Proof in number theory: the sum of all divisors of n

In summary, the proof in number theory regarding the sum of all divisors of a positive integer \( n \) involves the use of the divisor function \( \sigma(n) \). The function sums all positive divisors of \( n \), and its properties can be derived from the prime factorization of \( n \). If \( n \) can be expressed as \( n = p_1^{k_1} p_2^{k_2} \cdots p_m^{k_m} \), where \( p_i \) are distinct primes and \( k_i \) are their respective powers, then the sum of divisors is given by the formula \( \sigma(n) = (1 + p_1 + p_
  • #1
mathstudent1
18
0
let n be a positive integer show that if n is square then σ(n)( the sum of all divisors of n )is odd.
 
Physics news on Phys.org
  • #2
Do you have any ideas? What do you know about ##\sigma (n)##?

Before you start proving something you must make an inventory. E.g. there is a formula for ##\sigma (n)## if ##n## is given by its prime factor decomposition. If you can use that, then the proof is probably a lot easier as if you do not know this formula.
 
  • Like
Likes pinball1970
  • #3
fresh_42 said:
Do you have any ideas? What do you know about ##\sigma (n)##?

Before you start proving something you must make an inventory. E.g. there is a formula for ##\sigma (n)## if ##n## is given by its prime factor decomposition. If you can use that, then the proof is probably a lot easier as if you do not know this formula.
you mean this σ(p^a)=p^(a+1) -1/(p-1)? I have another question how can I write the form of square in the form of prime?
 
  • #4
mathstudent1 said:
you mean this σ(p^a)=p^(a+1) -1/(p-1)? I have another question how can I write the form of square in the form of prime?
I meant a more complicated formula with more primes.

Let's start with ##n## and what do we know. Say ##n=p_1^{r_1}\cdot p_2^{r_2}\cdots p_k^{r_k}## with pairwise distinct primes ##p_1,\ldots,p_k## and positive integers ##r_1,\ldots,r_k.##

What does it mean that ##n## is a square?
 
  • #5
fresh_42 said:
I meant a more complicated formula with more primes.

Let's start with ##n## and what do we know. Say ##n=p_1^{r_1}\cdot p_2^{r_2}\cdots p_k^{r_k}## with pairwise distinct primes ##p_1,\ldots,p_k## and positive integers ##r_1,\ldots,r_k.##

What does it mean that ##n## is a square?
n^2=p1^2r1*p2^2r2*..........*p^2rk?
 
  • #6
Yes, all powers of the prime factors are even. Btw., here is how to write formulas here at PF:
https://www.physicsforums.com/help/latexhelp/

Let's start easy: what is ##\sigma(p^{2r})##? Then how do we get to ##\sigma (p_1^{2r_1}\cdot p_2^{2r_2})##?
 
  • #7
fresh_42 said:
Yes, all powers of the prime factors are even. Btw., here is how to write formulas here at PF:
https://www.physicsforums.com/help/latexhelp/

Let's start easy: what is ##\sigma(p^{2r})##? Then how do we get to ##\sigma (p_1^{2r_1}\cdot p_2^{2r_2})##?
Do you mean we can use this form p^a+1 -1/p-1 ? and since sigma is a multiplicative function, we can distribute the sigma?
 
  • #8
mathstudent1 said:
Do you mean we can use this form p^a+1 -1/p-1 ? and since sigma is a multiplicative function, we can distribute the sigma?
Use the formula you gave us:
$$
\sigma (p^a)=\dfrac{p^{a+1}-1}{p-1}
$$
All we know is that ##a## is even. This is not much. Can you perform the long division ##(p^{a+1}-1)\, : \,(p-1)##?

Proceed step by step. If we know that it is true with one prime, then we can use multiplicativity, yes.
 
  • #9
fresh_42 said:
Use the formula you gave us:
$$
\sigma (p^a)=\dfrac{p^{a+1}-1}{p-1}
$$
All we know is that ##a## is even. This is not much. Can you perform the long division ##(p^{a+1}-1)\, : \,(p-1)##?

Proceed step by step. If we know that it is true with one prime, then we can use multiplicativity, yes.
so we can use this form? 1+p+P^2+.......+p^a?
 
