MICROWAVE PARAMAGNETIC RESONANCE ABSORPTION 479 



It would however have been a great deal more difficult to obtain the 

 entire bucking of +22 db at 10"^ W since a set-up like in Fig. 8 would 

 have to be used. Thus for the sake of simplicity the extra factor in signal 

 to noise of 2 or 3 was abandoned. The amount of negative bucking at 

 the higher powers will be limited by the stability of the bridge. A prac- 

 tical limit of (40-50) db was characteristic of our set-up. We see from 

 Fig. 9 that the agreement between the experimentally determined sensi- 

 tivity and the theoretically predicted sensitivity is satisfactory. 



The experimental results on the superheterodj'ne scheme agrees 

 again verj^ well with the predicted values up to a power level of 10"^ W. 

 (This corresponds to less than 10'^ spins in D.P.H.) Above this level Ti 

 (see fig. 10) has to be balanced to better than 40 db to keep the IF carrier 

 amplitude within the required value. Instabilities in the bridge due to 

 mechanical vibrations and thermal drifts start to contribute to the 

 noise. Thus at high power levels the superhet scheme starts to loose 

 some of its advantages unless special precautions are being taken to 

 eliminate the above mentioned noise factors. A great deal in this direc- 

 tion could probably be accomplished by shock-mounting the micro- 

 wave components and better temperature stabilit}^ for slow drifts. 

 Since we were mainly interested in powers below 1 mW, our efforts were 

 limited to controlling the temperature of the room to ±1°C. 



Since the superhet scheme was found to be the most sensitive one, 

 it might be w'orthwhile to discuss it in more detail. A block diagram of the 

 set up is shown in Fig. 10. 



The signal generator feeds into the magic T, where its power is split 

 between arm 2 and 3. Arm 2 has the reflection cavity with the sample, 

 the reflected voltage being bucked out with the aid of arm 3. For this 

 purpose arm 3 has a phase shifter and attenuator, an arrangement 

 which was found to be more satisfactory than a shde screw tuner as far 

 as stability and ease of operation goes. The desired signal appears then 

 in arm 4. It is fed into a balanced mixer which receives the local oscil- 

 lator power from the stabilized klystron II. The output of the balanced 

 mixer is then fed through the IF amplifier, detector, audio ampHfier and 

 lock-in detector. The circuits of each of those components is fairlj^ 

 standard and will not be dwelled upon further. The microwave power is 

 measured in arm 2 of the magic T. The power reflected from the cavity 

 is also monitored in arm 2. This is of great help in finding the cavity 

 when klystron I is sw^ept in frequency by means of a sawtooth voltage on 

 its reflector. Since the klj'stron mode itself might have some dips in it, 

 (which might be mistaken for the cavity), it proved helpful to display on 

 the scope the klj'^stron mode simultaneously with the reflected power 



