Each result cited is universally quantified over the data in its own statement, and is applied here to the data named below. We use silently that the order of R \mathbb{R} R is reflexive, transitive and antisymmetric, that equal real numbers satisfy ≤ \le ≤ in both directions, and that x < y x<y x < y entails x ≤ y x\le y x ≤ y . We use repeatedly the following consequence of claims 2 and 3 of Elementary Arithmetic in an Ordered Field : if a ≤ b a\le b a ≤ b and c ≤ d c\le d c ≤ d then a + c ≤ b + d a+c\le b+d a + c ≤ b + d , since 0 ≤ b − a 0\le b-a 0 ≤ b − a and 0 ≤ d − c 0\le d-c 0 ≤ d − c give 0 ≤ ( b + d ) − ( a + c ) 0\le(b+d)-(a+c) 0 ≤ ( b + d ) − ( a + c ) .
(N0) The absolute value of a nonnegative real. Let x x x be a real number with 0 ≤ x 0\le x 0 ≤ x . Then ∣ x ∣ = x |x|=x ∣ x ∣ = x . Indeed x ≤ ∣ x ∣ x\le|x| x ≤ ∣ x ∣ by claim 3 of Properties of the Absolute Value in an Ordered Field . Also − x ≤ − 0 -x\le -0 − x ≤ − 0 by claim 4 of Elementary Order Arithmetic in an Ordered Field , and − 0 = 0 -0=0 − 0 = 0 by claim 4 of Additive Cancellation and Elementary Additive Identities in a Field , so − x ≤ 0 -x\le 0 − x ≤ 0 and hence − x ≤ x -x\le x − x ≤ x by transitivity, while x ≤ x x\le x x ≤ x ; hence ∣ x ∣ ≤ x |x|\le x ∣ x ∣ ≤ x by claim 6 of Properties of the Absolute Value in an Ordered Field applied with c = x c=x c = x . Antisymmetry gives ∣ x ∣ = x |x|=x ∣ x ∣ = x .
(N1) Norms, squares and square roots. Let E E E be a real inner product space with inner product ⟨ ⋅ , ⋅ ⟩ E \langle\,\cdot\,,\cdot\,\rangle_{E} ⟨ ⋅ , ⋅ ⟩ E , norm ∥ ⋅ ∥ E \lVert\,\cdot\,\rVert_{E} ∥ ⋅ ∥ E and distance d E d_{E} d E . By Real Inner Product Spaces (ii) , 0 ≤ ∥ X ∥ E 0\le\lVert X\rVert_{E} 0 ≤ ∥ X ∥ E and the product ∥ X ∥ E ⋅ ∥ X ∥ E \lVert X\rVert_{E}\cdot\lVert X\rVert_{E} ∥ X ∥ E ⋅ ∥ X ∥ E equals ⟨ X , X ⟩ E \langle X,X\rangle_{E} ⟨ X , X ⟩ E ; by Properties of Real Powers (ii) that product is the power ∥ X ∥ E 2 \lVert X\rVert_{E}^{2} ∥ X ∥ E 2 of Real Powers (i) , so ∥ X ∥ E 2 = ⟨ X , X ⟩ E \lVert X\rVert_{E}^{2}=\langle X,X\rangle_{E} ∥ X ∥ E 2 = ⟨ X , X ⟩ E . By Properties of Real Powers (vi) , for nonnegative reals s s s and t t t we have ( t 2 ) 1 / 2 = t (t^{2})^{1/2}=t ( t 2 ) 1/2 = t , ( t 1 / 2 ) 2 = t (t^{1/2})^{2}=t ( t 1/2 ) 2 = t , and s 2 ≤ t s^{2}\le t s 2 ≤ t if and only if s ≤ t 1 / 2 s\le t^{1/2} s ≤ t 1/2 ; consequently s 2 ≤ t 2 s^{2}\le t^{2} s 2 ≤ t 2 implies s ≤ t s\le t s ≤ t . By Properties of Real Powers (i) a power of a nonnegative real is nonnegative, and 0 1 / 2 = 0 0^{1/2}=0 0 1/2 = 0 . By Real Inner Product Spaces (iii) , d E ( X , Y ) = ∥ X − Y ∥ E d_{E}(X,Y)=\lVert X-Y\rVert_{E} d E ( X , Y ) = ∥ X − Y ∥ E . We apply all of this to L 2 ( T n ) L^{2}(\mathbb{T}^{n}) L 2 ( T n ) , which is a real inner product space by L²(μ) Is a Real Hilbert Space (i) , and to H 1 ( T n ) H^{1}(\mathbb{T}^{n}) H 1 ( T n ) , which is one by The Sobolev Space of Once Weakly Differentiable Square-Integrable Classes on the Torus (ii) , without further mention.
(N2) The decomposition of the Sobolev norm and distance. Let U , U ′ ∈ H 1 ( T n ) U,U'\in H^{1}(\mathbb{T}^{n}) U , U ′ ∈ H 1 ( T n ) . By The Sobolev Space of Once Weakly Differentiable Square-Integrable Classes on the Torus (ii) and (N1),
∥ U ∥ H 1 2 = ∥ U ∥ L 2 2 + ∑ j = 1 n ∥ ∂ j U ∥ L 2 2 . \lVert U\rVert_{H^{1}}^{2}=\lVert U\rVert_{L^{2}}^{2}+\sum_{j=1}^{n}\lVert\partial_{j}U\rVert_{L^{2}}^{2}. ∥ U ∥ H 1 2 = ∥ U ∥ L 2 2 + j = 1 ∑ n ∥ ∂ j U ∥ L 2 2 .
Moreover U − U ′ ∈ H 1 ( T n ) U-U'\in H^{1}(\mathbb{T}^{n}) U − U ′ ∈ H 1 ( T n ) , H 1 ( T n ) H^{1}(\mathbb{T}^{n}) H 1 ( T n ) being a linear subspace of L 2 ( T n ) L^{2}(\mathbb{T}^{n}) L 2 ( T n ) by The Sobolev Space of Once Weakly Differentiable Square-Integrable Classes on the Torus (i) , and ∂ j ( U − U ′ ) = ∂ j U − ∂ j U ′ \partial_{j}(U-U')=\partial_{j}U-\partial_{j}U' ∂ j ( U − U ′ ) = ∂ j U − ∂ j U ′ for every j ∈ [ n ] j\in[n] j ∈ [ n ] by Elementary Properties of the Weak Partial Derivative on the Torus (ii) applied with c = − 1 c=-1 c = − 1 . Applying the display to U − U ′ U-U' U − U ′ and using (N1) therefore gives
d H 1 ( U , U ′ ) 2 = d L 2 ( U , U ′ ) 2 + ∑ j = 1 n d L 2 ( ∂ j U , ∂ j U ′ ) 2 . d_{H^{1}}(U,U')^{2}=d_{L^{2}}(U,U')^{2}+\sum_{j=1}^{n}d_{L^{2}}(\partial_{j}U,\partial_{j}U')^{2}. d H 1 ( U , U ′ ) 2 = d L 2 ( U , U ′ ) 2 + j = 1 ∑ n d L 2 ( ∂ j U , ∂ j U ′ ) 2 .
