Bicomplex Hardy Classes of Solutions to Higher-Order Vekua Equations

By Theorem 3.6, if f \in H^p(D,\mathbb{B}), then (f^+)^*, f^- \in H^p(D). By Theorem 2.14, since f_{nt} \in L^s(\partial D,\mathbb{B}), it follows that f^\pm_{nt} \in L^s(\partial D). Since

for all z \in D, it follows that (f^+)^*_{nt}, f^-_{nt} \in L^s(\partial D). By Theorem 2.26 statement (3), (f^+)^*,f^- \in H^s(D). Therefore, by Theorem 3.6, f \in H^s(D,\mathbb{B}).

Suppose there exists constants C>0, \alpha \geq 0, such that

for 0 \leq r < 1. So, by (?),

for all 0 \leq r < 1. Since f \in Hol(D,\mathbb{B}), it follows by Theorem 3.6 that (f^+)^*,f^- \in Hol(D). Since |(f^+)^*(re^{i\theta})| = |f^+(re^{i\theta})|, for all re^{i\theta}, it follows, by Theorem 2.5, that (f^+)^*_b, f^-_b exist. Since

for \phi \in C^\infty(\partial D,\mathbb{R}), it follows that f^+_b exists. Thus, for every \phi \in C^\infty(\partial D),

and the sum of limits on the right hand exists. Hence, f_b exists.

Suppose f_b exists. Since, for every \phi \in C^\infty(\partial D), we have

for every r, and

for every r, it follows that

and the limit on the right hand side exists. Thus, f^\pm_b exist. By similar reasoning as above, (f^+)^*_b exists. Since (f^+)^*, f^- \in Hol(D), it follows, by Theorem 2.5, that

and

Since f= p^+ f^+ + p^- f^-, it follows that

Finally, suppose that

Since (f^+)^*, f^- \in Hol(D) and are representable as distributional pairings against the Poisson integral, it follows, by Theorem 2.5, that there exist constants C_\pm > 0 and \alpha_\pm \geq 0 such that

for every re^{i\theta} \in D, and

for every re^{i\theta} \in D. Thus,

where C:= 2\max\{2^{-1/2}C_+, 2^{-1/2}C_-\} and \alpha := \max\{\alpha_+, \alpha_-\}, for every re^{i\theta} \in D.

By hypothesis, \bar\partial f= Af + B\overline{f} \in L^1(D, \mathbb{B}), so by Proposition 4.3,

Suppose that \varphi \in H^p(D,\mathbb{B}). Then, for r \in (0,1), we have

where C_p is a constant that depends on only p. By Theorem 2.18, since Af+B\overline{f}\in L^q(D,\mathbb{B}), q>2, it follows that T_\mathbb{B}(Af+B\overline{f}) \in C^{0,\alpha}(\overline{D},\mathbb{B}). Thus, there exists M>0 such that

for every z \in \overline{D}. Hence,

Since there is no dependence on r in the right hand side of the above inequality, we have

If 1 < q \leq 2 and p satisfies p < \frac{q}{2-q}, then returning to (Proof) we have

by Theorem 2.18. Since the right hand side has no dependence on r, we have

In either case f \in H^p_{A,B}(D,\mathbb{B}).

Now, suppose that f = \varphi + T_\mathbb{B}(Af + B\overline{f}) \in H^p_{A,B}(D,\mathbb{B}). Observe that, for r \in (0,1), we have

where C_p is a constant that depends on only p and M is a constant that bounds ||T_\mathbb{B}(Af + B\overline{f})(z)||_\mathbb{B}^p, for all z \in \overline{D}. So,

and \varphi \in H^p(D, \mathbb{B}). If 1 < q \leq 2 and p satisfies p < \frac{q}{2-q}, then returning to (?) and appealing to Theorem 2.18 once more, we have

Therefore,

and \varphi \in H^p(D,\mathbb{B}).

The representation is a consequence of Theorem 14 of [FundBicomplex]. See also Remark 38 of [BCBergman].

The inclusion of \phi in the specified class is a consequence of Theorem 2.18.

Suppose that f = \varphi e^\phi \in H_{A,0}(D,\mathbb{B}) and \varphi \in H^p(D,\mathbb{B}). Then, for each r \in (0,1), we have

Since e^\phi \in C^{0,\alpha}(\overline{D},\mathbb{B}), it follows that there exists M>0 such that ||e^{\phi(re^{i\theta})}||_{\mathbb{B}} < M, for all re^{i\theta} \in \overline{D}. Hence,

Thus,

and f \in H^p_{A,0}(D,\mathbb{B}).

