That doesn't look like a spectrum at all. Either you get light into the detector or the Photomultiplier is broken. The shape should have no big influence on the spectrum.

For me your picture is pure noise. No photopeak at all. I suggest defect electronics, my impression is no contact of PM with detector. I attend a typical spectrum, the upper curve is NaI. You will find an estimate for resolution. 

For me, this is just the background. I think you need to check the configuration of your test. For resolution, it is around 5%~6%(my experience) for 1173.2 keV. 

 as i now the resolution of Nai(Tl) detector is about 4-5% for the energy of Co-60 (1173) keV, and  form the picture i found it is background 

I´m not sure with background, because in every background spectrum you see at least some hints to peaks from the natural radiation which can easily been distinguished from noise. And especially the sharp step at the higher end of your curve looks  like a defect.

Annular detectors will have a difference in efficiency and a small change in resolution. You have no indication of photopeaks, only noise. Check all connections. Substitute a working detector to ensure electronics are working up to the photomultiplier. 

this is how Co-60 and background spectras should look like in annular NaI(Tl) detector.

it is obviously something wrong with electronic components, cables, photomultipliers .... Try to reconnect everything, test photomultipliers....

The shape as such is not expected to have any effect on the resolution.  What is the thickness of the crystal? Please make sure that due to the shape, the net thickness is not too less that Co-60 gammas are simply penetrating through without getting registered in the detector.  

There is even no background ramp in the graph. I suggest you check if the high voltage (HV) is switch on in counting because it looks like the photomultiplier tube (PMT) didn't pick up any lights from the NaI crystal. Forgetting turning on the HV is a common mistake in operating gamma spectrometers. Another reason for this is the PMT is broken.

 @Erik Strub

Thank you!

I had tried several days,but,any changes happened.I

The picture of crystal attached.The crystal is 9 in in diameter,2 in in length.The diameter of Central hole is 7 in.

Thickness of MgO layer is approx 2 mm,warpped by stainless steel is 0.5 mm  in thickness.

I've tried different photomultipliers,its function well.

The shape in the article named An NaI(Tl).............  is roughly the same with mine.

I connect the original signals into an oscilloscope.The  voltage value of noise  is about 40 mV,however,the voltage value is 200 mV when i put Co-60 in the center of the crystal.

Try to connect signal from amplifier to oscilloscope. Maybe problem is with amplifier.

@Hanno Krieger

Thank you for your typical spectrum.

I used an DigiBASE  manufactureed by ORTEC copmany.The result is the same.

I've tried different photomultipliers,its function well..

I connect the original signals into an oscilloscope.The  voltage value of noise  is about 40 mV,however,the voltage value is 200 mV when i put Co-60 in the center of the crystal.

@Liu Jiajian

Thank you!

I have an 2 in × 2in crystal,the enery resolution is around 7% for 0.662 MeV.and a cubic crystal Rectangle one,which in 7 in × 7 in ×16 in ,it's enery resolution is around 8%,worse than the cylinder crystal is  2 in in diameter .

The picture of my semi-annular crystal attached.

Are you able to switch photomultipliers connected to the annular NaI? How is the light pipe coupled? The photo shows a 2-inch cylinder attached to the semi-annulus. Is that the mount for the photomultiplier? How does the window contact the crystal?

Try Jan´s tip.

@Jan Hansman

Thank you！

I access the original signals,which is from the anode of Photomultiplier into an oscilloscope, .The  voltage value of noise  is about 40 mV,however,the voltage value is 200 mV when i put Co-60 in the center of the crystal.

Signal has to be bigger with source inside/on detector. That is ok.  I`m more worried about that sharp step at the end of spectrum. I had similar problem and it turns out that problem was with bad amplifier. Be careful,some amplifiers do not work well with NaI(Tl) detectors.

Where are you keeping the source with respect to the crystal to get the reported spectrum?

@Jan Hansman  @Joseph L Alvarez @Vrinda Devi K V

I‘ve tried digiBASE and digiBASE-E and 926,all spectrums are terrbile.no photopeaks.

I've connected signal from the anode of photomultipliers to oscilloscope.

The voltage value of noise is about 40 mV,however,the voltage value is 200 mV when i keep the source(Co-60) near the crystal in different direction.

The high voltage is about 700 volte.

The signal is terrible form the anode of photomultipliers.

Several pictures of crystal attached.

If you have a chance take a X-ray picture of your detector. Eventually you find a broken structure. One possible error yould be an open device. NaI don´t like water or liquids, they get blind, so you can´t detect real signals, just a white noise which is simular to your showed spectrum.

Is your detector sealed in an airtight housing?  What is the active area of the detector?  Are you sure it is not picking up any background radiation from the annulus?  Try blocking the annulus completely with any shielding material and if that helps in reducing the noise, you may need to put a collimator with limited slit width (which will expose the source) to prevent large active area of the detector from picking up the noise. 

HV of 700 V for this type of detector can be little too low. Check the manual and consider about adjusting HV maybe to 900-1000 V.

Also you can have too much signals for the electronics that you use. How much is dead time of the detector? Try to adjust Rise/Fall Time, Flat Top Time, Gate, Trigger etc.

And the worst thing, like @Hanno Krieger  write...it`s possible that crystal is broken or photomultiplier....

Dear Hatiden

You must review the entire range spectrometry system, from the connections, the instrumentation and grounding problems. It may be a problem generated by the earthing system no damp or false contacts in the parts of your system, unless you have controlled environmental conditions in your lab, or it may be due to faulty amplifier, you can greatly help an oscilloscope to see the signal at the output of each stage of your system.

hi Hatiden Smith，

I think what the persons above said are all right. i believe you should check all the system including HV, PMT, AM, PAM, Crystal. If you can not get the peak, even just Cs-137 the only gamma ray source. It is meaningless to do other measurement. Good luck!

@Hanno Krieger

I will try your tips.

I suspected that the semi-annular NaI detector have crystal defects.Some mistakes may happened in Crystal capsulation.

The photo numbers reduced a lot in the crystal.it shoule be 3,8000 photos per MeV.Actually.I measured a few.

It maybe somethings wrong with the MgO layer.or the crystal itself.

Thank you !

The reason shold be the crystal itslf. irrespective of anything else.

The photos loss donated the voltage value which is low.

Dear Hatiden,

Low voltage could arise from insulation problems around the cristal due to humidity. I hope you solve your problem. Best success.

Dear Hatiden,

i believe you have received a lot of good advice from all the responders. However, looking at your GammaVision MCA output, I notice that you have too many channels for a NaI spectrum. Please try the following: use 512 channels, drop the gain on your amplifier and take a new Gamma spectrum. I think your initial setting for your spectroscopic measurement was set up for the use of a Germanium detector. Am I correct? They have totally different energy resolutions.

Hope this will help you. 

Dea, it is not like a spectrum , it is look like background . check the system which included  HV, PMT, AM, PAM, Crystal. if  you use Cs-137 ans accumulated a spectrum for a period of time , check the position and the peak of the Cs-137. Good luck  

Dear Brother ... I sent you an email .... Please check and reply me ... Regards