(N3) Convergence and the sequence of distances. Let ( Y , d ) (Y,d) ( Y , d ) be a metric space, let ( X m ) m ∈ N (X_{m})_{m\in\mathbb{N}} ( X m ) m ∈ N be a sequence in Y Y Y and let X ∈ Y X\in Y X ∈ Y . Then ( X m ) (X_{m}) ( X m ) converges to X X X in ( Y , d ) (Y,d) ( Y , d ) if and only if the sequence of real numbers ( d ( X m , X ) ) m ∈ N \bigl(d(X_{m},X)\bigr)_{m\in\mathbb{N}} ( d ( X m , X ) ) m ∈ N converges to 0 0 0 . Indeed the two conditions are the same assertion about ε \varepsilon ε and N N N , because 0 ≤ d ( X m , X ) 0\le d(X_{m},X) 0 ≤ d ( X m , X ) by Metric Space and hence ∣ d ( X m , X ) − 0 ∣ = d ( X m , X ) |d(X_{m},X)-0|=d(X_{m},X) ∣ d ( X m , X ) − 0∣ = d ( X m , X ) by (N0), the difference d ( X m , X ) − 0 d(X_{m},X)-0 d ( X m , X ) − 0 being d ( X m , X ) d(X_{m},X) d ( X m , X ) by claim 4 of Additive Cancellation and Elementary Additive Identities in a Field .
(N4) A convergence criterion in the Sobolev space. Let U U U and, for every m ∈ N m\in\mathbb{N} m ∈ N , U m U_{m} U m belong to H 1 ( T n ) H^{1}(\mathbb{T}^{n}) H 1 ( T n ) , and suppose that ( U m ) (U_{m}) ( U m ) converges to U U U in L 2 ( T n ) L^{2}(\mathbb{T}^{n}) L 2 ( T n ) and that, for every j ∈ [ n ] j\in[n] j ∈ [ n ] , ( ∂ j U m ) (\partial_{j}U_{m}) ( ∂ j U m ) converges to ∂ j U \partial_{j}U ∂ j U in L 2 ( T n ) L^{2}(\mathbb{T}^{n}) L 2 ( T n ) . Then ( U m ) (U_{m}) ( U m ) converges to U U U in H 1 ( T n ) H^{1}(\mathbb{T}^{n}) H 1 ( T n ) .
To see this, put b m = d L 2 ( U m , U ) b_{m}=d_{L^{2}}(U_{m},U) b m = d L 2 ( U m , U ) and b j , m = d L 2 ( ∂ j U m , ∂ j U ) b_{j,m}=d_{L^{2}}(\partial_{j}U_{m},\partial_{j}U) b j , m = d L 2 ( ∂ j U m , ∂ j U ) for j ∈ [ n ] j\in[n] j ∈ [ n ] . By (N3) the real sequences ( b m ) m (b_{m})_{m} ( b m ) m and ( b j , m ) m (b_{j,m})_{m} ( b j , m ) m converge to 0 0 0 , so by claim 2 of Arithmetic of Limits of Real Sequences the sequences ( b m 2 ) m (b_{m}^{2})_{m} ( b m 2 ) m and ( b j , m 2 ) m (b_{j,m}^{2})_{m} ( b j , m 2 ) m converge to 0 ⋅ 0 = 0 0\cdot 0=0 0 ⋅ 0 = 0 , the squares being products by Properties of Real Powers (ii) . By Finite Sums of Real Numbers: Nonnegativity, Domination by the Sum, and Limits (iii) , applied with N = n N=n N = n and the constant limits 0 0 0 , the sequence whose m m m th term is ∑ j = 1 n b j , m 2 \sum_{j=1}^{n}b_{j,m}^{2} ∑ j = 1 n b j , m 2 converges to ∑ j = 1 n 0 \sum_{j=1}^{n}0 ∑ j = 1 n 0 , which is 0 0 0 by the second assertion of Finite Sums of Real Numbers: Nonnegativity, Domination by the Sum, and Limits (i) . By claim 1 of Arithmetic of Limits of Real Sequences and (N2), the sequence whose m m m th term is d H 1 ( U m , U ) 2 d_{H^{1}}(U_{m},U)^{2} d H 1 ( U m , U ) 2 converges to 0 + 0 = 0 0+0=0 0 + 0 = 0 . Applying Properties of Real Powers (vii) with exponent 1 / 2 1/2 1/2 to this sequence of nonnegative reals, the sequence whose m m m th term is ( d H 1 ( U m , U ) 2 ) 1 / 2 \bigl(d_{H^{1}}(U_{m},U)^{2}\bigr)^{1/2} ( d H 1 ( U m , U ) 2 ) 1/2 converges to 0 1 / 2 0^{1/2} 0 1/2 ; by (N1) that m m m th term is d H 1 ( U m , U ) d_{H^{1}}(U_{m},U) d H 1 ( U m , U ) and 0 1 / 2 = 0 0^{1/2}=0 0 1/2 = 0 . By (N3) again, ( U m ) (U_{m}) ( U m ) converges to U U U in H 1 ( T n ) H^{1}(\mathbb{T}^{n}) H 1 ( T n ) .
Proof of claim 1. Let U , U ′ ∈ H 1 ( T n ) U,U'\in H^{1}(\mathbb{T}^{n}) U , U ′ ∈ H 1 ( T n ) and let j ∈ [ n ] j\in[n] j ∈ [ n ] . For k ∈ [ n ] k\in[n] k ∈ [ n ] put t k = ∥ ∂ k U ∥ L 2 2 t_{k}=\lVert\partial_{k}U\rVert_{L^{2}}^{2} t k = ∥ ∂ k U ∥ L 2 2 , a nonnegative real by (N1), and put Σ = ∑ k = 1 n t k \Sigma=\sum_{k=1}^{n}t_{k} Σ = ∑ k = 1 n t k .
By Finite Sums of Real Numbers: Nonnegativity, Domination by the Sum, and Limits (i) , 0 ≤ Σ 0\le\Sigma 0 ≤ Σ . By (N2), ∥ U ∥ H 1 2 − ∥ U ∥ L 2 2 = Σ \lVert U\rVert_{H^{1}}^{2}-\lVert U\rVert_{L^{2}}^{2}=\Sigma ∥ U ∥ H 1 2 − ∥ U ∥ L 2 2 = Σ , so claim 3 of Elementary Arithmetic in an Ordered Field gives ∥ U ∥ L 2 2 ≤ ∥ U ∥ H 1 2 \lVert U\rVert_{L^{2}}^{2}\le\lVert U\rVert_{H^{1}}^{2} ∥ U ∥ L 2 2 ≤ ∥ U ∥ H 1 2 , and (N1) gives ∥ U ∥ L 2 ≤ ∥ U ∥ H 1 \lVert U\rVert_{L^{2}}\le\lVert U\rVert_{H^{1}} ∥ U ∥ L 2 ≤ ∥ U ∥ H 1 .
