Electrochemical state estimation of a lithium-ion battery provides crucial information to the battery management system about its state of charge and state of health. Insight into nonmeasurable electrochemical states, such as bulk and surface concentration, holds the key to improving safety, predicting degradation, and facilitating a wider range of battery operation. Although strategies to estimate concentration in a single electrode are well-researched, the topic of simultaneous concentration estimation in both electrodes is not fully investigated. Real-time knowledge of lithium concentration in both electrodes will be essential in setting operational limits specific to the electrode chemistry, enabling estimation of available power in each electrode, and developing algorithms to monitor or minimize degradation mechanisms occurring at either electrode. To that end, an interconnected sliding mode observer is proposed for the concurrent estimation of bulk and surface concentration in the cathode and anode. An experimentally validated enhanced single particle model forms the basis for the electrochemical-model-based observer design. The novel bidirectional structure of the interconnected observer circumvents the issue associated with the combined concentration estimation in both electrodes. The convergence of the interconnected observer's error dynamics is proved using Lyapunov's stability theory and its performance is verified under measurement noise and parametric uncertainties.