Existing sensor platforms have limited sensitivity, specificity, and portability. With a new algorithm for the coupled dipole approximation of Maxwell's equations, we examine near- and far-field features of electromagnetism (EM) coupled with localized surface plasmons on subwavelength, solid-state nanoparticle (NP) structures measured using spectroscopy, microscopy, and calorimetry. <i xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Near-field</i> extinction efficiency, blue/redshifts, and full-width at half-maximum are optimized using a new ¿bottom-up¿ NP assembly method that tunes particle size and spacing to enhance sensitivity and produce molecule-specific ¿tenfold surface-enhanced Raman spectroscopy enhancements. <i xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Far-field</i> plasmon-photon resonances are identified, which offer ¿ 10 <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">6</sup> -fold higher sensitivity. Solid-state NP structures increase stability, reduce power consumption, and improve response time and optothermal transduction up to tenfold for better portability and throughput relative to aggregation-prone NP suspensions. Sample rate is increased ¿tenfold by inducing transverse hydrodynamic diffusion adjacent to sensor interfaces. These results guide development of next-generation chemical and biological sensors based on EM-coupled UV, Raman, or terahertz modes that improve sensitivity, biospecificity, stability, and portability to distinguish biological molecules and species at high throughputs.