By Finite Sums of Real Numbers: Nonnegativity, Domination by the Sum, and Limits (ii) , applied with N = n N=n N = n , the index n n n in place of its j j j and the index j j j in place of its i i i , we get t j ≤ Σ t_{j}\le\Sigma t j ≤ Σ . Also ∥ U ∥ H 1 2 − Σ = ∥ U ∥ L 2 2 \lVert U\rVert_{H^{1}}^{2}-\Sigma=\lVert U\rVert_{L^{2}}^{2} ∥ U ∥ H 1 2 − Σ = ∥ U ∥ L 2 2 by (N2), which is nonnegative by (N1), so Σ ≤ ∥ U ∥ H 1 2 \Sigma\le\lVert U\rVert_{H^{1}}^{2} Σ ≤ ∥ U ∥ H 1 2 by claim 3 of Elementary Arithmetic in an Ordered Field . By transitivity ∥ ∂ j U ∥ L 2 2 ≤ ∥ U ∥ H 1 2 \lVert\partial_{j}U\rVert_{L^{2}}^{2}\le\lVert U\rVert_{H^{1}}^{2} ∥ ∂ j U ∥ L 2 2 ≤ ∥ U ∥ H 1 2 , and (N1) gives ∥ ∂ j U ∥ L 2 ≤ ∥ U ∥ H 1 \lVert\partial_{j}U\rVert_{L^{2}}\le\lVert U\rVert_{H^{1}} ∥ ∂ j U ∥ L 2 ≤ ∥ U ∥ H 1 .
For the distance inequalities, apply the two norm inequalities just proved to U − U ′ U-U' U − U ′ , which lies in H 1 ( T n ) H^{1}(\mathbb{T}^{n}) H 1 ( T n ) by (N2). Using (N1) and the identity ∂ j ( U − U ′ ) = ∂ j U − ∂ j U ′ \partial_{j}(U-U')=\partial_{j}U-\partial_{j}U' ∂ j ( U − U ′ ) = ∂ j U − ∂ j U ′ of (N2),
d L 2 ( U , U ′ ) = ∥ U − U ′ ∥ L 2 ≤ ∥ U − U ′ ∥ H 1 = d H 1 ( U , U ′ ) , d_{L^{2}}(U,U')=\lVert U-U'\rVert_{L^{2}}\le\lVert U-U'\rVert_{H^{1}}=d_{H^{1}}(U,U'), d L 2 ( U , U ′ ) = ∥ U − U ′ ∥ L 2 ≤ ∥ U − U ′ ∥ H 1 = d H 1 ( U , U ′ ) ,
d L 2 ( ∂ j U , ∂ j U ′ ) = ∥ ∂ j ( U − U ′ ) ∥ L 2 ≤ ∥ U − U ′ ∥ H 1 = d H 1 ( U , U ′ ) . d_{L^{2}}(\partial_{j}U,\partial_{j}U')=\lVert\partial_{j}(U-U')\rVert_{L^{2}}\le\lVert U-U'\rVert_{H^{1}}=d_{H^{1}}(U,U'). d L 2 ( ∂ j U , ∂ j U ′ ) = ∥ ∂ j ( U − U ′ ) ∥ L 2 ≤ ∥ U − U ′ ∥ H 1 = d H 1 ( U , U ′ ) .
Proof of claim 2. By The Sobolev Space of Once Weakly Differentiable Square-Integrable Classes on the Torus (ii) , H 1 ( T n ) H^{1}(\mathbb{T}^{n}) H 1 ( T n ) with ⟨ ⋅ , ⋅ ⟩ H 1 \langle\,\cdot\,,\cdot\,\rangle_{H^{1}} ⟨ ⋅ , ⋅ ⟩ H 1 is a real inner product space whose distance is d H 1 d_{H^{1}} d H 1 , and d H 1 d_{H^{1}} d H 1 is a metric on it by The Norm Metric of a Real Inner Product Space: Triangle Inequalities, Limits and Continuity (iii) . By Real Hilbert Spaces (i) it therefore suffices to prove that the metric space ( H 1 ( T n ) , d H 1 ) (H^{1}(\mathbb{T}^{n}),d_{H^{1}}) ( H 1 ( T n ) , d H 1 ) is complete .
Let ( U m ) m ∈ N (U_{m})_{m\in\mathbb{N}} ( U m ) m ∈ N be a Cauchy sequence in ( H 1 ( T n ) , d H 1 ) (H^{1}(\mathbb{T}^{n}),d_{H^{1}}) ( H 1 ( T n ) , d H 1 ) and let j ∈ [ n ] j\in[n] j ∈ [ n ] be arbitrary; the argument that follows uses nothing about j j j beyond j ∈ [ n ] j\in[n] j ∈ [ n ] , so its conclusion holds for every j ∈ [ n ] j\in[n] j ∈ [ n ] . Given a real ε > 0 \varepsilon>0 ε > 0 , choose N ∈ N N\in\mathbb{N} N ∈ N with d H 1 ( U m , U ℓ ) < ε d_{H^{1}}(U_{m},U_{\ell})<\varepsilon d H 1 ( U m , U ℓ ) < ε for all m , ℓ ∈ N m,\ell\in\mathbb{N} m , ℓ ∈ N with m ≥ N m\ge N m ≥ N and ℓ ≥ N \ell\ge N ℓ ≥ N . By claim 1, for such m m m and ℓ \ell ℓ we have d L 2 ( U m , U ℓ ) ≤ d H 1 ( U m , U ℓ ) < ε d_{L^{2}}(U_{m},U_{\ell})\le d_{H^{1}}(U_{m},U_{\ell})<\varepsilon d L 2 ( U m , U ℓ ) ≤ d H 1 ( U m , U ℓ ) < ε and d L 2 ( ∂ j U m , ∂ j U ℓ ) ≤ d H 1 ( U m , U ℓ ) < ε d_{L^{2}}(\partial_{j}U_{m},\partial_{j}U_{\ell})\le d_{H^{1}}(U_{m},U_{\ell})<\varepsilon d L 2 ( ∂ j U m , ∂ j U ℓ ) ≤ d H 1 ( U m , U ℓ ) < ε . Hence ( U m ) (U_{m}) ( U m ) and ( ∂ j U m ) (\partial_{j}U_{m}) ( ∂ j U m ) are Cauchy sequences in L 2 ( T n ) L^{2}(\mathbb{T}^{n}) L 2 ( T n ) with the distance d L 2 d_{L^{2}} d L 2 .
