rXXXX American Chemical Society A dx.doi.org/10.1021/es103486s | Environ. Sci. Technol. XXXX, XXX, 000–000 ARTICLE pubs.acs.org/est Increasing Secondary and Renewable Material Use: A Chance Constrained Modeling Approach To Manage Feedstock Quality Variation Elsa A. Olivetti,* ,† Gabrielle G. Gaustad, ‡ Frank R. Field, † and Randolph E. Kirchain † † Materials Systems Laboratory, Engineering Systems Division, Massachusetts Institute of Technology, 77 Massachusetts Avenue, Cambridge, Massachusetts 02139, United States ‡ Golisano Institute of Sustainability at Rochester Institute of Technology, 111 Lomb Memorial Drive, Rochester, New York 14623, United States b S Supporting Information ’ INTRODUCTION It is estimated that the use of nonfuel materials in the US exceeds 60 kg per person per day. 1 While most of the rest of world uses much less, usage there is growing at twice the rate. 2 Although these estimates are fraught with uncertainty, they point to an emerging global challenge in dealing with the effects of unprecedented levels of materials use. Key strategies that will likely play a role in meeting that challenge are increasing the reliance on both secondary (i.e., recycled) and renewable re- sources. For recycled resources, energy benefits are well docu- mented and in some cases exceptional. 3 Regarding renewables, the ultimate benefits, which may include reduced energy use, nonrenewable depletion, and carbon burden, remain controver- sial for many current applications. Nevertheless, the trend in renewables technology development is promising. 4 Although rarely discussed together, these two strategies share a common economic barrier to implementation  increased quality varia- tion compared to more conventional resources (i.e., primary mineral ores or synthetics). The existence of increased variability creates an inherent economic disincentive for using recycled or renewable raw materials. Unfortunately, the most prevalent implementations of batch process planning tools (a key tool for materials pro- ducers and recyclers) overestimate this disincentive and thereby undervalue and underutilize such raw materials. This paper examines the benefits, in terms of cost and variable raw material usage, of one approach to batch plan- ning tools that explicitly considers raw material variability, a chance-constrained (CC) model formulation. Speci fically, this paper explores the generality of the benefits of a CC- based model, the drivers of that benefit, and the conditions that maximize benefit. Before moving to the analysis, the next two sections examine the pervasiveness of quality variability in both recycled and renewable materials contexts and discuss the literature on managing variation in general and, within process industries, on the use of blending models. Received: October 27, 2010 Accepted: March 22, 2011 Revised: March 7, 2011 ABSTRACT: The increased use of secondary (i.e., recycled) and renewable resources will likely be key toward achieving sustainable materials use. Unfortunately, these strategies share a common barrier to economical implementation  increased quality variation compared to their primary and synthetic counterparts. Current deterministic process-planning models overestimate the economic impact of this increased variation. This paper shows that for a range of industries from biomaterials to inorganics, managing variation through a chance-constrained (CC) model enables increased use of such variable raw materials, or heterogeneous feedstocks (hF), over conven- tional, deterministic models. An abstract, analytical model and a quantitative model applied to an industrial case of aluminum recycling were used to explore the limits and benefits of the CC formulation. The results indicate that the CC solution can reduce cost and increase potential hF use across a broad range of production conditions through raw materials diversification. These benefits increase where the hFs exhibit mean quality performance close to that of the more homogeneous feedstocks (often the primary and synthetic materials) or have large quality variability. In terms of operational context, the relative performance grows as intolerance for batch error increases and as the opportunity to diversify the raw material portfolio increases.