  • #10
mathstudent1 said:
so we can use this form? 1+p+P^2+.......+p^a?
Yes. How many terms are there? You probably have to distinguish between odd and even ##p## if you want to answer whether the sum is even or odd.

What is ##1+p+p^2+\ldots+p^a \pmod{2}##?
 
Last edited:
  • #11
fresh_42 said:
Yes. How many terms are there? You probably have to distinguish between odd and even ##p## if you want to answer whether the sum is even or odd.
a+1 terms and since a must be even we will have even+1= odd and since sigma is multiplicative, we can distribute the sigma so we have this odd*odd*.....= odd that is why sigma is odd right?
 
  • #12
mathstudent1 said:
a+1 terms and since a must be even we will have even+1= odd and since sigma is multiplicative, we can distribute the sigma so we have this odd*odd*.....= odd that is why sigma is odd right?
That's the idea but you must be a bit more precise. You should write it like

\begin{align*}
\sigma (p^{2r})=\dfrac{p^{2r+1}-1}{p-1}=\underbrace{1+p^1+p^2+\ldots+p^{2r}}_{2r+1\text{ terms}}
=\begin{cases}
\equiv 1 &\text{ if } p\equiv 0 \pmod{2}\\
\equiv (2r+1)\cdot 1\equiv 1 &\text{ if } p\equiv 1 \pmod{2}
\end{cases}
\end{align*}

because it has two different reasons why the sum is odd depending on whether ##p=2## or not.

Finally, write it step by step.

\begin{align*}
n&=p_1^{r_1}\cdot\ldots\cdot p_k^{r_k}\text{ with pairwise distinct primes }p_i \text{ and } r_i\equiv 0\pmod{2}\\
\sigma (n)&=\sigma (p_1^{r_1}\cdot\ldots\cdot p_k^{r_k})\\
&=\sigma (p_1^{r_1})\cdot\ldots\cdot \sigma (p_k^{r_k})\text{ by multiplicativity of }\sigma \\
&\equiv \underbrace{1\cdot 1\ldots \cdot 1}_{k\text{-times }}\\
&\equiv 1\pmod{2}
\end{align*}

You have had all the ingredients (the formula for ##p^a##, the multiplicativity of ##\sigma ##, and the result of the long division), you only had to put them into the right order. A general guideline is to start with a list of what you have and make conclusions step by step.

Not sure whether this helps (I have written better ones) ...
https://www.physicsforums.com/insights/how-most-proofs-are-structured-and-how-to-write-them/
 
  • #13
fresh_42 said:
That's the idea but you must be a bit more precise. You should write it like

\begin{align*}
\sigma (p^{2r})=\dfrac{p^{2r+1}-1}{p-1}=\underbrace{1+p^1+p^2+\ldots+p^{2r}}_{2r+1\text{ terms}}
=\begin{cases}
\equiv 1 &\text{ if } p\equiv 0 \pmod{2}\\
\equiv (2r+1)\cdot 1\equiv 1 &\text{ if } p\equiv 1 \pmod{2}
\end{cases}
\end{align*}

because it has two different reasons why the sum is odd depending on whether ##p=2## or not.

Finally, write it step by step.

\begin{align*}
n&=p_1^{r_1}\cdot\ldots\cdot p_k^{r_k}\text{ with pairwise distinct primes }p_i \text{ and } r_i\equiv 0\pmod{2}\\
\sigma (n)&=\sigma (p_1^{r_1}\cdot\ldots\cdot p_k^{r_k})\\
&=\sigma (p_1^{r_1})\cdot\ldots\cdot \sigma (p_k^{r_k})\text{ by multiplicativity of }\sigma \\
&\equiv \underbrace{1\cdot 1\ldots \cdot 1}_{k\text{-times }}\\
&\equiv 1\pmod{2}
\end{align*}

You have had all the ingredients (the formula for ##p^a##, the multiplicativity of ##\sigma ##, and the result of the long division), you only had to put them into the right order. A general guideline is to start with a list of what you have and make conclusions step by step.

Not sure whether this helps (I have written better ones) ...
https://www.physicsforums.com/insights/how-most-proofs-are-structured-and-how-to-write-them/
thank you
 
Back
Top