By The Flat Torus: Standing Notation (iv) the space L 2 ( T n ) L^{2}(\mathbb{T}^{n}) L 2 ( T n ) is a real Banach space , so by that clause its metric space is complete; by the preamble that metric is d L 2 d_{L^{2}} d L 2 . Hence there is U ∈ L 2 ( T n ) U\in L^{2}(\mathbb{T}^{n}) U ∈ L 2 ( T n ) to which ( U m ) (U_{m}) ( U m ) converges, and for each j ∈ [ n ] j\in[n] j ∈ [ n ] there is G j ∈ L 2 ( T n ) G_{j}\in L^{2}(\mathbb{T}^{n}) G j ∈ L 2 ( T n ) to which ( ∂ j U m ) (\partial_{j}U_{m}) ( ∂ j U m ) converges; the n n n choices of G j G_{j} G j are finite in number and require no appeal to a choice principle beyond selecting one object for each of finitely many indices.
Fix j ∈ [ n ] j\in[n] j ∈ [ n ] . The classes U m U_{m} U m lie in L 2 ( T n ) L^{2}(\mathbb{T}^{n}) L 2 ( T n ) and ∂ j U m \partial_{j}U_{m} ∂ j U m is the j j j -th weak partial derivative of U m U_{m} U m for every m m m , and ( U m ) (U_{m}) ( U m ) and ( ∂ j U m ) (\partial_{j}U_{m}) ( ∂ j U m ) converge in L 2 ( T n ) L^{2}(\mathbb{T}^{n}) L 2 ( T n ) to U U U and to G j G_{j} G j ; so Elementary Properties of the Weak Partial Derivative on the Torus (v) shows that G j G_{j} G j is the j j j -th weak partial derivative of U U U . As this holds for every j ∈ [ n ] j\in[n] j ∈ [ n ] , The Sobolev Space of Once Weakly Differentiable Square-Integrable Classes on the Torus (i) gives U ∈ H 1 ( T n ) U\in H^{1}(\mathbb{T}^{n}) U ∈ H 1 ( T n ) , and ∂ j U = G j \partial_{j}U=G_{j} ∂ j U = G j by the uniqueness recorded in The Weak Partial Derivative on the Torus (ii) . By (N4), ( U m ) (U_{m}) ( U m ) converges to U U U in H 1 ( T n ) H^{1}(\mathbb{T}^{n}) H 1 ( T n ) . Every Cauchy sequence therefore converges, as required.
Proof of claim 3. Let k k k be a natural number. By Rescaling a Mollifier Kernel , ρ 1 / k \rho_{1/k} ρ 1/ k is a mollifier kernel of radius R = ( 1 / k ) δ R=(1/k)\,\delta R = ( 1/ k ) δ on R n \mathbb{R}^{n} R n , and 0 < R 0<R 0 < R by claim 5 of Elementary Order Arithmetic in an Ordered Field . By Mollifier Kernel of Radius δ \delta δ on R n \mathbb{R}^n R n the map ρ 1 / k \rho_{1/k} ρ 1/ k is smooth, satisfies 0 ≤ ρ 1 / k ( y ) 0\le\rho_{1/k}(y) 0 ≤ ρ 1/ k ( y ) for every y ∈ R n y\in\mathbb{R}^{n} y ∈ R n , vanishes at every y y y with R < ∥ y ∥ R<\lVert y\rVert R < ∥ y ∥ , and has ∫ R n ρ 1 / k d λ n = 1 \int_{\mathbb{R}^{n}}\rho_{1/k}\,d\lambda_{n}=1 ∫ R n ρ 1/ k d λ n = 1 ; being smooth it is continuous by claim 3 of Euclidean Space Is Open in Itself . So ρ 1 / k \rho_{1/k} ρ 1/ k meets the standing hypotheses imposed on the kernel in Properties of Periodic Convolution on the Torus and in The Lebesgue Bound for Periodic Convolution and Convergence of Mollifications , with this R R R .
We record once that C p e r ∞ ⊆ C p e r 1 C^{\infty}_{\mathrm{per}}\subseteq C^{1}_{\mathrm{per}} C per ∞ ⊆ C per 1 : a smooth map on R n \mathbb{R}^{n} R n is of class C 1 C^{1} C 1 by claim 2 of Euclidean Space Is Open in Itself , and periodicity is the same condition for the two classes by Lattice-Periodic Functions and the Periodic Function Classes (iii) .
Part (a). Since u ∈ L 2 ( T n ) u\in\mathcal{L}^{2}(\mathbb{T}^{n}) u ∈ L 2 ( T n ) and ρ 1 / k \rho_{1/k} ρ 1/ k is smooth, Properties of Periodic Convolution on the Torus (i) gives ρ 1 / k ⋆ u ∈ C p e r ∞ \rho_{1/k}\star u\in C^{\infty}_{\mathrm{per}} ρ 1/ k ⋆ u ∈ C per ∞ , hence ρ 1 / k ⋆ u ∈ C p e r 1 \rho_{1/k}\star u\in C^{1}_{\mathrm{per}} ρ 1/ k ⋆ u ∈ C per 1 . By The Sobolev Space of Once Weakly Differentiable Square-Integrable Classes on the Torus (i) , U k ∈ H 1 ( T n ) U_{k}\in H^{1}(\mathbb{T}^{n}) U k ∈ H 1 ( T n ) , and by Elementary Properties of the Weak Partial Derivative on the Torus (i) the j j j -th weak partial derivative of U k U_{k} U k is the class of ( ∂ j ( ρ 1 / k ⋆ u ) ) ∣ Q \bigl(\partial_{j}(\rho_{1/k}\star u)\bigr)|_{Q} ( ∂ j ( ρ 1/ k ⋆ u ) ) ∣ Q for every j ∈ [ n ] j\in[n] j ∈ [ n ] . By Elementary Properties of the Weak Partial Derivative on the Torus (iv) , applied to U U U , to ∂ j U \partial_{j}U ∂ j U , to the representatives u u u and g j g_{j} g j , and to the kernel ρ \rho ρ with parameter 1 / k 1/k 1/ k , we have ∂ j ( ρ 1 / k ⋆ u ) = ρ 1 / k ⋆ g j \partial_{j}(\rho_{1/k}\star u)=\rho_{1/k}\star g_{j} ∂ j ( ρ 1/ k ⋆ u ) = ρ 1/ k ⋆ g j as maps on R n \mathbb{R}^{n} R n . Hence ∂ j U k = [ ( ρ 1 / k ⋆ g j ) ∣ Q ] \partial_{j}U_{k}=[(\rho_{1/k}\star g_{j})|_{Q}] ∂ j U k = [( ρ 1/ k ⋆ g j ) ∣ Q ] .