Now, suppose that f = \varphi e^\phi \in H^p_{A,0}(D,\mathbb{B}). Since \phi \in C^{0,\alpha}(\overline{D},\mathbb{B}), by Theorem 4.5, it follows that (e^\phi)^+, (e^\phi)^- \in C^{0,\alpha}(\overline{D}), where e^\phi = p^+ (e^\phi)^+ + p^- (e^\phi)^-. Since continuous functions on compact sets are bounded, it follows that there exist M_+,M_-, m_+, m_- > 0 such that

and

for all re^{i\theta} \in \overline{D}. Since e^v is not a zero divisor for all v \in \mathbb{B}, it follows that e^v is invertible, for all v \in\mathbb{B}, see [FundBicomplex]. Let m:= \min\{m_+,m_-\} and M:= \max\{M_+,M_-\}. Then,

and

for all re^{i\theta} \in \overline{D}. Hence,

and

for all re^{i\theta} \in \overline{D}. Thus, for r \in (0,1), we have

Therefore,

where C is a constant that depends only on p and A, and \varphi \in H^p(D,\mathbb{B}).

By Theorem 4.6, if w \in H^p_{A,0}(D,\mathbb{B}), then w = e^\phi \varphi and \varphi \in H^p(D,\mathbb{B}). Since \varphi \in H^p(D,\mathbb{B}), it follows, by Proposition 3.8, that \varphi \in L^m(D,\mathbb{B}), for every 0 < m < 2p. Since A \in L^q(D), q>2, it follows that e^\phi \in C^{0, \alpha}(\overline{D},\mathbb{B}). So, there exists a positive real number M such that

for all z \in \overline{D}. Hence,

for every 0 < m < 2p. Therefore, w \in L^m(D,\mathbb{B}), for every 0 < m < 2p.

By Theorem 4.4, every f \in H^p_{A,B}(D,\mathbb{B}) such that \bar\partial f= Af + B\overline{f} \in L^q(D, \mathbb{B}), q>2, is representable as f = \varphi + T_\mathbb{B}(Af + B\overline{f}) with \varphi \in H^p(D,\mathbb{B}). By Proposition 3.8, \varphi \in L^m(D,\mathbb{B}), for every 0 < m < 2p. If q>2, then T_\mathbb{B}(Af + B\overline{f}) \in C^{0, \alpha}(\overline{D},\mathbb{B}) \subset L^m(D,\mathbb{B}), for every 0 < m < 2p. If 1 < q \leq 2, then T_\mathbb{B}(Af + B\overline{f}) \in L^\gamma(D, \mathbb{B}), for 1 < \gamma < \frac{2q}{2-q}, by Theorem 2.18. Since in this case we assume that p < \frac{q}{2-q}, it follows that every m < 2p < \frac{2q}{q-2} and T_\mathbb{B}(Af+ B\overline{f}) \in L^m(D,\mathbb{B}), for every 0 < m < 2p. Thus, in either case, we have

where C_m is a constant that depends only on m. Therefore, f \in L^m(D), for every 0 < m < 2p.

By Definition 2.15,

By Proposition 2.11,

where w^\pm, A^\pm, B^\pm : D\to\mathbb{C}. Observe that

So, if w solves \bar\partial w= Aw + B\overline{w}, then

Hence, by identifying the components of the idempotent representations on the left and right hand sides, we have

and

Since \frac{\partial w^+}{\partial z} = A^+ w^+ + B^+ (w^+)^*, it follows that

The other direction is clear by reversing the above computation.

By Theorem 4.9, w \in H_{A,B}(D,\mathbb{B}) if and only if (w^+)^* \in H_{(A^+)^*,(B^+)^*}(D) and w^- \in H_{A^-,B^-}(D).

Suppose that w \in H^p_{A,B}(D,\mathbb{B}). Since

for all z \in D, by (?), it follows that

Since |w^+(z)| = |(w^+)^*(z)|, for all z, it follows that (w^+)^* \in H^p_{(A^+)^*,(B^+)^*}(D) and w^- \in H^p_{A^-,B^-}(D).

