The first application of DNP for sensitivity enhancement in solid-state MAS NMR was demonstrated in the 1980's by a group at Delft Technical University. That research was focused on the study of coals, and used the radicals naturally present in such materials:
Wind RA, Duijvestijn MJ, van der Lugt C, Manenschijn A, Vriend J. Applications of dynamic nuclear polarization in 13 C C NMR in solids. Progr NMR Spectr 17, 33 (1985),
Later, more general DNP/MAS techniques for general samples (with explicit radical doping) and in rather higher fields have been developed at the Massachusetts Institute of Technology. An important part of that research involves the chemical synthesis of bi-radicals, to improve the efficiency of the DNP process (see SREP Team website).
Technically, the big difference with dissolution-DNP is, well, that there is no dissolution. The DNP process and the NMR observation are done in situ at the same field and temperature. The upper limit to the field is given by the availability of suitable microwave sources. The lower limit to the temperature is given by the practical requirements of a MAS system driven by something cheaper than He: starting with cold N2 gas, the actual sample temperature may be of the order of 100 K. At that temperature, the electron spin-lattice relaxation times become very short; which implies that powerful microwave sources are needed: of the order of 10 W (instead of the 20 - 100 mW used in dissolution-DNP).
The coupling of the microwaves remains rather inefficient. This requires relatively powerful, high-frequency microwave sources (gyrotrons).
Gyrotrons belong to the family of coherent radiation sources known as electron cyclotron masers (ECM). The physical mechanism of a gyrotron is based on the relativistic dependence of the electron cyclotron frequency on the electron energy and an associated instability known as a 'negative mass instability'. When an ensemble of electrons are injected into a magnetic field and interact with an electromagnetic wave whose frequency is equal to the relativistic cyclotron frequency, the resulting wave particle interaction induces an instability (the 'negative mass instability') that is manifested by orbital bunching of the electrons initially uniformly distributed on a Larmor radius. If this occurs in a millimeter-wave cavity designed to support a resonant mode with a frequency slightly higher than the relativistic cyclotron frequency, it will cause the electron bunch to decelerate rapidly and emit a substantial fraction of its rotational kinetic energy as millimeter-wave radiation. Figure1 gives an illustration of all key elements of a high power gyrotron that uses this process. The negative mass instability was first predicted theoretically both in terms of a classical [1,2] and quantum theory  in the late 1950s and the first experimental demonstration followed a few years later, demonstrating the potential for this source to be used as an efficient heating scheme for magnetized plasmas. But going from this early proof-of principle to the performances achieved for gyrotrons to be used in ITER [4,5] has required a substantial effort in both technology and theoretical understanding . In the last decade in parallel to the development of gyrotrons at MW-power level for fusion applications a second line of development was pursued with a diagnostic oriented goal. These type of gyrotrons are small to medium power gyrotrons (10-1000W) operating at frequencies ranging from (100GHz to 1THz) with applications to various fields such as plasma diagnostics, far-infrared spectroscopy, ESR  spectroscopy and DNP enhanced NMR . Compared to gyrotron for fusion applications, the 'diagnostic' gyrotron for spectroscopy applications are technologically significantly simpler since they must not handle large RF powers, however, their rf spectral properties are more stringent in terms of frequency stability and line-width. The gyrotrons presently under design at CRPP for DNP enhanced NMR spectroscopy will operate at frequencies of 196GHz and 392GHz and will provide rf powers in excess of 50W.
Figure 1. Schematic of a high power high-frequency gyrotron for fusion applications. An annular cross-section electron beam is guided by a strong magnetic field produced by a superconducting magnet and injected into a radiofrequency resonant cavity. As the beam passes through this cavity it experiences a negative mass instability, whereby electrons within it, initially uniformly distributed on a Larmor radius, undergo orbital bunching (see upper inset), causing them to decelerate rapidly and emit coherent rf radiation. The lower inset is a photograph of a cross-section of the resonator cavity of a typical 1-MW-class gyrotron; the annular electron beam in the resonator has an average radius and thickness of 10mm and 0.5mm, respectively. Compared to a gyrotron for fusion application, the characteristic mechanical and electrical parameters for a gyrotron to be used for DNP enhanced NMR spectroscopy are between one and two order of magnitude smaller.
 R.Q. Twiss, Radiation Transfer and the Possibility of Negative Absorption in Radio Astronomy, Aust. J. Phys. 11, 564 (1958).
 A. V. Gaponov, Interaction between electron fluxes and electromagnetic waves in waveguides, Izv. VUZ. Radiofizika 2, 450 (1959).
 J. Schneider, Stimulated Emission of Radiation by Relativistic Electrons in a Magnetic Field, Phys. Rev. Lett. 2, 504 (1958).
 K. Sakamoto, A. Kasugai, K. Takahashi, R. Minami, N. Kobayashi, K. Kajiwara, Achievement of robust high-efficiency 1 MW oscillation in the hard-self-excitation region by a 170 GHz continuous-wave gyrotron, Nature Phys. 3, 411 (2007).
 J.P. Hogge, S. Alberti, A. Arnold, D. Bariou, P. Benin, T. Bonicelli, A. Bruschi, R. Chavan, S. Cirant, O. Dumbrajs, D. Fasel, F. Gandini, E. Giguet, T. Goodman, R. Heidinger, M. Henderson, S. Illy, J. Jin, C. Lievin, R. Magne, P. Marmillod, P.L. Mondino, A. Perez, B. Piosczyk, L. Porte, T. Rzesnicki, M. Santinelli, M. Thumm, M.Q. Tran, I. Yovchev, Development of a 2-MW, CW Coaxial Gyrotron at 70 GHz and Test Facility for ITER, J. Phys. Conf. Series 25, 33 (2005).
 G.S. Nusinovich, Introduction to the Physics of Gyrotrons (John Hopkins Univ. Press, Maryland, USA, 2004)
 S. Mitsudo, T. Higuchi, K. Kanazawa, T. Idehara, I. Ogawa, M. Chiba, High Field ESR Measurements Using Gyrotron FU Series as Radiation Sources, J. Phys. Soc. Jpn. 72, Suppl. B 172 (2003).
 C.D. Joye, R.G. Griffin, M.K. Hornstein, Hu Kan-Nian, K.E. Kreischer, M. Rosay, M.A. Shapiro, J.R. Sirigiri, R.J. Temkin, P.P. Woskov, Operational characteristics of a 14-W 140-GHz gyrotron for dynamic nuclear polarization, IEEE Trans. Plasma Sci. 34, 518 (2006).