Part (b). Let K K K be the integral of ∣ ρ 1 / k ∣ |\rho_{1/k}| ∣ ρ 1/ k ∣ over R n \mathbb{R}^{n} R n . Since 0 ≤ ρ 1 / k ( y ) 0\le\rho_{1/k}(y) 0 ≤ ρ 1/ k ( y ) for every y y y , (N0) gives ∣ ρ 1 / k ( y ) ∣ = ρ 1 / k ( y ) |\rho_{1/k}(y)|=\rho_{1/k}(y) ∣ ρ 1/ k ( y ) ∣ = ρ 1/ k ( y ) for every y y y , so ∣ ρ 1 / k ∣ |\rho_{1/k}| ∣ ρ 1/ k ∣ and ρ 1 / k \rho_{1/k} ρ 1/ k are the same map and K = 1 K=1 K = 1 . Applying The Lebesgue Bound for Periodic Convolution and Convergence of Mollifications (i) with p = 2 p=2 p = 2 to u u u and to each g j g_{j} g j , and using 1 ⋅ t = t 1\cdot t=t 1 ⋅ t = t ,
∥ ( ρ 1 / k ⋆ u ) ∣ Q ∥ 2 ≤ ∥ u ∥ 2 , ∥ ( ρ 1 / k ⋆ g j ) ∣ Q ∥ 2 ≤ ∥ g j ∥ 2 ( j ∈ [ n ] ) . \bigl\lVert(\rho_{1/k}\star u)|_{Q}\bigr\rVert_{2}\le\lVert u\rVert_{2},\qquad\bigl\lVert(\rho_{1/k}\star g_{j})|_{Q}\bigr\rVert_{2}\le\lVert g_{j}\rVert_{2}\quad(j\in[n]). ( ρ 1/ k ⋆ u ) ∣ Q 2 ≤ ∥ u ∥ 2 , ( ρ 1/ k ⋆ g j ) ∣ Q 2 ≤ ∥ g j ∥ 2 ( j ∈ [ n ]) .
By The Lebesgue Space Lᵖ(μ) (iii) and L²(μ) Is a Real Hilbert Space (i) these are the inequalities ∥ U k ∥ L 2 ≤ ∥ U ∥ L 2 \lVert U_{k}\rVert_{L^{2}}\le\lVert U\rVert_{L^{2}} ∥ U k ∥ L 2 ≤ ∥ U ∥ L 2 and ∥ ∂ j U k ∥ L 2 ≤ ∥ ∂ j U ∥ L 2 \lVert\partial_{j}U_{k}\rVert_{L^{2}}\le\lVert\partial_{j}U\rVert_{L^{2}} ∥ ∂ j U k ∥ L 2 ≤ ∥ ∂ j U ∥ L 2 , by part (a). Squaring, which preserves ≤ \le ≤ between nonnegative reals by Properties of Real Powers (v) , and comparing the two sums term by term by claim 1 of Comparison and Absolute Value Bounds for Finite Sums of Real Numbers ,
∥ U k ∥ L 2 2 ≤ ∥ U ∥ L 2 2 , ∑ j = 1 n ∥ ∂ j U k ∥ L 2 2 ≤ ∑ j = 1 n ∥ ∂ j U ∥ L 2 2 . \lVert U_{k}\rVert_{L^{2}}^{2}\le\lVert U\rVert_{L^{2}}^{2},\qquad\sum_{j=1}^{n}\lVert\partial_{j}U_{k}\rVert_{L^{2}}^{2}\le\sum_{j=1}^{n}\lVert\partial_{j}U\rVert_{L^{2}}^{2}. ∥ U k ∥ L 2 2 ≤ ∥ U ∥ L 2 2 , j = 1 ∑ n ∥ ∂ j U k ∥ L 2 2 ≤ j = 1 ∑ n ∥ ∂ j U ∥ L 2 2 .
Adding these two inequalities and using (N2) gives ∥ U k ∥ H 1 2 ≤ ∥ U ∥ H 1 2 \lVert U_{k}\rVert_{H^{1}}^{2}\le\lVert U\rVert_{H^{1}}^{2} ∥ U k ∥ H 1 2 ≤ ∥ U ∥ H 1 2 , whence ∥ U k ∥ H 1 ≤ ∥ U ∥ H 1 \lVert U_{k}\rVert_{H^{1}}\le\lVert U\rVert_{H^{1}} ∥ U k ∥ H 1 ≤ ∥ U ∥ H 1 by (N1).
Convergence. By The Lebesgue Bound for Periodic Convolution and Convergence of Mollifications (ii) with p = 2 p=2 p = 2 , applied to u u u , the sequence whose k k k th term is the class of ( ρ 1 / k ⋆ u ) ∣ Q (\rho_{1/k}\star u)|_{Q} ( ρ 1/ k ⋆ u ) ∣ Q , that is U k U_{k} U k , converges to [ u ] = U [u]=U [ u ] = U in L 2 ( T n ) L^{2}(\mathbb{T}^{n}) L 2 ( T n ) ; applied to g j g_{j} g j , the sequence whose k k k th term is the class of ( ρ 1 / k ⋆ g j ) ∣ Q (\rho_{1/k}\star g_{j})|_{Q} ( ρ 1/ k ⋆ g j ) ∣ Q , that is ∂ j U k \partial_{j}U_{k} ∂ j U k by part (a), converges to [ g j ] = ∂ j U [g_{j}]=\partial_{j}U [ g j ] = ∂ j U . By (N4), ( U k ) k ∈ N (U_{k})_{k\in\mathbb{N}} ( U k ) k ∈ N converges to U U U in H 1 ( T n ) H^{1}(\mathbb{T}^{n}) H 1 ( T n ) .
Proof of claim 4. Write S 0 = { [ w ∣ Q ] : w ∈ C p e r ∞ } \mathcal{S}_{0}=\{[w|_{Q}]:w\in C^{\infty}_{\mathrm{per}}\} S 0 = {[ w ∣ Q ] : w ∈ C per ∞ } . Since C p e r ∞ ⊆ C p e r 1 C^{\infty}_{\mathrm{per}}\subseteq C^{1}_{\mathrm{per}} C per ∞ ⊆ C per 1 , as recorded in the proof of claim 3, The Sobolev Space of Once Weakly Differentiable Square-Integrable Classes on the Torus (i) gives S 0 ⊆ H 1 ( T n ) \mathcal{S}_{0}\subseteq H^{1}(\mathbb{T}^{n}) S 0 ⊆ H 1 ( T n ) .
Let U ∈ H 1 ( T n ) U\in H^{1}(\mathbb{T}^{n}) U ∈ H 1 ( T n ) and let ε \varepsilon ε be a real number with 0 < ε 0<\varepsilon 0 < ε . By Existence of Mollifier Kernels of Every Radius there is a mollifier kernel ρ \rho ρ of radius 1 1 1 on R n \mathbb{R}^{n} R n . Choose a representative u u u of U U U and, for each j ∈ [ n ] j\in[n] j ∈ [ n ] , a representative g j g_{j} g j of ∂ j U \partial_{j}U ∂ j U . By claim 3, applied with δ = 1 \delta=1 δ = 1 , the sequence ( U k ) k ∈ N (U_{k})_{k\in\mathbb{N}} ( U k ) k ∈ N converges to U U U in H 1 ( T n ) H^{1}(\mathbb{T}^{n}) H 1 ( T n ) , so by Convergent Sequence in a Metric Space there is k ∈ N k\in\mathbb{N} k ∈ N with d H 1 ( U k , U ) < ε d_{H^{1}}(U_{k},U)<\varepsilon d H 1 ( U k , U ) < ε . By part (a) of claim 3, ρ 1 / k ⋆ u ∈ C p e r ∞ \rho_{1/k}\star u\in C^{\infty}_{\mathrm{per}} ρ 1/ k ⋆ u ∈ C per ∞ , so U k ∈ S 0 U_{k}\in\mathcal{S}_{0} U k ∈ S 0 .