Now, suppose that (w^+)^* \in H^p_{(A^+)^*,(B^+)^*}(D) and w^- \in H^p_{A^-,B^-}(D). Since

for every z \in D, by (?), it follows that, for each re^{i\theta} \in D, we have

where C_p is a constant that depends only on p. Since the right hand side of the above inequality has no dependce on r, it follows that

Therefore, w \in H^p_{A,B}(D,\mathbb{B}).

By Theorems 4.5 and 4.6, if A \in L^q(D,\mathbb{B}), q>2 and f \in H^p_{A,0}(D), then f = \varphi e^\phi and \varphi \in H^p(D,\mathbb{B}). By Theorem 3.7, since we assume that \varphi^+ \in H^p_w(D) and w and p satisfy the relationship that w \in L^\gamma(D) with \gamma>2 or p < \frac{\gamma}{2-\gamma}, it follows that \varphi has a nontangential boundary value \varphi_{nt} \in L^p(\partial D,\mathbb{B}) and

Since f = \varphi e^\phi, \varphi_{nt} exists and is in L^p(\partial D,\mathbb{B}), and e^\phi \in C^{0, \alpha}(\overline{D},\mathbb{B}) by Theorem 4.5, it follows that f_{nt} is in L^p(\partial D,\mathbb{B}). Also, for r \in (0,1), we have

where C_p is a constant that depends on only p and M is a constant that depends on only p and A. Since

it follows that

So,

Observe that the first integral summand in the right hand side of the above inequality will converge to 0 as r \nearrow 1 by Theorem 3.7, and the second and third integral summands in the right hand side of the above inequality will converge to 0 by Lebesgue’s Dominated Convergence Theorem as r \nearrow 1. Thus,

By Theorem 4.6, f = e^\phi \varphi, and \varphi \in H^p(D,\mathbb{B}). Since A \in L^q(D), q>2, it follows that e^\phi \in C^{0, \alpha}(\overline{D},\mathbb{B}). So, there exist real numbers m>0 and M>0 such that

for all z \in \overline{D}. So,

and \varphi_{nt} \in L^s(\partial D, \mathbb{B}). By Theorem 3.9, \varphi \in H^s(D, \mathbb{B}). By Theorem 4.6, f \in H^s_{A,0}(D,\mathbb{B}).

By Theorem 4.4, if f \in H^p_{A,B}(D,\mathbb{B}) with \bar\partial f = Af + B\overline{f} \in L^q(D), where q and p satisfy the hypothesis, then f = \varphi + T_\mathbb{B}(Af + B\overline{f}) and \varphi \in H^p(D,\mathbb{B}). By Theorem 3.7, \varphi has a nontangential boundary value \varphi_{nt} \in L^p(\partial D,\mathbb{B}) and

By Theorem 2.18, if p and q satisfy the hypothesis, then T_\mathbb{B}(Af + B\overline{f}) \in L^p(\partial D,\mathbb{B}) and

So, f_{nt} exists and is in L^p(\partial D,\mathbb{B}). Also, for r \in (0,1), we have

Therefore,

Let w \in H^p_{A,B}(D,\mathbb{B}). By Theorem 4.10, w = p^+w^+ + p^- w^-, where (w^+)^* \in H^p_{(A^+)^*,(B^+)^*}(D,\mathbb{B}) and w^- \in H^p_{A^-,B^-}(D). Suppose also that w^+ \in H^p_g(D), for some g \in L^\ell(D), \ell > 2 (or if 1 < \ell \leq 2, then p satisfies p < \frac{\ell}{2-\ell}). By Theorem 2.23 and statement (3) of Theorem 2.26, it follows that w^+ and w^- have nontangential boundary values w^+_{nt} and w^-_{nt}, respectively, w^+_{nt}, w^-_{nt} \in L^p(\partial D), and,

and

By linearity, w_{nt} = p^+ w^+_{nt} + p^- w^-_{nt} exists and is in L^p(\partial D,\mathbb{B}). Observe that, for each r \in (0,1),

where C_p is a constant that depends only on p. Since

and

it follows that

Therefore,

First, suppose w \in H_{A,B}(D,\mathbb{B}), w_b exists, and w_b \in L^p(\partial D). Since w_b exists and is in L^p(\partial D,\mathbb{B}), it follows, by Proposition 2.14, that w_b^\pm exist and w^\pm \in L^p(\partial D, \mathbb{C}). Since

