Concept:
- Elementary row operations never change the rank of a matrix, and the identity matrix has full rank $3$. So a matrix can be obtained from the identity matrix by row operations only if it also has full rank $3$, that is, its rows are linearly independent. Any full-rank $3\times3$ matrix can likewise be reduced back to the identity matrix, so full rank is exactly the condition needed.
- A row is linearly dependent on the other two if it can be written as $R_3=aR_1+bR_2$ for some numbers $a,b$. This can often be spotted directly, without computing a full determinant.
Step 1: Test option (A).
All three rows are $[1,1,1]$, so $R_2=R_1$ directly. The rows are dependent, rank is less than $3$, so this matrix cannot be obtained from the identity matrix.
Step 2: Test option (C).
Rows are $R_1=[1,1,1]$, $R_2=[2,3,4]$, $R_3=[2,5,8]$. Try $R_3=aR_1+bR_2$:
$a+2b=2$ and $a+3b=5$, so subtracting gives $b=3$, then $a=2-2(3)=-4$.
Check the third entry: $a+4b=-4+12=8$, which matches the third entry of $R_3$. So $R_3=-4R_1+3R_2$, the rows are dependent, rank is less than $3$. This matrix cannot be obtained from the identity matrix.
Step 3: Test option (D).
Rows are $R_1=[1,1,1]$, $R_2=[-1,1,2]$, $R_3=[0,2,3]$. Adding the first two rows directly: $R_1+R_2=[0,2,3]$, which is exactly $R_3$. So $R_3=R_1+R_2$, the rows are dependent, rank is less than $3$. This matrix cannot be obtained from the identity matrix.
Step 4: Test option (B).
Rows are $R_1=[1,1,1]$, $R_2=[2,3,4]$, $R_3=[1,2,1]$. Try $R_3=aR_1+bR_2$:
$a+2b=1$ and $a+3b=2$, so subtracting gives $b=1$, then $a=1-2(1)=-1$.
Check the third entry: $a+4b=-1+4=3$, but the third entry of $R_3$ is $1$, not $3$. The values of $a,b$ that fit the first two entries fail the third, so no such combination exists.
Step 5: Conclude for option (B).
Since $R_3$ cannot be written as a combination of $R_1$ and $R_2$, the three rows are linearly independent, giving full rank $3$. This matrix can be obtained from the identity matrix by elementary row operations.
Final Answer: Option (B)