GL_NV_depth_buffer_float provides a format of texture that can store depth buffer information not only in 32-bit but also in float. This is for use with a FBO (framebuffer object)
Not sure if you can set the system (window) framebuffer’s depth buffer to 32-bits, but more than likely you can configure a framebuffer object (FBO) with a depth buffer that has 32-bits, render to that, and then merely copy the resulting color buffer to the system color buffer.
Here you can see that ARB_depth_texture provides GL_DEPTH_COMPONENT32 (32-bit int depth buffer). Further, you can see this is part of the OpenGL 3.3 or 4.0 specs. So if you’re using one of those versions, or your OpenGL driver provides ARB_depth_texture, you should be able to create a 32-bit depth texture and render to it with an FBO (assuming you have ARB_framebuffer_object, or are running OpenGL 3.0+).