for every real-valued \phi \in C^\infty(\partial D) and re^{i\theta} \in D, and w^+_b exists, it follows that (w^+)^*_b exists. By Theorem 4.10, (w^+)^* \in H_{(A^+)^*, (B^+)^*}(D) and w^- \in H_{A^-,B^-}(D). Since (w^+)^* \in H_{(A^+)^*, (B^+)^*}(D) and w^- \in H_{A^-,B^-}(D), (w^+)^*_b and w^-_b both exist, and (w^+)^*_b,w^-_b \in L^p(\partial D), it follows by statement (5) of Theorem 2.26 that (w^+)^* \in H^p_{(A^+)^*, (B^+)^*}(D) and w^- \in H^p_{A^-,B^-}(D). By Theorem 4.10, w \in H^p_{A,B}(D,\mathbb{B}).

Now, suppose w \in H^p_{A,B}(D,\mathbb{B}). By Theorem 4.10, (w^+)^* \in H^p_{(A^+)^*, (B^+)^*}(D) and w^- \in H^p_{A^-,B^-}(D). By statement (3) of Theorem 2.26, the nontangential boundary values (w^+)^*_{nt} and w^-_{nt} exist and are in L^p(\partial D). By statement (4) of Theorem 2.26, (w^+)^*, w^- \in L^1(D). By part (3) of Theorem 2.26 again,

and

By Theorem 2.20, since (w^+)^*, w^- \in L^1(D), (w^+)^*_{nt}, w^-_{nt} \in L^p(\partial D) \subset L^1(\partial D), and

and

it follows that (w^+)^*_{b} and w^-_{b} exist. Also, by Theorem 2.20 (w^+)^*_{nt} = (w^+)^*_{b} and w^-_{nt} = w^-_{b}.

By Theorem 4.5, every w \in H_{A,0}(D,\mathbb{B}), where A \in L^1(D), has a representation

where \varphi \in Hol(D,\mathbb{B}) and, if q>2, \phi \in C^{0, \alpha}(\overline{D},\mathbb{B}) \subset L^\infty(\overline{D},\mathbb{B}).

Suppose \varphi_b exists. By Theorem 3.10, there exists C> 0 and \alpha \geq 0 such that

for all re^{i\theta} \in D. Observe that

for every re^{i\theta} \in D. Letting \widetilde{C}:= C\sqrt{2}\,||e^{\phi}||_{L^\infty}, we have

for every re^{i\theta} \in D.

Now, assume there are constants C>0 and \alpha \geq 0 such that

for every re^{i\theta} \in D. Since e^\phi \in C^{0,\alpha}(\overline{D},\mathbb{B}) and e^\phi is neither zero nor a zero divisor, it follows that \left|\left|\frac{1}{e^{\phi(re^{i\theta})}}\right|\right|_{\mathbb{B}} is bounded, i.e., there exists N>0 such that

for all re^{i\theta} \in D. Therefore,

for every re^{i\theta} \in D. By Theorem 3.10, \varphi_b exists.

Observe that functions of the form f(z) = \sum_{k = 0}^{n-1} \widehat{z^*}^k \varphi_k(z), where \varphi_k \in Hol(D,\mathbb{B}), for every k, are elements of P^n(D,\mathbb{B}) by direct computation.

We prove the other direction by induction. The result is clear for n = 1. Suppose that, for every n satisfying 1 \leq n \leq m-1, that if f \in P^n(D, \mathbb{B}), then there exists a collection \{\varphi_k\}_{k=0}^{n-1} such that \varphi_k\in Hol(D,\mathbb{B}), for every k, and

Let g \in P^m(D, \mathbb{B}). So, \bar\partial g \in P^{m-1}(D, \mathbb{B}), and there exists a collection \{\varphi_k\}_{k=0}^{m-2} such that \varphi_k \in Hol(D,\mathbb{B}), for every k, and

Observe that

Thus, g - \sum_{k=0}^{m-2} \frac{1}{k+1}\widehat{z^*} ^{k+1} \varphi_k \in Hol(D,\mathbb{B}). Let

Let \phi_0 := \gamma and \phi_{k+1} := \frac{1}{k+1} \varphi_k, for every 0 \leq k \leq m-2. Then we rearrange to have

Suppose that \{\varphi_k\}_{k=0}^{n-1}\subset H^p(D,\mathbb{B}). Then

solves

and, for each r \in (0,1),

where C is a constant that depends on only p and n. Thus,

and f \in P^{n,p}(D, \mathbb{B}).

