(Disclosure: I'm a physicist)
You could just as well ask: "how can an electric field line just stop somewhere?", and thereby conclude that there can be no such thing as an "electric monopole" (a positively- or negatively-charged particle). As long as the universe has no net electric or magnetic charge, all lines will terminate somewhere. If the universe did have a net charge the point is subtle, but that's irrelevant: the paper talks above pairs of opposite-pole monopoles created together, like a particle and its antiparticle. So this argument doesn't hold water.
Monopoles aren't impossible in principle (it would just be an extra term in Maxwell's equations) and are predicted in some theories, but fundamental-particle monopoles have never been observed. The summaries of this paper are confusing a lot of people: the authors are describing a crystal system with excitations that look like monopoles. They are NOT describing discovery of a new fundamental particle, but rather a new kind of solid-state phenomenon.
It's worth noting that Planck doesn't actually use a lot of liquid helium to cool itself down. It's cryogenic system is based upon "cryogen-free" mechanical refrigerators - the satellite launches warm, then cools itself down electrically and by radiating to space. The satellite lifetime isn't limited by running out of liquid helium.
Herschel, in contrast, does have a giant liquid helium tank. It launches full of helium, and eventually warms up when the tank runs out.
The "E-modes" and "B-modes" referred to in the article aren't quite the same as electric and magnetic fields. Here's the basic story.
Suppose you try to map the polarization of the microwave background across the sky. Each direction on the sky has some polarization magnitude and direction, which we can represent by a little headless arrows on the sky (headless because flipping the polarization 180 degrees doesn't change it). A map of the CMB polarization thus looks like a bunch of little line segments of varying sizes and orientations all across the sky.
Now imagine looking at the pattern of polarization directions near some point on the sky. This arrangement of lines can be "curl-free" if the lines are oriented radially or circumferentially around the central point; this is called an "E-mode" pattern. The polarization pattern might instead have a curl component, which is called a "B-mode" pattern. another way of looking at it: an E-mode pattern looks locally the same when mirror-reversed, while a B-mode pattern does not. Any field on the sky can be written as the sum of an E-mode pattern and a B-mode pattern.
This technicality is important because of how polarization is generated in the microwave background. It turns out that all kinds of relatively mundane processes can generate E-modes - they're still very interesting and informative, but we know they're there (and have even detected them). B-mode patterns are much more unusual - it turns out that normal CMB physics cannot generate large-scale B-modes. Inflation, however, generates a background of gravity waves in the early universe that produce a B-mode contribution to the CMB. This is incredibly tiny and difficult to detect, but it's a smoking gun for inflation.
It's worth noting that more than one such telescope hopes to probe CMB polarization on a similar timescale. Caltech and JPL are leading the BICEP2 and SPIDER collaborations (also with NIST), which will also be deploying in a few months (the former at the South Pole, the latter on a high-flying balloon) to probe E-mode and B-mode CMB polarization. The Princeton experiment mentioned in this article isn't that different - it just apparently has better press!