I'm guessing that the chemical rocket envelope in the plot encompasses points representing actually-built rocket engines, rather than theoretical ones, hence some of the irregularity of the shape is due to historical accident.
10N is quite small for a chemical rocket engine. Such units are mainly used for attitude control of small spacecraft rather than making significant maneuvers, so reliability, simplicity, and light weight are more critical than specific impulse. I'm not sure if it's not actually possible to make high-Isp small thrusters, or if it's just that no one bothers.
In particular, I note that Aerojet's catalog of bipropellant (MMH+NTO) thrusters (which get around 300 seconds Isp) extends down to 22N; below that, they offer catalyzed hydrazine monoprop thrusters (around 220 seconds) down to 1N; the engineering simplicity of requiring only a single propellant tank pays for the loss in specific impulse (and it's probably tricky to get good bipropellant mixing in such a small combustion chamber). These two categories of thruster contribute to the left two-thirds of the chemical engine envelope in the plot.
Further up and to the right, small hydrogen-oxygen engines stake out the high end of the Isp envelope: the Chinese YF-73 at 44kN and 420 seconds, then a bunch of engines in the neighborhood of the US RL10: 65-100 kN and 440-460 seconds. That's the high-water specific impulse mark for production chemical engines, the space shuttle main engine RS-25 is off the right hand side of the chart at 2200kN and 452 seconds. Again, I'm not sure whether it's possible to make high-Isp hydrogen-oxygen engines smaller than ~40kN or whether it's just not done.