Now, suppose that f \in P^{n,p}(D,\mathbb{B}). By Theorem 5.3, it follows that there exist \varphi_k \in Hol(D,\mathbb{B}), for every k, such that

Observe that the explicit form of \varphi_k can be given by

for each k (see [l2poly] or [itvek, itvekbvp, WBD] for the motivation for this formula in the complex setting). Hence, for each k and r \in (0,1), we have

where C_p is a constant that depends on only p and n. Since f \in P^{n,p}(D, \mathbb{B}), it follows that

for all \ell satisfying 0 \leq \ell \leq n-1. Thus,

so

Therefore, \varphi_k \in H^p(D,\mathbb{B}), for every k.

By Theorem 5.4, every f \in P^{n,p}(D,\mathbb{B}) is representable as f = \sum_{k = 0}^{n-1} \widehat{z^*}^k f_k, where f_k \in H^p(D,\mathbb{B}), for every k. Since f_k \in H^p(D,\mathbb{B}), it follows by Proposition 3.8 that f_k \in L^m(D,\mathbb{B}), for every 0 < m < 2p. Observe that

where C is a constant that depends on n and m. Thus, f \in L^m(D,\mathbb{B}).

If \varphi_k^+ \in H^p_{w_k}(D) and w_k \in L^{q_k}(D), where q_k>2 or 1 < q_k \leq 2 and p < \frac{q_k}{2-q_k}, then, by Theorem 3.7, \varphi_k has a nontangential boundary value \varphi_{k,nt} \in L^p(\partial D, \mathbb{B}) and

Since each \varphi_k has a nontangential boundary value \varphi_{k,nt} and f = \sum_{k = 0}^{n-1} \widehat{z^*}^k \varphi_k, it follows that f_{nt} exists and is

Observe that

where C_p is a constant that depends on only p and n. Hence, f_{nt} \in L^p(\partial D, \mathbb{B}). Now, for r \in (0,1), we have

where C_p is a constant that depends only on p and is not necessarily the same from line to line. Observe that, for every k,

By Lebesgue’s Dominated Convergence Theorem,

for every k. Since

for every k, by Theorem 3.7, and

where C_{p,k} is a constant that depends on only p and k, for every k, it follows that

Thus,

Observe that functions of the form f(z) = \sum_{k = 0}^{n-1} \widehat{z^*}^k \varphi_k(z), where \varphi_k \in H_{A,0}(D,\mathbb{B}), for every k, are elements of V^n_{A,0}(D,\mathbb{B}) by direct computation.

We prove the other direction by induction. The result is clear for n = 1. Suppose that, for every n satisfying 1 \leq n \leq m-1, that if f \in V^n_{A,0}(D, \mathbb{B}), then there exists a collection \{\varphi_k\}_{k=0}^{n-1} such that \varphi_k\in H_{A,0}(D, \mathbb{B}), for every k, and

Let g \in V^m_{A,0}(D, \mathbb{B}). So, (\bar\partial-A) g \in V^{m-1}_{A,0}(D, \mathbb{B}), and there exists a collection \{\varphi_k\}_{k=0}^{m-2} such that \varphi_k \in H_{A,0}(D,\mathbb{B}), for every k, and

Observe that

Thus, g - \sum_{k=0}^{m-2} \frac{1}{k+1}\widehat{z^*} ^{k+1} \varphi_k \in H_{A,0}(D,\mathbb{B}). Let

Let \phi_0 := \gamma and \phi_{k+1} := \frac{1}{k+1} \varphi_k, for every 0 \leq k \leq m-2. Then we rearrange to have

First, suppose that f \in V^{n,p}_{A,0}(D,\mathbb{B}). By Theorem 6.3, there exists \varphi_{k} \in H_{A,0}(D,\mathbb{B}) such that

Observe that

for every k. For this formula in the complex setting, see Remark 3.2 of [WBD] (or Theorem 1 of [itvek] or Theorem 4.1 of [itvekbvp]). Now, we have

where C_{n,p} is a constant that depends only on n and p. Hence, \varphi_k \in H^p_{A,0}(D).