Thus for every real ε > 0 \varepsilon>0 ε > 0 there is an element of S 0 \mathcal{S}_{0} S 0 at d H 1 d_{H^{1}} d H 1 -distance less than ε \varepsilon ε from U U U ; by condition 3 of Characterization of the Closure in a Metric Space by Open Balls , U U U belongs to the closure of S 0 \mathcal{S}_{0} S 0 in ( H 1 ( T n ) , d H 1 ) (H^{1}(\mathbb{T}^{n}),d_{H^{1}}) ( H 1 ( T n ) , d H 1 ) . As U U U was arbitrary, that closure is H 1 ( T n ) H^{1}(\mathbb{T}^{n}) H 1 ( T n ) , so S 0 \mathcal{S}_{0} S 0 is dense in H 1 ( T n ) H^{1}(\mathbb{T}^{n}) H 1 ( T n ) .
Proof of claim 5. By Existence of Mollifier Kernels of Every Radius fix a mollifier kernel ρ \rho ρ of radius 1 1 1 on R n \mathbb{R}^{n} R n , and for k ∈ N k\in\mathbb{N} k ∈ N let ρ 1 / k \rho_{1/k} ρ 1/ k be its rescaling, a mollifier kernel of radius R k = 1 / k R_{k}=1/k R k = 1/ k on R n \mathbb{R}^{n} R n by Rescaling a Mollifier Kernel , with 0 < R k 0<R_{k} 0 < R k . Let D D D be the set of maps furnished by Separability of the Lebesgue Spaces of the Torus for p = 2 p=2 p = 2 ; by Separability of the Lebesgue Spaces of the Torus (ii) every s ∈ D s\in D s ∈ D lies in L 2 ( T n ) \mathcal{L}^{2}(\mathbb{T}^{n}) L 2 ( T n ) and the set { [ s ] : s ∈ D } \{[s]:s\in D\} {[ s ] : s ∈ D } is countable . Put
S = { [ ( ρ 1 / k ⋆ s ) ∣ Q ] : s ∈ D , k ∈ N } . \mathcal{S}=\bigl\{\,[(\rho_{1/k}\star s)|_{Q}]\ :\ s\in D,\ k\in\mathbb{N}\,\bigr\}. S = { [( ρ 1/ k ⋆ s ) ∣ Q ] : s ∈ D , k ∈ N } .
S \mathcal{S} S is a subset of H 1 ( T n ) H^{1}(\mathbb{T}^{n}) H 1 ( T n ) . As in part (a) of the proof of claim 3, ρ 1 / k ⋆ s ∈ C p e r ∞ ⊆ C p e r 1 \rho_{1/k}\star s\in C^{\infty}_{\mathrm{per}}\subseteq C^{1}_{\mathrm{per}} ρ 1/ k ⋆ s ∈ C per ∞ ⊆ C per 1 for every s ∈ D s\in D s ∈ D and every k k k , so The Sobolev Space of Once Weakly Differentiable Square-Integrable Classes on the Torus (i) applies.
S \mathcal{S} S is countable. By Properties of Periodic Convolution on the Torus (ii) the map ρ 1 / k ⋆ s \rho_{1/k}\star s ρ 1/ k ⋆ s depends only on the class [ s ] [s] [ s ] , so there is a well-defined map Φ \Phi Φ from { [ s ] : s ∈ D } × N \{[s]:s\in D\}\times\mathbb{N} {[ s ] : s ∈ D } × N to S \mathcal{S} S with Φ ( [ s ] , k ) = [ ( ρ 1 / k ⋆ s ) ∣ Q ] \Phi([s],k)=[(\rho_{1/k}\star s)|_{Q}] Φ ([ s ] , k ) = [( ρ 1/ k ⋆ s ) ∣ Q ] , and S \mathcal{S} S is exactly the set of its values. The set N \mathbb{N} N is countable by claim 1 of Basic Properties of Countable Sets , so the Cartesian product { [ s ] : s ∈ D } × N \{[s]:s\in D\}\times\mathbb{N} {[ s ] : s ∈ D } × N is countable by claim 1 of Products and Powers of Countable Sets , and S \mathcal{S} S is countable by claim 4 of Basic Properties of Countable Sets .
S \mathcal{S} S is dense. Let U ∈ H 1 ( T n ) U\in H^{1}(\mathbb{T}^{n}) U ∈ H 1 ( T n ) and let ε \varepsilon ε be a real number with 0 < ε 0<\varepsilon 0 < ε . Choose a representative u u u of U U U and, for each j ∈ [ n ] j\in[n] j ∈ [ n ] , a representative g j g_{j} g j of ∂ j U \partial_{j}U ∂ j U . By claim 3 applied with δ = 1 \delta=1 δ = 1 , the sequence ( U k ) k ∈ N (U_{k})_{k\in\mathbb{N}} ( U k ) k ∈ N converges to U U U in H 1 ( T n ) H^{1}(\mathbb{T}^{n}) H 1 ( T n ) , so there is k ∈ N k\in\mathbb{N} k ∈ N with d H 1 ( U k , U ) < ε / 2 d_{H^{1}}(U_{k},U)<\varepsilon/2 d H 1 ( U k , U ) < ε /2 , where U k = [ ( ρ 1 / k ⋆ u ) ∣ Q ] U_{k}=[(\rho_{1/k}\star u)|_{Q}] U k = [( ρ 1/ k ⋆ u ) ∣ Q ] . Fix this k k k for the rest of the argument.
For j ∈ [ n ] j\in[n] j ∈ [ n ] , the map ∂ j ρ 1 / k \partial_{j}\rho_{1/k} ∂ j ρ 1/ k is continuous and vanishes at every y y y with R k < ∥ y ∥ R_{k}<\lVert y\rVert R k < ∥ y ∥ , by Convolution of a Locally Integrable Function with a Compactly Supported Kernel (iv) applied to the kernel ρ 1 / k \rho_{1/k} ρ 1/ k , which is of class C 1 C^{1} C 1 by claim 2 of Euclidean Space Is Open in Itself . Let K j K_{j} K j be the integral of ∣ ∂ j ρ 1 / k ∣ |\partial_{j}\rho_{1/k}| ∣ ∂ j ρ 1/ k ∣ over R n \mathbb{R}^{n} R n , a nonnegative real number by that clause, and put
C = ( 1 + ∑ j = 1 n K j 2 ) 1 / 2 . C=\Bigl(1+\sum_{j=1}^{n}K_{j}^{2}\Bigr)^{1/2}. C = ( 1 + j = 1 ∑ n K j 2 ) 1/2 .
