You will probably have to create your own texture. See http://www.trainingtutorial.com/TAT5.htm.
I use the following workflow to create the bitmap:
1-Model the texture as an inventor part (as you've done with your attached file)
2-Rotate the part so that you are looking at the surface
3-Save a copy as bitmap
4-Open the bitmap with image editor of your choice (I use GIMP)
5-Crop the image to a tileable square
6-Select the background, and change the color to magenta (Inventor will treat magenta regions as transparent)
7-Make sure the very bottom right pixel is magenta (this will tell Inventor that the image has transparency)
I looked at some of my custom textures, and it appears you're right. I think you need to "tell" Inventor that the texture is transparent by naming the file as "whatever&.bmp" rather than "whatever.bmp" -- so the ampersand is the key. Then the color of the bottom right pixel defines the color in the bitmap that will be treated as transparent.
I like Alewar's approach to getting the BMP in the first place. I'll add a couple of thoughts:
1) By setting the background color prior to grabbing the screen capture as Alewar describes, you could likely save yourself a little image-editing time.
2) I create a specific material for the bar grating. Use your new image, and then set the density appropriately. i.e. 1 1/4" x 1/8" bar grating weighs something like 5.94 #/SF. This translates to something like .03311 #/in^3. Doing this gives you the weight of the grating.
3) As Brian says.... the lower right corner does not have to be magenta. The most common trick is:
- save the image with the _& name
- size the image 1px bigger than normal. if you've got a 64px image, set it to 65px square. Then set your lower right pixel to be whatever transparent color you want. AD just happens to use magenta, so it's really obvious when you're picking colors in the Styles box, which ones are transparent. No need to follow suit, but it seems like a reasonable convention. I stick with it.
4) Play with the image-size and scaling. Smaller images, will, I suspect (never tested it) result in better performance, but will definitely show more pixelation, obviously. The texture here is 64px, scaled up to 800% to approximate the spacing. As alewar says, you could also model full-size, grab the screen capture, and scale 100%.
You do get some funny wrapping of images around corners, etc, and you don't get a true sense of depth of the object. But at least you get the ability to see through the grating in the model, and it doesn't kill the performance so much.
Also, if you render these, they'll render reasonably (more or less). Light passes through them, they cast shadows, etc, but they basically behave as planar objects. So it's not 'correct', but it's not horrible, either.