Next, suppose that \varphi_k \in H^p_{A,0}(D), for every k. Since

for every 0 \leq k \leq n-1, where C_{kj}= k!{j}\choose{j-k}, it follows that

where C_{n,p} is a constant that depends on only n and p.

By Theorem 6.4, every f \in V^{n,p}_{A,0}(D,\mathbb{B}) is representable as f = \sum_{k = 0}^{n-1} \widehat{z^*}^k f_k, where f_k \in H^p_{A,0}(D,\mathbb{B}). By Theorem 4.7, f_k \in L^m(D,\mathbb{B}), for every 0 < m < 2p and every k. Hence, we have

where C is a constant that depends on n and m. Therefore, f \in L^m(D,\mathbb{B}).

Let f \in V^{n,p}_{A,0}(D,\mathbb{B}). By Theorem 6.4,

where \varphi_k \in H^p_{A,0}(D,\mathbb{B}), for every k. If \varphi_k^+ \in H^p_{w_k}(D), where w_k \in L^{q_k}(D) with q_k>2 or 1 < q_k \leq 2 and p< \frac{q_k}{2-q_k}, for every k, it follows that \varphi_{k,nt} exists, is in L^p(\partial D), and

Since f = \sum_{k = 0}^{n-1} \widehat{z^*}^k \varphi_k, it follows that

exists, and, for r \in (0,1), we have

where C_{p,n} and \widetilde{C_{p,n}} are constants that depend on only p and n. Observe that

So,

where C_{p,n}, \widetilde{C_{p,n}}, and \widetilde{\widetilde{C_{p,n}}} are constants that depend on only p and n. Observe that, as r \nearrow 1, the first sum in the last line of the inequality converges to zero because of (Proof), and the second and third sums in the last line of the inequality converge to zero by Lebesgue’s Dominated Convergence Theorem. Therefore,

By Theorem 6.4, if f \in V^{n,p}_{A,0}(D,\mathbb{B}), then f = \sum_{k = 0}^{n-1} \widehat{z^*}^k f_k and f_k \in H^p_{A,0}(D,\mathbb{B}), for every k. Since every f_k \in H^p_{A,0}(D,\mathbb{B}) and each f_{k,nt} \in L^{s_k}(\partial D,\mathbb{B}), for some s_k > p, it follows, by Theorem 4.13, that f_k \in H^{s_k}_{A,0}(D,\mathbb{B}), for each k. Hence, every f_k \in H^s_{A,0}(D,\mathbb{B}), for every k, where s:= \min_{k}\{s_k\}. By Theorem 6.4, since every f_k \in H^s_{A,0}(D,\mathbb{B}), it follows that f \in V^{n,s}_{A,0}(D,\mathbb{B}).

Observe that functions of the form f(z) = \sum_{k = 0}^{n-1} (\widehat{z} + \widehat{z^*})^k \varphi_k(z), where \varphi_k \in H_{A,B}(D,\mathbb{B}), for every k, are elements of V^n_{A,B}(D,\mathbb{B}) by direct computation.

We prove the other direction by induction. The result is clear for n = 1. Suppose that, for every n satisfying 1 \leq n \leq m-1, that if f \in V^n_{A,B}(D, \mathbb{B}), then there exists a collection \{\varphi_k\}_{k=0}^{n-1} such that \varphi_k\in H_{A,B}(D,\mathbb{B}), for every k, and

Let g \in V^m_{A,B}(D, \mathbb{B}). So, (\bar\partial-A-BC(\cdot)) g \in V^{m-1}_{A,B}(D, \mathbb{B}), and there exists a collection \{\varphi_k\}_{k=0}^{m-2} such that \varphi_k \in H_{A,B}(D,\mathbb{B}), for every k, and

Observe that

Thus, g - \sum_{k=0}^{m-2} \frac{1}{k+1}(\widehat{z} + \widehat{z^*})^{k+1} \varphi_k \in H_{A,B}(D,\mathbb{B}). Let

Let \phi_0 := \gamma and \phi_{k+1} := \frac{1}{k+1} \varphi_k, for every 0 \leq k \leq m-2. Then we rearrange to have

First, suppose that f \in V^{n,p}_{A,B}(D,\mathbb{B}). By Theorem 6.8,

where

for every k. Observe that each \varphi_k is representable as

for every k. For this formula in the complex setting, see Remark 3.2 of [WBD] (or Theorem 1 of [itvek] or Theorem 4.1 of [itvekbvp]). Using this representation, we can see that, for r \in (0,1),

where C_{n,k,p} and C_{n,p} are constants that depend on only n, k, and p. Hence,

for every k, and \varphi_k \in H^p_{A,B}(D, \mathbb{B}), for every k.

