There are many cons of that:

1. Cluster controlling becomes hard

2. Cluster head selection process becomes more difficult

3. If you increase network radius by increasing the distance between two nodes, energy consumption of each node increases

4. MAC layer protocols becomes more complex

5. Bandwidth issue will be there if FDMA is there. In case of CDMA its not a problem

Increasing the cluster radius may result more nodes join the cluster and distance of edge nodes to the cluster head will increase. Thus increasing energy consumption.

Also, if the CH creates a TDMA schedule for the nodes in its cluster. Each node has to wait more to send their data packets. Because every node should sent in their own slot which is assigned by CH.

I recommed to you to study LEACH protocol. In the paper, some analysis about optimal number of CH is done.  

Increasing cluster radius/size.

Advantages: Avoid the data redundancy,if  2 nodes may have same data and both data are collected twice by 2 CH's i.e both are the members of different CH, So if this data is collected by single CH then it avoid redundancy

Disadvantage: TDMA slot may not sufficient for communication between CH and member, the chances of  more energy dissipation by CH  and even loss of data  

Thank you all,

Is there any formula to find the optimal number of clusters or the size of clusters?

Yes, if you read LEACH protocol paper, you can also find a mathematical definition of finding the optimal number of clusters.

>Number of Clusters = No. of Cluster Head elected for 1 round, Cluster Size= No. of  members in each CH.

> In LEACH No. CH depends on Probability P, where P plays an vital role in finding threshold T(n). if P=0.5, then 50% of the nodes will be CH. If total nodes =100, then 50 will CH and the cluster size may be very small.

>Using this P & T(n) you can specify your own constraint to choose your desire CH and its  size. refer this link- > ( http://ieeexplore.ieee.org/xpl/abstractAuthors.jsp?reload=true&arnumber=6749371) 

Basically with an increase in cluster size the number of nodes in the cluster increases, along with increase in distance to the cluster head, so immediate effect is on the energy drained out on reaching the cluster head. Moreover, delay to send the packets to the destination will also increase along with the chances of replicating data. So the lifetime on individual nodes with be decreased and overall network lifetime also reduced. On the other hand less radius implies less cluster members, higher lifetime, as the cost of an increase in the number of cluster heads. So there is a tradeoff in finding the optimal size of the radius which keep the clusters size within limit along with maintaining the energy balance. The actual LEACH paper provides some mathematical formulation on this.