Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

How can we find $\pi_{d}(U(N+M)/U(N) \times U(M))$ ? Is there any way to visualize the target space ? I am specifically interested in the $d=1,2$ & $(N,M)=(2,2),(2,4)$ cases. Thanks.

share|cite|improve this question
Have you tried looking at the long exact sequence for the fibration $U(N)\times U(M)\hookrightarrow U(N+M)\to U(N+M)/U(N)\times U(M)$? That is usually good enough to get the low degree groups. – Mariano Suárez-Alvarez Sep 8 '13 at 4:57
Please include the definitions of what you are asking when you ask a question. – Vishal Gupta Sep 8 '13 at 5:16
What are $M,N, M +N, U$? How does $d$ get into the act? – Ronnie Brown Sep 8 '13 at 10:18

Let us write $X=U(N+M)/U(N)\times U(M)$. The fibration $$U(N)\times U(M)\hookrightarrow U(N+M)\to U(N+M)/U(N)\times U(M)$$ gives as a long exact sequence (because all spaces are connected) \begin{align} \tag{$\star$} 0\leftarrow\pi_1(X)\leftarrow\pi_1(U(N+M))&\leftarrow\pi_1(U(N)\times U(M))\leftarrow \\ &\leftarrow\pi_2(X)\leftarrow\pi_2(U(N+M))\leftarrow\cdots \end{align} If $K\geq1$, there is a fibration $$U(K-1)\hookrightarrow U(K)\to U(K)/U(K-1)=S^{2K-1}$$ from which we get again a long exact sequence \begin{align} 0\leftarrow\pi_1(S^{2K-1})&\leftarrow\pi_1(U(K))\leftarrow\pi_1(U(K-1))\leftarrow \\ &\leftarrow\pi_2(S^{2K-1})\leftarrow\pi_2(U(K))\leftarrow\pi_2(U(K-1))\leftarrow\pi_3(S^{2K-1})\cdots \end{align} Now:

  • If $K\geq2$, then $2K-1\geq3$ and therefore $\pi_1(S^{2K-1})=\pi_2(S^{2K-1})=0$, and so we see that the inclusion $U(K-1)\to U(K)$ induces an isomorphism in $\pi_1$. It follows by induction that $\pi_1(U(K))=\pi_1(U(1))\cong\mathbb Z$ for $K\geq1$
  • If $K\geq3$, then $2K-1\geq5$ and $\pi_2(S^{2K-1})=\pi_3(S^{2K-1})=0$, and the exact sequence shows that the inclusion $U(K-1)\to U(K)$ induces an isomorphism in $\pi_2$. It follows by induction that $\pi_2(U(K))\cong\pi_2(U(2))$ for $K\geq2$.
  • When $K=2$, the long exact sequence tells us that $$\pi_2(S^3)\leftarrow\pi_2(U(2))\leftarrow\pi_2(U(1))$$ is exact, and we know that the first and third groups are zero, so $\pi_2(U(2))=0$. The previous point tells us then that $\pi_2(U(K))=0$ for all $K\geq1$.

We now go back to the sequence $(\star)$ and let us suppose that $N,M\geq1$. Using the computations for $\pi_1$ and $\pi_2$ of $U(K)$ that we have now, and the fact that homotopy groups play nicely with products. We are left with the sequence $$ 0\leftarrow\pi_1(X)\leftarrow\pi_1(U(N+M))\leftarrow\pi_1(U(N))\times\pi_1(U(M))\xleftarrow{\partial}\pi_2(X)\leftarrow0 $$ We know that the arrow $\pi_1(U(N+M))\leftarrow\pi_1(U(N))\times\pi_1(U(M))$ is sujective (because already the restriction to th efirst factor, $\pi_1(U(N+M))\leftarrow\pi_1(U(N))$ is surjective according to the first point above!) It follows that $\pi_1(X)=0$ and $\pi_2(X)\cong\mathbb Z$.

share|cite|improve this answer
I come from a Physics background. The above Homotopy group is relevant in the context of 'Topological defects' in Condensed matter physics. I am only comfortable with thinking about mappings from circle / sphere to sphere / torus. I am finding your answer highly technical. Can you please suggest a textbook where I find such stuff (basically the idea of fibration) ? Thanks a lot for your answer. – Akshay Kumar Sep 8 '13 at 16:44
Well, computing homottopy groups is a rather technical enterprise! You can learn all this in books like Switzer's or Whitehead's on homotopy theory. – Mariano Suárez-Alvarez Sep 8 '13 at 19:16
For the case of Z classification, can we write down a formula that gives me the particular integer class (topological invariant) to which a given mapping belongs to ? – Akshay Kumar Sep 9 '13 at 5:58
The map $\partial$ in the last exact sequence is quite explicit (because our fibration is very explicit!) and can in principle be computed. You should read a bit about the long exact sequence for homotopy groups. – Mariano Suárez-Alvarez Sep 9 '13 at 6:14

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.