Now, suppose that f=\sum_{k = 0}^{n-1} (\widehat{z} + \widehat{z^*})^k \varphi_k solves (\bar\partial - A - BC(\cdot))^n f = 0 and \varphi_k \in H^p_{A,B}(D,\mathbb{B}), for every k. Since

where C_{j,n} is a constant that depends on only j and k, for every 0 \leq j \leq n-1, it follows that, for r \in (0,1),

where C_{n,p} and C are constants that depend on only n and p. Thus,

for every 0 \leq j\leq n-1, and f \in V^{n,p}_{A,B}(D,\mathbb{B}).

By Theorem 6.9, if f \in V^{n,p}_{A,B}(D,\mathbb{B}), then \sum_{k = 0}^{n-1} (\widehat{z} + \widehat{z^*})^k f_k, where every f_k \in H^p_{A,B}(D,\mathbb{B}). Since f_k \in H^p_{A,B}(D,\mathbb{B}) and \bar\partial \varphi_k \in L^{q_k}(D,\mathbb{B}), q_k>2 or 1 < q_k \leq 2 and p < \frac{q_k}{2 - q_k}, for every k, it follows by Theorem 4.8 that f_k \in L^m(D,\mathbb{B}), for every 0 < m < 2p. Thus,

where C is a constant that depends on n and m. Therefore, f \in L^m(D,\mathbb{B}).

By Theorem 6.9, if f \in V^{n,p}_{A,B}(D,\mathbb{B}), then f = \sum_{k = 0}^{n-1} (\widehat{z} + \widehat{z^*})^k \varphi_k and \varphi_k \in H^p_{A,B}(D,\mathbb{B}), for every k. By Theorem 4.4, if \bar\partial \varphi_k \in L^{q_k}(D,\mathbb{B}), q_k > 2, \varphi_k is representable as

where h_k \in H^p(D,\mathbb{B}). By Theorem 4.14, if h_k^+ \in H^p_{\gamma_k}(D), \gamma_k \in L^{\sigma_k}(D), \sigma_k > 2 or 1 < \sigma_k \leq 2 and p < \frac{\sigma_k}{2-\sigma_k}, then h_k has a nontangential boundary value \varphi_{k,nt} \in L^p(\partial D, \mathbb{B}) and

By our hypotheses, this holds for every \varphi_k, 0 \leq k \leq n-1. So, f_{nt} = \sum_{k = 0}^{n-1} (\widehat{e^{i\theta}} + \widehat{(e^{i\theta})^*})^k \varphi_{k,nt} exists and is in L^p(\partial D, \mathbb{B}). Also, for r \in (0,1), we have

where C_p, C_{p,n}, and C_{p,n,k} are all constants that only depend on their subscripts. Observe that

So, for each k,

Hence,

Observe that the first sum in the last line of the inequality converges to zero as r \nearrow 1, by (Proof), and the second sum converges to zero as r\nearrow 1, by Lebesgue’s Dominated Convergence Theorem. Therefore,

By Theorem 6.9, if f \in V^{n,p}_{A,B}(D,\mathbb{B}), then f = \sum_{k = 0}^{n-1} (\widehat{z} + \widehat{z^*})^k \varphi_k and \varphi_k \in H^p_{A,B}(D,\mathbb{B}), for every k. By Theorem 4.10, if \varphi_k \in H^p_{A,B}(D,\mathbb{B}), for every k, then \varphi_k = p^+ \varphi_k^+ + p^- \varphi_k^-, for every k. Since \varphi_k^+ \in H^p_{\gamma_k}(D), where \gamma_k \in L^{\sigma_K}(D), \sigma_k > 2 or 1 < \sigma_k \leq 2 and p < \frac{\sigma_k}{2-\sigma_k}, for every k, it follows, by Theorem 4.15, that the nontangential boundary value \varphi_{k,nt} of \varphi_k exists, is in L^p(\partial D), and

for every k. The rest of the details are precisely the same as in the proof of Theorem 6.11 after (Proof).