Context : Galaxy evolution is known to depend on environment since it differs in clusters and in the field , but studies are sometimes limited to the relatively nearby Universe ( z < 0.5 ) .
It is still necessary to increase our knowledge of cluster galaxy properties above z = 0.5 .

Aims : In a previous paper we have detected several cluster candidates at z > 0.5 as part of a systematic search for clusters in the Canada France Hawaii Telescope Legacy Survey by applying the Adami & MAzure Cluster FInder ( AMACFI ) , based on photometric redshifts .
We focus here on two of them , located in the Deep 3 ( hereafter D3 ) field : D3-6 and D3-43 .

Methods : We have obtained spectroscopy with Gemini/GMOS instrument and measured redshifts for 23 and 14 galaxies in the two structures .
These redshifts were combined with those available in the literature .
A dynamical and a weak lensing analysis were also performed , together with the study of X-ray Chandra archive data .

Results : Cluster D3-6 is found to be a single structure of height spectroscopically confirmed members at an average redshift z = 0.607 , with a velocity dispersion of 423 km s ^ { -1 } .
It appears to be a relatively low-mass cluster .
D3-43-S3 has 46 spectroscopically confirmed members at an average redshift z = 0.739 .
The cluster can be decomposed into two main substructures , having a velocity dispersion of about 600 and 350 km s ^ { -1 } .
An explanation of the fact that D3-43-S3 is detected through weak lensing ( only marginally , at the \sim 3 \sigma level ) but not in X-rays could be that the two substructures are just beginning to merge more or less along the line of sight .
We also show that D3-6 and D3-43-S3 have similar global galaxy luminosity functions , stellar mass functions , and star formation rate ( SFR ) distributions .
The only differences are that D3-6 exhibits a lack of faint early-type galaxies , a deficit of extremely high stellar mass galaxies compared to D3-43-S3 , and an excess of very high star formation rate galaxies .

Conclusions : This study shows the power of techniques based on photometric redshifts to detect low to moderately massive structures , even at z \sim 0.75 .
Combined-approach cluster surveys such as EUCLID are crucial to find and study these clusters at these relatively high redshifts .
Finally , we show that photometric redshift techniques are also well suited to study the galaxy content and properties of the clusters ( galaxy types , star formation rates , etc… ) .