF (Fig1A-D) and LGE presence in the septum in both DCM groups (Fig2) was comparable (8.2[4.6; 10.3] versus 7.6[6.1;15.3]%, p=0.82 and 19 (31.7%) versus 17 (34%) patients in non-LBBB DCM group respectively, p=0.8) Moreover, the presence of LGE and midwall fibrosis in septum was not related to the QRS duration (p=0.93 and 0.59). In non-LBBB DCM patients the percentage of CS was significantly higher (4.0[1.55; 11.68]% versus 1.4[0.05;8.5]% in LBBB-DCM patients, p=0.047), whereas percentage of GZ and total fibrosis in both DCM groups was comparable (Fig.3A,B). DIR value was higher in patients with idiopathic LBBB than in HV (0.54±0.09 versus 0.34±0.1, p<0.001) while the percentage of GZ and CS was extremely low and did not differ from its amount in HV (Fig.3C,D). Conclusion: Neither focal (including midwall striae) nor interstitial fibrosis is associated with LBBB in patients with DCM. Diffuse inflammation in LBBB-DCM patients may contribute to the progression of systolic dysfunction but is not the reason for LBBB. The increased value of interstitial fibrosis in patients with LBBB may reflect latent diffuse process in myocardium inexorably leading to DCM development.

Charting the Unknown: First Passage Time Probabilities for Pearson Diffusion Process and Application to Options Risk Management

The first passage time probabilities have applications in many fields, including Finance, Marketing, Economics, Physics, and Statistics. In this paper, we study the first passage time probabilities for a Pearson diffusion process and obtain the lower and upper bounds of the first passage time density. We show that the density may be approximated by the upper bound with an error of approximately five percent. We present an application by modelling the profit and loss function of the S&P 500, FTSE 100 and DAX 40 index options using a Pearson diffusion process. Further, we establish the relation between first passage time probabilities and MaxVaR, i.e., the intra-horizon risk and obtain the MaxVaR for various index options based on first passage time probabilities. This is important as MaxVaR can capture the risk and potential losses incurred at any time of the trading horizon. In addition, we conduct a sensitivity analysis on the parameters for the purpose of robustness.

Chemically reactive and aging macromolecular mixtures. II. Phase separation and coarsening.

In a companion paper, we put forth a thermodynamic model for complex formation via a chemical reaction involving multiple macromolecular species, which may subsequently undergo liquid-liquid phase separation and a further transition into a gel-like state. In the present work, we formulate a thermodynamically consistent kinetic framework to study the interplay between phase separation, chemical reaction, and aging in spatially inhomogeneous macromolecular mixtures. A numerical algorithm is also proposed to simulate domain growth from collisions of liquid and gel domains via passive Brownian motion in both two and three spatial dimensions. Our results show that the coarsening behavior is significantly influenced by the degree of gelation and Brownian motion. The presence of a gel phase inside condensates strongly limits the diffusive transport processes, and Brownian motion coalescence controls the coarsening process in systems with high area/volume fractions of gel-like condensates, leading to the formation of interconnected domains with atypical domain growth rates controlled by size-dependent translational and rotational diffusivities.

Microstructural Aspects on Brazing Stainless Steels for High Temperature Applications

A wide range of applications request components that work in an environment where they are subject to high temperatures, combined with a corrosive action of the same environment. The components with a complex geometry can be obtained only using some assembling processes, among which are welding and brazing. One of the most important advantages of these processes is the possibility to obtain a sealed joint, that guarantees it to be waterproof. In comparison with welding process, brazing process has also an important technological advantage: it is a process that is able to produce a quality joint in a very small, narrow and tight places., where is very difficult or almost impossible to reach with other welding processes (e.g. GMAW/MIG – Gas Metal Arc Welding, GTAW/TIG – Gas Tungsten Arc Welding, SMAW – Shield Metal Arc Welding, FCAW - Flux Cored Arc Welding, SAW – Submerged Arc Welding). The research in this direction carried out in University Politehnica Timisoara, was focused on using brazing as a joining process to obtain a complex geometry part working in these environments. The brazed joint will create a dissimilar joint, putting in contact stainless steel with a Ag-Cu brazing alloy, which creates diffusion processes at microstructural level as well as phase transformations (due to thermal cycle and diffusion) which have a large impact on operating behavior of these joints. This paper presents some results on investigation phase and microstructural constituents transformations that took place in these brazed joints.