PWM control for half-bridge driver

Hello! I am building an induction heater device controlled by half-bridge driver. I've already successfully tested the Mazilli ZVS flyback driver, and now I would like to achieve heating control with a new scheme. Driver schematics will be shown below. MOSFETs are controlled by optocouplers which receive PWM signals from a microcontroller. PWM signals must have 53 KHz frequency and 180-degree phase shift. I would like to use my Arduino Nano to achieve this. For PWM I am planning to use pins 3 and 11 and therefore mess with Timer2. For now, I just want to create a code for testing the driver with an adjustable duty cycle by encoder rotation. I used Nano before with a MAX6675 scheme and thermocouple, SSD1306 OLED display and rotary encoder and I would like to keep them. My current questions are:

  1. Is it suitable to use Arduino Nano's timer2 to generate two PMW signals with desired frequency and phase shift? Link1 - according to this topic I have to use timer1 to achieve that. Another topic with Mega2560 and only phase shift looks useful too link2.

  2. Which PWM mode is preferable for my task?

  3. Is it correct that the duty cycle can only be within the 0-50% range for half-bridge driver control?

I will be glad if you could suggest a better approach if there is any.

Schematics

Some initial code without encoder part:

#define PWM1_PIN 3
#define PWM2_PIN 11

void setup() {
  pinMode(PWM1_PIN, OUTPUT);
  pinMode(PWM2_PIN, OUTPUT);

  // Phase shift
  TCCR2A |= (1 << COM2B0);

 // Code to set PWM frequency
 //  

  // Initial PWM duty cycle for both signals:
  int dutyCycle = 0;  // can be only from 0  to 127 ?
  OCR2A = (OCR2A * dutyCycle) / 255;
  OCR2B = (OCR2B * dutyCycle) / 255;
}

void loop() {
  // Some control logic with encoder
}

Thank you in advance!

Your problem seems to call for a full H-bridge driver. Why not?

did you check the optocoupler datasheet to see if they can cope with that frequency?

The pull-up FET Q4 needs to be bootstrapped..The TLP-250 driving it needs a bootstrap power supply..there are different ways to do that..Look at some application notes..I prefer TLP-350s..

Did you see U3 in the diagram?

yes.. for Q4 to drive the 'resonant load' to 22v, its opto-driver has to have isolated PS..aka bootstrap..

Hello everyone again! First of all, thank you for the answers! They gave me a direction for my research. It was true, I've select wrong optocouple at first. In the end I was able to build working half-bridge and consequently full bridge inverter scheme. I will share half-bridge scheme below. It utileses ESP32 as PWM signal source.

Schematics