The number 1 + ∑ j = 1 n K j 2 1+\sum_{j=1}^{n}K_{j}^{2} 1 + ∑ j = 1 n K j 2 is positive: the sum is nonnegative by Finite Sums of Real Numbers: Nonnegativity, Domination by the Sum, and Limits (i) and 0 < 1 0<1 0 < 1 by claim 6 of Elementary Order Arithmetic in an Ordered Field , so claim 3 of Elementary Order Arithmetic in an Ordered Field applies. Hence 0 < C 0<C 0 < C by Properties of Real Powers (i) , and C 2 = 1 + ∑ j = 1 n K j 2 C^{2}=1+\sum_{j=1}^{n}K_{j}^{2} C 2 = 1 + ∑ j = 1 n K j 2 by (N1).
By Separability of the Lebesgue Spaces of the Torus (iii) , applied with the positive real ε / ( 4 C ) \varepsilon/(4C) ε / ( 4 C ) , there is s ∈ D s\in D s ∈ D with ∥ [ u ] − [ s ] ∥ L 2 ( T n ) ≤ ε / ( 4 C ) \bigl\lVert[u]-[s]\bigr\rVert_{L^{2}(\mathbb{T}^{n})}\le\varepsilon/(4C) [ u ] − [ s ] L 2 ( T n ) ≤ ε / ( 4 C ) . Put w = u − s w=u-s w = u − s , which lies in L 2 ( T n ) \mathcal{L}^{2}(\mathbb{T}^{n}) L 2 ( T n ) by Minkowski's Inequality (ii) , and note [ w ] = [ u ] − [ s ] [w]=[u]-[s] [ w ] = [ u ] − [ s ] by The Lebesgue Space Lᵖ(μ) (ii) , so that ∥ w ∥ 2 ≤ ε / ( 4 C ) \lVert w\rVert_{2}\le\varepsilon/(4C) ∥ w ∥ 2 ≤ ε / ( 4 C ) by The Lebesgue Space Lᵖ(μ) (iii) .
Put V = [ ( ρ 1 / k ⋆ s ) ∣ Q ] V=[(\rho_{1/k}\star s)|_{Q}] V = [( ρ 1/ k ⋆ s ) ∣ Q ] , an element of S \mathcal{S} S , and W = [ ( ρ 1 / k ⋆ w ) ∣ Q ] W=[(\rho_{1/k}\star w)|_{Q}] W = [( ρ 1/ k ⋆ w ) ∣ Q ] . By Properties of Periodic Convolution on the Torus (iii) applied with c = − 1 c=-1 c = − 1 we have ρ 1 / k ⋆ w = ρ 1 / k ⋆ u − ρ 1 / k ⋆ s \rho_{1/k}\star w=\rho_{1/k}\star u-\rho_{1/k}\star s ρ 1/ k ⋆ w = ρ 1/ k ⋆ u − ρ 1/ k ⋆ s as maps on R n \mathbb{R}^{n} R n ; all three maps lie in C p e r ∞ C^{\infty}_{\mathrm{per}} C per ∞ and hence in C p e r C_{\mathrm{per}} C per , so Continuous Periodic Functions are Power-Integrable and Dense on the Torus (iii) gives W = U k − V W=U_{k}-V W = U k − V in L 2 ( T n ) L^{2}(\mathbb{T}^{n}) L 2 ( T n ) . As U k U_{k} U k and V V V lie in H 1 ( T n ) H^{1}(\mathbb{T}^{n}) H 1 ( T n ) , so does W W W , and d H 1 ( U k , V ) = ∥ W ∥ H 1 d_{H^{1}}(U_{k},V)=\lVert W\rVert_{H^{1}} d H 1 ( U k , V ) = ∥ W ∥ H 1 by (N1).
We bound ∥ W ∥ H 1 \lVert W\rVert_{H^{1}} ∥ W ∥ H 1 . By Properties of Periodic Convolution on the Torus (i) , ∂ j ( ρ 1 / k ⋆ w ) = ( ∂ j ρ 1 / k ) ⋆ w \partial_{j}(\rho_{1/k}\star w)=(\partial_{j}\rho_{1/k})\star w ∂ j ( ρ 1/ k ⋆ w ) = ( ∂ j ρ 1/ k ) ⋆ w for every j ∈ [ n ] j\in[n] j ∈ [ n ] , so by Elementary Properties of the Weak Partial Derivative on the Torus (i) we have ∂ j W = [ ( ( ∂ j ρ 1 / k ) ⋆ w ) ∣ Q ] \partial_{j}W=[((\partial_{j}\rho_{1/k})\star w)|_{Q}] ∂ j W = [(( ∂ j ρ 1/ k ) ⋆ w ) ∣ Q ] . Applying The Lebesgue Bound for Periodic Convolution and Convergence of Mollifications (i) with p = 2 p=2 p = 2 , first with the kernel ρ 1 / k \rho_{1/k} ρ 1/ k , whose mass is 1 1 1 as computed in part (b) of the proof of claim 3, and then with the kernel ∂ j ρ 1 / k \partial_{j}\rho_{1/k} ∂ j ρ 1/ k , whose mass is K j K_{j} K j ,
∥ W ∥ L 2 ≤ ∥ w ∥ 2 , ∥ ∂ j W ∥ L 2 ≤ K j ∥ w ∥ 2 ( j ∈ [ n ] ) . \lVert W\rVert_{L^{2}}\le\lVert w\rVert_{2},\qquad\lVert\partial_{j}W\rVert_{L^{2}}\le K_{j}\lVert w\rVert_{2}\quad(j\in[n]). ∥ W ∥ L 2 ≤ ∥ w ∥ 2 , ∥ ∂ j W ∥ L 2 ≤ K j ∥ w ∥ 2 ( j ∈ [ n ]) .
Squaring by Properties of Real Powers (v) , using Properties of Real Powers (iii) to write ( K j ∥ w ∥ 2 ) 2 = K j 2 ∥ w ∥ 2 2 \bigl(K_{j}\lVert w\rVert_{2}\bigr)^{2}=K_{j}^{2}\lVert w\rVert_{2}^{2} ( K j ∥ w ∥ 2 ) 2 = K j 2 ∥ w ∥ 2 2 , comparing the sums term by term by claim 1 of Comparison and Absolute Value Bounds for Finite Sums of Real Numbers , and adding the resulting inequalities, (N2) gives
∥ W ∥ H 1 2 ≤ ∥ w ∥ 2 2 + ∑ j = 1 n K j 2 ∥ w ∥ 2 2 . \lVert W\rVert_{H^{1}}^{2}\le\lVert w\rVert_{2}^{2}+\sum_{j=1}^{n}K_{j}^{2}\,\lVert w\rVert_{2}^{2}. ∥ W ∥ H 1 2 ≤ ∥ w ∥ 2 2 + j = 1 ∑ n K j 2 ∥ w ∥ 2 2 .
By claims 1 and 4 of Properties of a Sum over a Finite Index Set , applied to the map j ↦ K j 2 j\mapsto K_{j}^{2} j ↦ K j 2 on [ n ] [n] [ n ] with the scalar ∥ w ∥ 2 2 \lVert w\rVert_{2}^{2} ∥ w ∥ 2 2 , the sum on the right equals ∥ w ∥ 2 2 ∑ j = 1 n K j 2 \lVert w\rVert_{2}^{2}\sum_{j=1}^{n}K_{j}^{2} ∥ w ∥ 2 2 ∑ j = 1 n K j 2 , so the right-hand side is ( 1 + ∑ j = 1 n K j 2 ) ∥ w ∥ 2 2 = C 2 ∥ w ∥ 2 2 \bigl(1+\sum_{j=1}^{n}K_{j}^{2}\bigr)\lVert w\rVert_{2}^{2}=C^{2}\lVert w\rVert_{2}^{2} ( 1 + ∑ j = 1 n K j 2 ) ∥ w ∥ 2 2 = C 2 ∥ w ∥ 2 2 , which equals ( C ∥ w ∥ 2 ) 2 \bigl(C\lVert w\rVert_{2}\bigr)^{2} ( C ∥ w ∥ 2 ) 2 by Properties of Real Powers (iii) . By (N1), ∥ W ∥ H 1 ≤ C ∥ w ∥ 2 \lVert W\rVert_{H^{1}}\le C\lVert w\rVert_{2} ∥ W ∥ H 1 ≤ C ∥ w ∥ 2 . Since 0 ≤ C 0\le C 0 ≤ C and ∥ w ∥ 2 ≤ ε / ( 4 C ) \lVert w\rVert_{2}\le\varepsilon/(4C) ∥ w ∥ 2 ≤ ε / ( 4 C ) , claim 5 of Elementary Arithmetic in an Ordered Field gives C ∥ w ∥ 2 ≤ C ⋅ ε / ( 4 C ) = ε / 4 C\lVert w\rVert_{2}\le C\cdot\varepsilon/(4C)=\varepsilon/4 C ∥ w ∥ 2 ≤ C ⋅ ε / ( 4 C ) = ε /4 .
Finally, d H 1 d_{H^{1}} d H 1 is a metric by The Norm Metric of a Real Inner Product Space: Triangle Inequalities, Limits and Continuity (iii) , so by the triangle inequality of Metric Space ,
d H 1 ( U , V ) ≤ d H 1 ( U , U k ) + d H 1 ( U k , V ) < ε / 2 + ε / 4 , d_{H^{1}}(U,V)\le d_{H^{1}}(U,U_{k})+d_{H^{1}}(U_{k},V)<\varepsilon/2+\varepsilon/4, d H 1 ( U , V ) ≤ d H 1 ( U , U k ) + d H 1 ( U k , V ) < ε /2 + ε /4 ,
and ε / 2 + ε / 4 < ε \varepsilon/2+\varepsilon/4<\varepsilon ε /2 + ε /4 < ε because ε / 4 \varepsilon/4 ε /4 is positive. Hence d H 1 ( U , V ) < ε d_{H^{1}}(U,V)<\varepsilon d H 1 ( U , V ) < ε with V ∈ S V\in\mathcal{S} V ∈ S .
Since ε \varepsilon ε was an arbitrary positive real, condition 3 of Characterization of the Closure in a Metric Space by Open Balls shows that U U U lies in the closure of S \mathcal{S} S ; since U U U was arbitrary, S \mathcal{S} S is dense in H 1 ( T n ) H^{1}(\mathbb{T}^{n}) H 1 ( T n ) by Dense Subset of a Topological Space . Being also countable, it witnesses that ( H 1 ( T n ) , d H 1 ) (H^{1}(\mathbb{T}^{n}),d_{H^{1}}) ( H 1 ( T n ) , d H 1 ) is separable .
Proof of claim 6. Let S 0 = { [ w ∣ Q ] : w ∈ C p e r ∞ } \mathcal{S}_{0}=\{[w|_{Q}]:w\in C^{\infty}_{\mathrm{per}}\} S 0 = {[ w ∣ Q ] : w ∈ C per ∞ } as in claim 4. By The Lebesgue Bound for Periodic Convolution and Convergence of Mollifications (iii) with p = 2 p=2 p = 2 , S 0 \mathcal{S}_{0} S 0 is dense in L 2 ( T n ) L^{2}(\mathbb{T}^{n}) L 2 ( T n ) , so its closure in L 2 ( T n ) L^{2}(\mathbb{T}^{n}) L 2 ( T n ) is L 2 ( T n ) L^{2}(\mathbb{T}^{n}) L 2 ( T n ) by Dense Subset of a Topological Space . By claim 4, S 0 ⊆ H 1 ( T n ) \mathcal{S}_{0}\subseteq H^{1}(\mathbb{T}^{n}) S 0 ⊆ H 1 ( T n ) , and H 1 ( T n ) ⊆ L 2 ( T n ) H^{1}(\mathbb{T}^{n})\subseteq L^{2}(\mathbb{T}^{n}) H 1 ( T n ) ⊆ L 2 ( T n ) by The Sobolev Space of Once Weakly Differentiable Square-Integrable Classes on the Torus (i) . Hence claim 5 of The Closure is the Smallest Closed Superset , applied in the topological space L 2 ( T n ) L^{2}(\mathbb{T}^{n}) L 2 ( T n ) , gives
L 2 ( T n ) = cl ( S 0 ) ⊆ cl ( H 1 ( T n ) ) ⊆ L 2 ( T n ) , L^{2}(\mathbb{T}^{n})=\operatorname{cl}(\mathcal{S}_{0})\subseteq\operatorname{cl}\bigl(H^{1}(\mathbb{T}^{n})\bigr)\subseteq L^{2}(\mathbb{T}^{n}), L 2 ( T n ) = cl ( S 0 ) ⊆ cl ( H 1 ( T n ) ) ⊆ L 2 ( T n ) ,
the closures being taken in L 2 ( T n ) L^{2}(\mathbb{T}^{n}) L 2 ( T n ) . Therefore the closure of H 1 ( T n ) H^{1}(\mathbb{T}^{n}) H 1 ( T n ) in L 2 ( T n ) L^{2}(\mathbb{T}^{n}) L 2 ( T n ) is L 2 ( T n ) L^{2}(\mathbb{T}^{n}) L 2 ( T n ) , that is, H 1 ( T n ) H^{1}(\mathbb{T}^{n}) H 1 ( T n ) is dense in L 2 ( T n ) L^{2}(\mathbb{T}^{n}) L 2 ( T n ) by Dense Subset of a Topological Space .